{"id":365522,"date":"2017-03-02T22:37:26","date_gmt":"2017-03-03T06:37:26","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-event&p=365522"},"modified":"2017-05-25T02:54:16","modified_gmt":"2017-05-25T09:54:16","slug":"microsoft-research-asia-symposium-on-collaborative-research","status":"publish","type":"msr-event","link":"https:\/\/www.microsoft.com\/en-us\/research\/event\/microsoft-research-asia-symposium-on-collaborative-research\/","title":{"rendered":"Microsoft Research Asia Symposium on Collaborative Research"},"content":{"rendered":"

Venues<\/strong>: Microsoft Research Asia<\/a><\/p>\n

Contact us<\/strong>: If you have any questions about this event, please send us an email at\u00a0ruil@microsoft.com<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

Venues: Microsoft Research Asia Contact us: If you have any questions about this event, please send us an email at\u00a0ruil@microsoft.com<\/p>\n","protected":false},"featured_media":330716,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr_startdate":"2017-05-18","msr_enddate":"2017-05-18","msr_location":"Beijing, China","msr_expirationdate":"","msr_event_recording_link":"","msr_event_link":"","msr_event_link_redirect":false,"msr_event_time":"","msr_hide_region":false,"msr_private_event":false,"footnotes":""},"research-area":[],"msr-region":[197903],"msr-event-type":[210063],"msr-video-type":[],"msr-locale":[268875],"msr-program-audience":[],"msr-post-option":[],"msr-impact-theme":[],"class_list":["post-365522","msr-event","type-msr-event","status-publish","has-post-thumbnail","hentry","msr-region-asia-pacific","msr-event-type-workshop","msr-locale-en_us"],"msr_about":"

Venues<\/strong>: Microsoft Research Asia<\/a><\/p>\r\n

Contact us<\/strong>: If you have any questions about this event, please send us an email at\u00a0ruil@microsoft.com<\/a><\/p>","tab-content":[{"id":0,"name":"Home","content":"

Welcome to 2017 Microsoft Research Asia Symposium on Collaborative Research. This symposium is organized by Microsoft Research Asia to not only share\u00a0the latest scientific achievements from cooperation with Academia in China, but also explore opportunities with great potential in the near future.<\/p>\r\n

Over the years, Microsoft Research Asia has been proactively collaborating with the academic community on joint research projects and programs in China. Relying on a world-class research teams, Microsoft Research Asia has built a productive and cooperative relationships with Universities and research institutions. Joined forces with top universities in China, Microsoft Research Asia has established several joint labs covering various of research directions.<\/p>\r\n

This symposium consists of keynotes, plenary sessions, break-out sessions, and technology demos and showcases. Specially, this time we would hold a workshop on the Microsoft Cognitive Toolkit<\/a> (previously known as CNTK), a\u00a0free, easy-to-use, open-source, commercial-grade toolkit that trains deep learning algorithms to learn like the human brain.<\/p>"},{"id":1,"name":"Agenda","content":"

May 17 2017 (Wednesday)<\/h2>\r\n\r\n\r\n\r\n\r\n\r\n
Time<\/th>\r\nSession<\/th>\r\nSpeaker<\/th>\r\n<\/tr>\r\n<\/thead>\r\n
14:00-17:00<\/td>\r\nWorkshop on the Microsoft Cognitive Toolkit<\/td>\r\nTaifeng Wang, Lead Researcher, Microsoft Research Asia<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>\r\n

May 18 2017 (Thursday)<\/h2>\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n
Time<\/th>\r\nSession<\/th>\r\nSpeaker<\/th>\r\n<\/tr>\r\n<\/thead>\r\n
09:20-09:50<\/td>\r\nWelcome Tea and Onsite Registration<\/td>\r\n<\/td>\r\n<\/tr>\r\n
09:50-09:55<\/td>\r\nOpening<\/td>\r\nTim Pan, Outreach\u00a0Director, Microsoft Research Asia<\/td>\r\n<\/tr>\r\n
10:00-10:40<\/td>\r\nKeynote Speech: Pushing the New Frontier of Artificial Intelligence<\/td>\r\nTie-Yan Liu, Principle Researcher, Microsoft Research Asia<\/td>\r\n<\/tr>\r\n
10:40-11:40<\/td>\r\nPanel Discussion: Turning Ideas into Reality<\/td>\r\nModerator<\/strong>: Tim Pan, Outreach\u00a0Director, Microsoft Research Asia\r\n\r\nPanelists<\/strong>:\r\nMinyi Guo, Professor, Shanghai Jiao Tong University\r\nMing Lei, CEO, Director of AI Innovation Center, Peking University; One of Seven Co-founders for Baidu\r\nBiao Cheng, Director of Incubation, Microsoft Research Asia\r\nJun Yan, Senior Researcher, Microsoft Research Asia<\/td>\r\n<\/tr>\r\n
11:40-12:00<\/td>\r\nGroup Photo with Tea Break<\/td>\r\n<\/td>\r\n<\/tr>\r\n
12:00-13:30<\/td>\r\nPoster\/Demo Sessions<\/td>\r\n<\/td>\r\n<\/tr>\r\n
13:30-15:10<\/td>\r\nTrack A<\/td>\r\nSpeakers<\/strong>:\r\nWenjun Zeng, MSRA\r\nJiangtao Wen, Tsinghua University\r\nYonghong Tian, Peking University\r\nJiwen Lu, Tsinghua University\r\nHong Cheng, UESTC<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack B<\/td>\r\nSpeakers<\/strong>:\r\nTao Qin, MSRA\r\nLiwei Wang, Peking University\r\nYang Yu, Nanjing University\r\nZhihua Zhang, Peking University\r\nYinghe Chen & Yongxia Shi, Beijing Normal University<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack C<\/td>\r\nSpeakers<\/strong>:\r\nXin Tong, MSRA\r\nZhouchen Lin, Peking University\r\nFeng Ye, Beijing Film Academy\r\nDongdong Weng, Beijing Institute of Technology\r\nJian Sun, Xi'an Jiaotong University<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack D<\/td>\r\nSpeakers<\/strong>:\r\nLintao Zhang, MSRA\r\nGuangyu Sun, Peking University\r\nChengchen Hu, Xi'an Jiaotong University\r\nXuanzhe Liu, Peking University\r\nJidong Zhai, Tsinghua University<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack E<\/td>\r\nSpeakers<\/strong>:\r\nTao Mei, MSRA\r\nYu-Gang Jiang, Fudan University\r\nXueming Qian, Xi'an Jiaotong University\r\nYong Hu, Beihang University\r\nWeiyao Lin, Shanghai Jiao Tong University<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack F<\/td>\r\nSpeakers<\/strong>:\r\nQiang Huo, MSRA\r\nNan Duan, MSRA\r\nJun Du, University of Science and Technology of China\r\nXin Zhang, South China University of Technology<\/td>\r\n<\/tr>\r\n
15:10-15:30<\/td>\r\nBreak<\/td>\r\n<\/td>\r\n<\/tr>\r\n
15:30-17:10<\/td>\r\nTrack A<\/td>\r\nSpeakers<\/strong>:\r\nFrank Soong, MSRA\r\nJun Zhu, Tsinghua University\r\nHaifeng Li, Harbin Institute of Technology\r\nBaoliang Lu, Shanghai Jiao Tong University\r\nZhaoxiang Zhang, Institute of Automation, Chinese Academy of Sciences<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack B<\/td>\r\nSpeakers<\/strong>:\r\nBin Shao, MSRA\r\nJie Tang, Tsinghua University\r\nXinbing Wang, Shanghai Jiao Tong University\r\nShuai Ma, Beihang University\r\nHongzhi Wang, Harbin Institute of Technology<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack C<\/td>\r\nSpeakers<\/strong>:\r\nChin-Yew Lin, MSRA\r\nJun Yan, MSRA\r\nYuanbin Wu, East China Normal University\r\nHailong Cao, Harbin Institute of Technology\r\nZhou Zhao, Zhejiang University<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack D<\/td>\r\nSpeakers<\/strong>:\r\nYongqiang Xiong, MSRA\r\nLiang Jeff Chen, MSRA\r\nHaisheng Tan, University of Science and Technology of China\r\nJiao Wang, Northeastern University<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack E<\/td>\r\nSpeakers<\/strong>:\r\nYu Zheng, MSRA\r\nYingcai Wu, Zhejiang University\r\nYihua Huang, Nanjing University\r\nLiqing Zhang, Shanghai Jiao Tong University\r\nLan Zhang, University of Science and Technology of China<\/td>\r\n<\/tr>\r\n
<\/td>\r\nTrack F<\/td>\r\nSpeakers<\/strong>:\r\nRuihua Song, MSRA\r\nZhiwen Yu, Northwestern Polytechnical University\r\nFuzhen Zhuang, Institute of Computing Technolgy, Chinese Academy of Sciences\r\nQi Liu, University of Science and Technology of China<\/td>\r\n<\/tr>\r\n
17:30-18:00<\/td>\r\nDeparture for banquet<\/td>\r\n<\/td>\r\n<\/tr>\r\n
18:00-20:00<\/td>\r\nBanquet<\/td>\r\n<\/td>\r\n<\/tr>\r\n
20:00-<\/td>\r\nReturn to hotel<\/td>\r\n<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>"},{"id":2,"name":"Abstracts","content":"[accordion]\r\n\r\n[panel header=\"Cross-lingual Word Pepresentation Learning Based on Distribution\"]\r\n\r\nSpeaker<\/strong>: Hailong Cao,\u00a0Harbin Institute of Technology\r\n\r\nIn this talk, I would like to introduce a distribution based model to learn bilingual word embeddings from monolingual data. It is simple, effective and does not require any parallel data or any seed lexicon. We take advantage of the fact that word embeddings are usually in form of dense real-valued low-dimensional vector and therefore the distribution of them can be accurately estimated. A novel cross-lingual learning objective is proposed which directly matches the distributions of word embeddings in one language with that in the other language. During the joint learning process, we dynamically estimate the distributions of word embeddings in two languages respectively and minimize the dissimilarity between them through standard back propagation algorithm. Our learned bilingual word embeddings allow to group each word and its translations together in the shared vector space. We demonstrate the utility of the learned embeddings on the task of finding word-to-word translations from monolingual corpora. Our model achieved encouraging performance on data in both related languages and substantially different languages.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"GraphView and Azure Cosmos DB Graph\"]\r\n\r\nSpeaker<\/strong>: Liang Jeff Chen, Microsoft Research Asia\r\n\r\nGraph data and applications are becoming ubiquitous. \u00a0Instead of following the research and industry trend of building graph databases and systems from scratch, we asked a simple yet fundamental question: what is the fundamental gap between graphs and old data formats and is it possible to bridge it? The answer to the question has since yielded a series of efforts and GraphView, a middleware that presents to applications a view of a graph database while internally compiles a graph program to data instructions that can be executed in existing database systems. As such, it re-uses the state-of-the-art technologies in today\u2019s database products and co-evolves with them for many years to come. This design philosophy not only inspires SQL Server\u2019s SQL Graph, but convinces Azure DocumentDB to take GraphView as the core and launch Azure Cosmos DB Graph, a multi-model database that empowers not only new graph customers but existing document-store and key-value-store customers. As of today, Azure Cosmos DB Graph is supporting a few key internal customers, as well as several external customers. We are expecting many more to come after its public announcement at \/\/Build.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Multisensor based Interaction and Learning for Human-Robot Hybrid Systems\"]\r\n\r\nSpeaker<\/strong>: Hong Cheng, University of Electronic Science and Technology of China (UESTC)\r\n\r\nRecently, human Robot Hybrid systems have been designed and developed to provide functional motion assistance to disabled and elderly people in daily activities. This talk will discuss key techniques of human-robot hybrid systems, which include ergonomics, multisensor based physical human-robot interaction, wearable computing, human intention estimation, multimodal interaction and cooperation. The related advances in UESTC exoskeletons will also be introduced in this talk, which includes reinforcement learning in pHRI and our exoskeleton systems, AIDER system for walking assistance and HUALEX system for human augmentation.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"The Deep Learning Based Speech Separation and Recognition\"]\r\n\r\nSpeaker<\/strong>: Jun Du, University of Science and Technology of China (USTC)\r\n\r\nWe will report our recent progress of deep learning based speech separation and recognition. We examine several critical problems in deep learning based\u807d speech separation, including synthesizing the noise data to improve the noise generalization, designing the objective function optimization under the probabilistic framework rather than the conventional MMSE, etc. For the recognition part, we will focus on the multi-channel case and share our thinking on how to fully utilize the multi-channel information under the deep learning framework. Specifically, the iterative mask estimation approach will be introduced as the core technology of our champion system in CHiME-4 challenge.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Building Informational Bot (InfoBot) with Question Generation & Answering\"]\r\n\r\nSpeaker<\/strong>: Nan Duan, Microsoft Research Asia\r\n\r\nIn this talk, I will briefly introduce how to build informational bots based on various genres of knowledge, using question generation (QG) and question answering (QA) technologies. Besides, I will present how infobot is applied in important Microsoft productions, such as Bing, Xiaoice, Customer Service Bot, and etc.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Rethink SDN: abstraction, overhead and development\"]\r\n\r\nSpeaker<\/strong>: Chengchen Hu, Xi'an Jiaotong University\r\n\r\nSoftware Defined Networking (SDN) greatly simplifies network management and introduces unprecedented flexibility by decoupling control functions from the network data plane. However, such a decoupling also opens a box of various open questions, which are not well addressed. This talk will briefly describe the work in XJTU towards more flexible SDN with programmable data plane, less overhead, and easier interoperability.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Deep Recommendation with Collaborative Filtering and Content Filtering\"]\r\n\r\nSpeaker<\/strong>: Yihua Huang, Nanjing University\r\n\r\nRecommendation system that recommends a user with the needed items plays a vital role in modern intelligent web and mobile applications. On one hand, good recommendation accuracy should take into consideration the auxiliary information from multi-sources other than conventional behavioral information. On the other hand, this also requires the innovation on recommendation model. We present RecNN, a pure deep neuron network (DNN) based framework, that seamlessly integrates both collaborative filtering and content based recommendation. The training data from multiple sources is considered as multiple views accordingly and each view contains information from user and item respectively. RecNN first utilizes different DNN structures to model the information such as the item description and user demographics into distributed embeddings. The inherent interactions are then modeled with two levels, intra-view and inter-view, in which each interaction is modeled as a DNN, and the results of intra-view DNNs are stacked together, followed by an inter-view DNN to make the final recommendation. The intra-view interaction aims to model the local user-item interaction, e.g., the collaborative filtering interaction is captured from the behavioral view, and the inter-view interaction models the global interaction among views. The evaluation against multiple recommendation tasks validates the effectiveness of RecNN.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Advances in Microsoft\u2019s Handwriting OCR Technology\"]\r\n\r\nSpeakers: Qiang Huo, Microsoft Research Asia\r\n\r\nOptical Character Recognition (OCR) is an important enabling technology to empower people to do more and achieve more. In Microsoft Research Asia (MSRA), we have been developing Microsoft\u2019s next generation OCR engines which can detect both printed and handwritten text in an image captured by a camera phone or glass, and recognize the detected text for follow-up actions. In this talk, I will give you a glimpse of what we have achieved so far for handwriting OCR and the road map ahead.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Dense Video Captioning\"]\r\n\r\nSpeaker<\/strong>: Yu-Gang Jiang, Fudan University\r\n\r\nIn this sharing talk, I will introduce our latest work on dense video captioning \u9225?automatically generating a paragraph of textual description for a given video. This goes one-step further beyond the recent works on generating a single sentence for a video clip.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Cognitive Principles for Audio Emotion Perception and Speech Emotion Recognition\"]\r\n\r\nSpeaker<\/strong>: Haifeng Li, Harbin Institute of Technology\r\n\r\n[\/panel]\r\n\r\n[panel header=\"From QA to Problem Solving\"]\r\n\r\nSpeakers: Chin-Yew Lin, Microsoft Research Asia\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Action Recognition and Summarization for Videos\"]\r\n\r\nSpeaker<\/strong>: Weiyao Lin, Shanghai Jiao Tong University\r\n\r\nAction recognition is an important yet challenging task in computer vision. In this talk, I will introduce a novel deep-based framework for action recognition, which improves the recognition accuracy by: 1) deriving more precise features for representing actions, and 2) reducing the asynchrony between different information streams. We first introduce a coarse-to-fine network which extracts shared deep features at different action class granularities and progressively integrates them to obtain a more accurate feature representation for input actions. We further introduce an asynchronous fusion network. It fuses information from different streams by asynchronously integrating stream-wise features at different time points, hence better leveraging the complementary information in different streams. Moreover, I will also introduce our work on summarizing long surveillance videos.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"The Shape Interaction Matrix-Based Affine Invariant Mismatch Removal for Partial-Duplicate Image Search\"]\r\n\r\nSpeaker<\/strong>: Zhouchen Lin, Peking University\r\n\r\nMismatch removal is a key step in many computer vision problems. In this paper, we handle the mismatch removal problem by adopting shape interaction matrix (SIM). Given the homogeneous coordinates of the two corresponding point sets, we first compute the SIMs of the two point sets. Then, we detect the mismatches by picking out the most different entries between the two SIMs. Even under strong affine transformations, outliers, noises, and burstiness, our method can still work well. Actually, this paper is the first non-iterative mismatch removal method that achieves affine invariance. Extensive results on synthetic 2D points matching data sets and real image matching data sets verify the effectiveness, efficiency, and robustness of our method in removing mismatches. Moreover, when applied to partialduplicate image search, our method reaches higher retrieval precisions with shorter time cost compared with the state-ofthe-art geometric verification methods.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Towards User-Friendly Mobile Web Computing\"]\r\n\r\nSpeaker<\/strong>: Xuanzhe Liu, Peking University\r\n\r\nWeb browsing is one of the most significant user requirements on mobile devices such as smartphones. However, the user\u00a0experience of mobile web browsing is undesirably sluggish because of the slow resource loading. We made a comprehensive measurement study to\u00a0uncover the resource update history and cache\u00a0configurations at the server side, and analyze the cache performance in various time granularities.\u00a0We investigate three main root causes:\u00a0Same\u00a0Content,\u00a0Heuristic Expiration<\/em>, and\u00a0Conservative Expiration Time.\u00a0<\/em>Based on these findings, we have developed two solutions to mitigate the imperfect resource-loading performance from different aspects. At the programming abstraction level, we propose\u00a0the ReWAP, which is based on an\u00a0efficient mechanism of resource packaging where stable resources are encapsulated and maintained into a package, and such a\u00a0package shall be loaded always from the local storage and updated by\u00a0explicitly refreshing.\u00a0Compared to the original mobile Web\u00a0apps with cache enabled, ReWAP can significantly reduce the data traffic, with the median saving up to 51%. In addition, ReWAP can\u00a0incur only very minor runtime overhead of the client-side browsers. At the system level, we propose the SWAROVsky,\u00a0a dual-proxy\u00a0system that comprises a remote cloud-side proxy and a local proxy on mobile devices. \u00a0Our system can be used with existing Web browsers and Web servers, and does not break the normal semantics\u00a0of a webpage. Evaluations with 50 websites show that on average our system can reduce the page load time by\u00a043.1%\u00a0and the\u00a0network data transmission by\u00a057.6%, while imposing marginal system overhead.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Emotion Recognition Using EEG and Eye Tracking Data\"]\r\n\r\nSpeaker<\/strong>: Bao-Liang Lu, Shanghai Jiao Tong University\r\n\r\nThe field of affective computing aspires to narrow the communicative gap between the highly emotional human and the emotionally challenged computers by developing computational systems that recognize and respond to human emotions. The detection and modeling of human emotions are the primary studies of affective computing. Among various approaches to emotion recognition, the electroencephalography (EEG)-based model is more reliable because of its high accuracy and objective evaluation in comparison with other external appearance clues like facial expression and gesture. Various psychophysiology studies have demonstrated the correlations between\u807d human emotions and EEG signals. In this talk, we will present our recent work on investigating critical frequency bands, critical channels, and the stable patterns over time, and developing emotion models with transfer learning and deep learning.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Deep Learning for Visual Analysis\"]\r\n\r\nSpeaker<\/strong>: Jiwen Lu, Tsinghua University\r\n\r\nIn this talk, we will briefly introduce some deep learning methods which are developed in our research group, including deep metric learning, deep hashing, multi-modal deep learning, and deep sharable feature learning. We also show their effectiveness in some vision applications such as face and person recognition, image and video search, and object tracking and recognition.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Enhancing Customer decisions: A Choice Modeling Perspective\"]\r\n\r\nSpeaker<\/strong>: Qi Liu, University of Science and Technology of China (USTC)\r\n\r\nThough personalized services (like recommender systems) are useful for handling information overload, it is still very challenging for customers to make the final choice, because the items in one consumption session are usually quite similar to each other. In this talk, we will briefly report our attempts to enhance the customer decisions via modeling the preferences and psychological traits of the customer in a particular session or sessions.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Object detection and retrieval in surveillance video\"]\r\n\r\nSpeaker<\/strong>: Xueming Qian, Xi'an Jiaotong University\r\n\r\nIn this talk, we will introduce our recent work on attribute based surveillance video retrieval systems. The main content of our talk includes the following parts: (1) Robust and efficient video objects detection and feature representation. The corresponding video objects include car, pedestrian, face. (2) Fast feature indexing and similarity measurement. (3) Demo.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Deep learning: challenges and opportunities\"]\r\n\r\nSpeaker<\/strong>: Tao Qin, Microsoft Research Asia\r\n\r\nDeep learning has achieved tremendous successes in recent years. It also faces multiple challenges, such as big-data challenge, big-model challenge, big-computation challenge, and so on. In this talk, I will first introduce those challenges and then discuss possible solutions and opportunities to address them.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"A World of Difference: Divergent Word Interpretations among People\"]\r\n\r\nSpeakers<\/strong>: Ruihua Song, Microsoft Research Asia\r\n\r\nA World of Difference: Divergent Word Interpretations among People\r\nAbstract: The divergent word usages reflect the differences among people. In this talk, we present a novel angle of studying word usage divergence \u2013 word interpretations. We propose an approach that quantifies semantic differences of interpretations among different groups of people. The effectiveness of our approach is validated by quantitative evaluations. The experiment results indicate that the divergences in word interpretations truly exist. We further apply the approach to two well studied types of differences between people \u2013 gender and region differences. The detected words with divergent interpretations reveal the unique features of specific groups of people. For the gender case, we discover that different interests, social attitudes, and characters between males and females result in their divergent interpretations of many words. For the region case, we find that specific interpretations of some words reveal the geographical and cultural features of different regions. Moreover, we further study the relation between word interpretation and frequency. Our results suggest that word interpretation and frequency, although both are effective indicators of word usages, are quite different.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Matching Speech in the Phonetic Space - Application to Voice Conversion and X-lingual TTS \"]\r\n\r\nSpeaker<\/strong>: Frank Soong, Microsoft Research Asia\r\n\r\nTo match speech segments from different speakers in the same or different languages, we need to equalize the acoustic difference between speakers as to measure the phonetic similarity at a relatively short, sub-phonemic, segment level. With a well-trained, speaker-independent, neural net (NN) based acoustic model, a speech segment is stochastically characterized by its sub-phonemic, \u201csenone\u201d posterior probability vector. An information-theoretic measure, Kullback-Leibler Divergence (KLD), is chosen to measure the phonetic \u201cdistance\u201d between two such derived posterior vectors. The proposed approach has many possible applications, including: 1. Voice Conversion\uff0ci.e., converting the voice timber from a source speaker to a target speaker but keeping the same word content of the sentence; 2. X-lingual TTS training, i.e., training TTS of a different (target) language by using the source speaker\u2019s monolingual speech data. We will present our NN-KLD based algorithms along with the voice conversion and cross-language TTS demos.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Towards Big Graph Search: Challenges and Techniques\"]\r\n\r\nSpeaker<\/strong>: Shuai Ma, Beihang University\r\n\r\nGraphs have more expressive power and are widely used today, and various applications of social computing trigger the pressing need of a new search paradigm. In this talk, we argue that graph search is the one filling this gap. We first introduce the application of graph search in various scenarios. We then formalize the graph search problem and briefly discuss its challenges. Finally, we introduce several useful query and data techniques towards efficient and effective big graph search.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Bridging Vision and Language with Deep Learning\"]\r\n\r\nSpeaker<\/strong>: Tao Mei, Microsoft Research Asia\r\n\r\nVisual recognition has been a fundamental challenge in computer vision for decades. Thanks to the recent development of deep learning techniques, researchers are striving to bridge vision (image and video) and natural language, which has become an emerging research area. We will present a few recent advances bridging vision and language with deep learning techniques, including image and video captioning, image and video chatting, storytelling, vision and language grounding, datasets, grand challenges, and open issues.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Microsoft Graph Engine and Its Applications\"]\r\n\r\nSpeaker<\/strong>: Bin Shao, Microsoft Research Asia\r\n\r\n[\/panel]\r\n\r\n[panel header=\" FP-DNN: An Automated Framework for Mapping Deep Neural Networks onto FPGAs with RTL-HLS Hybrid Templates\"]\r\n\r\nSpeaker<\/strong>: Guangyu Sun, Peking University\r\n\r\nDNNs (Deep Neural Networks) have demonstrated great success in numerous applications such as image classification, speech recognition, video analysis, etc. However, DNNs are much more computation-intensive and memory-intensive than previous shallow models. Thus, it is challenging to deploy DNNs in both large-scale data centers and real-time embedded systems. Considering performance, flexibility, and energy efficiency, FPGA-based accelerator for DNNs is a promising solution. Unfortunately, conventional accelerator design flows make it difficult for FPGA developers to keep up with the fast pace of innovations in DNNs.\r\n\r\nTo overcome this problem, we propose FP-DNN (Field Programmable DNN), an end-to-end framework that takes TensorFlow-described DNNs as input, and automatically generates the hardware implementations on FPGA boards with RTL-HLS hybrid templates. FP-DNN performs model inference of DNNs with our high-performance computation engine and carefully-designed communication optimization strategies. We implement CNNs, LSTM-RNNs, and Residual Nets with FPDNN, and experimental results show the great performance and flexibility provided by our proposed FP-DNN framework.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Deep Learning Approach for Model Learning in Image Processing and Analysis\"]\r\n\r\nSpeaker<\/strong>: Jian Sun, Xi'an Jiaotong University\r\n\r\nIn this talk, I will show that several mathematical models in imaging sciences, such as the sparsity-based models and statistical models, can be reformulated as deep learning models. We formulated Markov random field model in image prior modeling, iterative shrinkage in signal processing, compressive sensing model in MRI to be deep learning problems. These induced deep architectures are non-conventional, task-specific and achieved state-of-the-art results for solving image inverse problems, e.g., image restoration, compressive sensing MRI, etc.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Online Job Dispatching and Scheduling in Edge-Clouds\"]\r\n\r\nSpeaker<\/strong>: Haisheng Tan, University of Science and Technology of China\r\n\r\nIn edge-cloud computing, a set of edge servers are deployed near the mobile devices such that these devices can offload jobs to the servers with low latency. One fundamental and critical problem in edge-cloud systems is how to dispatch and schedule the jobs so that the job response time (defined as the interval between the release of a job and the arrival of the computation result at its device) is minimized. To study this problem, we propose a general model, where the jobs are generated in arbitrary order and times at the mobile devices and offloaded to servers with both upload and download delays. Our goal is to minimize the total weighted response time over all the jobs. The weight is set based on how latency sensitive the job is. We derive the first online job dispatching and scheduling algorithm in edge-clouds, called OnDisc, which is scalable <\/em>in the speed augmentation model. Moreover, OnDisc can be easily implemented in distributed systems. Extensive simulations on a real-world data-trace from Google show that OnDisc can reduce the total weighted response time dramatically compared with heuristic algorithms.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"MSRA\u56fe\u5f62\u7814\u7a76\u65b0\u8fdb\u5c55\"]\r\n\r\nSpeaker<\/strong>: Xin Tong, Microsoft Research Asia\r\n\r\n\u5728\u8fd9\u4e2a\u62a5\u544a\u4e2d\uff0c\u6211\u5c06\u4ecb\u7ecd\u6211\u4eec\u6700\u8fd1\u5728\u56fe\u5f62\u9886\u57df\u6240\u505a\u7684\u4e00\u4e9b\u7814\u7a76\u5de5\u4f5c\uff0c\u5305\u62ec\u4e09\u7ef4\u5185\u5bb9\u751f\u6210\uff0c\u51e0\u4f55\u5904\u7406\uff0c\u6750\u8d28\u5efa\u6a21\uff0c\u589e\u5f3a\u73b0\u5b9e\uff0c\u4ee5\u53ca\u53ef\u89c6\u5316\u65b9\u9762\u7684\u6700\u65b0\u8fdb\u5c55\uff0c\u5e76\u8ba8\u8bba\u548c\u5c55\u671b\u672a\u6765\u7684\u7814\u7a76\u65b9\u5411\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"AMiner: Mining Scientific Networks with AI\"]\r\n\r\nSpeaker<\/strong>: Jie Tang, Tsinghua University\r\n\r\nJie Tang is a Tenured associate professor with the Department of Computer Science and Technology at Tsinghua University, and was also visiting scholar at Cornell University, Hong Kong University of Science and Technology, and Southampton University. His interests include social network analysis, data mining, and machine learning. He has published more than 200 journal\/conference papers and holds 20 patents. His papers have been cited by more than 8,400 times. He served as PC Co-Chair of CIKM\u9225?6, WSDM\u9225?5, ASONAM\u9225?5, SocInfo\u9225?2, KDD-CUP\/Poster\/Workshop\/Local\/Publication Co-Chair of KDD\u9225?1-15, and Associate Editor-in-Chief of ACM TKDD, Editors of IEEE TKDE\/TBD and ACM TIST. He leads the project AMiner.org for academic social network analysis and mining, which has attracted more than 8 million independent IP accesses from 220 countries\/regions in the world. He was honored with the UK Royal Society-Newton Advanced Fellowship Award, CCF Young Scientist Award, and NSFC Excellent Young Scholar.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Big Graph Computation=Trinity+SQL\"]\r\n\r\nSpeaker<\/strong>: Feng Xiong, Harbin Institute of Technology\r\n\r\nBig graph computation becomes growingly popular since the explosion of data. To accomplish tradition tasks on it, researches start to design various parallel algorithms. However, the scalability and usability are circumscribed because algorithms can vary from different source schemas, processing techniques, tasks, etc. From solving this problem, we design a prototype system as a combination of distributed SQL engine and Trinity. In our system, big graphs are handled in a database way. Various common-used tasks are supported by our system. In this talk, I will introduce the system and its various applications such as community discovery, path matching and frequent subgraph mining.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Real-time multi-spectral camera using FPGA\"]\r\n\r\nSpeaker<\/strong>: Jiao Wang, Northeastern University\r\n\r\nWith the development of science and technology, multi-spectral camera has a great application in many fields, such as defense, medical, aerospace and aviation, etc. But a fatal defect in technology is that it will consume too much time to reconstruct spectral data cube on CPU or GPU. In our research , we design a prototype camera based on FPGA which can reach up to 20 fps under 200MHz, @256*256*15. It is the first multi-spectral camera can provide real-time performance in the world.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Realtime detection and tracking on embedded system\"]\r\n\r\nSpeaker<\/strong>: Yu Wang, Tsinghua University\r\n\r\n\u8fd1\u4e9b\u5e74\u6765\uff0c\u57fa\u4e8eCNN\u7684\u7269\u4f53\u68c0\u6d4b\u548c\u8ffd\u8e2a\u76f8\u8f83\u4e8e\u4f20\u7edf\u65b9\u6cd5\u5df2\u7ecf\u53d6\u5f97\u4e86\u5de8\u5927\u7684\u7a81\u7834\uff0c\u4ee5\u7269\u4f53\u68c0\u6d4b\u4e3a\u4f8b\uff0c\u4f20\u7edf\u68c0\u6d4b\u4e3b\u6d41\u7b97\u6cd5DPM\u5728VOC2007\u4e0a\u7684mAP\u53ea\u80fd\u8fbe\u523043%\uff0c\u4e0e\u4e4b\u76f8\u5bf9\u5e94\u7684\u662f\u57fa\u4e8eCNN\u7684faster R-CNN\u5728\u540c\u6837\u7684\u6570\u636e\u96c6\u4e0amAP\u8fbe\u5230\u4e8673%\uff0c\u6574\u6574\u63d0\u9ad8\u63d0\u9ad8\u4e8630\u4e2a\u767e\u5206\u70b9\u3002\u4f46\u662f\u7b97\u6cd5\u9ad8\u7cbe\u5ea6\u5f80\u5f80\u610f\u5473\u7740\u66f4\u5927\u7684\u8ba1\u7b97\u91cf\uff0c\u66f4\u5927\u7684\u53c2\u6570\u91cf\uff0c\u4f8b\u5982\uff0cfaster R-CNN\u7684\u8ba1\u7b97\u91cf\u8fbe\u5230\u4e86100G\uff0c\u53c2\u6570\u91cf\u8fbe\u5230\u4e86600M\uff08\u4f30\u8ba1\uff09\uff0c\u53d7\u9650\u4e8e\u7247\u4e0a\u6709\u9650\u7684\u8ba1\u7b97\u4e0e\u5b58\u50a8\u8d44\u6e90\uff0c\u6211\u4eec\u65e0\u6cd5\u5c06faster R-CNN\u76f4\u63a5\u79fb\u690d\u5230\u7247\u4e0a\uff0c\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e24\u4e2a\u4e3b\u8981\u7684\u51b2\u7a81\uff0c\u6211\u4eec\u5f00\u53d1DPU\u63d0\u9ad8\u7247\u4e0a\u7684\u8ba1\u7b97\u901f\u7387\uff0c\u91c7\u7528\u5b9a\u70b9\u538b\u7f29\u6280\u672f\u51cf\u5c11\u8ba1\u7b97\u91cf\uff0c\u6700\u7ec8\u5728\u7247\u4e0a\u6210\u529f\u90e8\u7f72\u4e863fps\u7684faster R-CNN\uff0c\u7cbe\u5ea6\u4e5f\u8fbe\u5230\u4e86\u4e16\u754c\u9886\u5148\u6c34\u5e73\u3002\u5728\u7269\u4f53\u8ffd\u8e2a\u65b9\u9762\uff0c\u6211\u4eec\u91c7\u7528\u7684\u662f\u76ee\u524d\u6d41\u884c\u7684KCF\u7b97\u6cd5\uff0c\u5e76\u4e3a\u4e86\u7247\u4e0a\u7684\u9ad8\u6548\u5b9e\u73b0\u5bf9\u7b97\u6cd5\u505a\u4e86\u4e00\u90e8\u5206\u6539\u5584\uff0c\u4f7f\u4e4b\u80fd\u591f\u5145\u5206\u5229\u7528\u7247\u4e0a\u7684\u8ba1\u7b97\u5b58\u50a8\u8d44\u6e90\uff0c\u901a\u8fc7\u8fd9\u79cd\u65b9\u5f0f\uff0c\u6211\u4eec\u5c06KCF\u5728ARM\u4e0a\u7684\u5355\u68465fps\u5b8c\u5584\u6210\u7247\u4e0a\u76845\u6846100fps\u3002\u4ee5\u6b64\u540c\u65f6\uff0c\u4e3a\u4e86\u66f4\u52a0\u9ad8\u6548\u5730\u8054\u5408\u7269\u4f53\u68c0\u6d4b\u548c\u8ffd\u8e2a\uff0c\u6211\u4eec\u8bbe\u8ba1\u4e86\u4e00\u79cd\u7b56\u7565\u4f7f\u5f97\u7269\u4f53\u68c0\u6d4b\u548c\u8ffd\u8e2a\u80fd\u591f\u66f4\u52a0\u9ad8\u6548\uff0c\u534f\u4f5c\u5730\u8fd0\u884c\uff0c\u6700\u7ec8\u5b9e\u73b0\u7684\u7247\u4e0a\u7684\u5b9e\u65f6\u7269\u4f53\u68c0\u6d4b\u4e0e\u8ffd\u8e2a\u7cfb\u7edf\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Machine Learning for Healthcare: Lung Cancer Detection via Deep Neural Networks\"]\r\n\r\nSpeaker<\/strong>: Liwei Wang, Peking University\r\n\r\nEarly detection of pulmonary cancer is the most promising way to enhance a patient\u9225\u6a9a chance for survival. Accurate pulmonary nodule detection in computed tomography (CT) images is a crucial step in diagnosing pulmonary cancer. In this talk, I will show our approach for pulmonary nodule detection based on DCNNs, which achieves the state of the art performance.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Paperbook: Design and Implementation\"]\r\n\r\nSpeaker<\/strong>: Xinbing Wang, Shanghai Jiao Tong University\r\n\r\nIn this talk, we conceptualize and design a novel academic system, paperbook or AceMap, to analyze the big scholarly data and present the results through a \u9225\u6e15ap\" approach. AceMap integrates several algorithms in the eld of network analysis and data mining, and then displays the information in a clear and intuitive way, aiming to help the researchers facilitate their work. After describing the big picture, we present achieved results and our work in progress. By far, AceMap has implemented the following functions: dynamic citation network display, paper clustering, academic genealogy, author and conference homepage, etc. We have also designed and performed distributed network analysis algorithms in a cutting-edge Spark system and utilized modern visualization tools to present the results. Finally, we conclude our paper by proposing the future outlooks.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"A universal low-latency real time optical flow based stereoscopic panoramic video communication system for AR\/VR\"]\r\n\r\nSpeaker<\/strong>: Jiaotao Wen, Tsinghua University\r\n\r\nWe introduce an optimized system for real time, low latency stereoscopic panoramic video communications that is camera agnostic. After intelligent camera calibration, the system is capable of stitching inputs from different cameras using a real time, low latency optical flow based algorithm that intelligently learns input video features over time to improve stitch quality. Depth information is also extracted in the process. The resulted stereoscopic panoramic video is then encoded with content-adaptive temporal and\/or spatial resolution to achieve low bitrate while maintaining good video quality. Various aspects of the system including the optimized stitching algorithm, parallelization and task scheduling, as well as encoding will be introduced with demos with conventional (non-panoramic) professional and consumer grade cameras as well as integrated panoramic cameras.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Extensible Wide Area Tracking System\"]\r\n\r\nSpeaker<\/strong>: Dongdong Weng, Beijing Institute of Technology\r\n\r\nThe existing optical tracking technology can be divided into active and passive systems. The active system is limited by its expensive cameras and not suitable for consumer market. Passive system, although the price is cheaper than the active one, but because of the signal interference between the base stations, its effective tracking area is not large. We proposed an extensible wide area tracking system which used the optical synchronous laser coding technology to prevent interference between scanning stations. In our system, the number of scanning base stations can be expanded from the current 2 ( HTC VIVE ) to dozens and the tracking area can reach 100 square meters.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"SmartAdp: Finding Optimal Billboard Locations from Large-Scale Taxi Trajectories\"]\r\n\r\nSpeaker<\/strong>: Yingcai Wu, Zhejiang University\r\n\r\nThe problem of formulating solutions immediately and comparing them rapidly for billboard placements has plagued advertising planners for a long time, owing to the lack of efficient tools for in-depth analyses to make informed decisions. In this talk, I will present our recent work that employs visual analytics combining the state-of-the-art mining and visualization techniques to tackle this problem using large-scale GPS trajectory data. In particular, we present SmartAdp, an interactive visual analytics system that deals with the two major challenges including finding good solutions in a huge solution space and comparing the solutions in a visual and intuitive manner. An interactive framework that integrates a novel visualization-driven data mining model enables advertising planners to effectively and efficiently formulate good candidate solutions.\u807d The presented approach can be adapted for other location selection problems such as selecting locations of retail stores or restaurants using trajectory data. More information about this work can be found here: http:\/\/www.ycwu.org\/projects\/smartadp.html<\/a>\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Realtime detection and tracking on embedded system\"]\r\n\r\nSpeaker<\/strong>: Yu Wang, Tsinghua University\r\n\r\n\u8fd1\u4e9b\u5e74\u6765\uff0c\u57fa\u4e8eCNN\u7684\u7269\u4f53\u68c0\u6d4b\u548c\u8ffd\u8e2a\u76f8\u8f83\u4e8e\u4f20\u7edf\u65b9\u6cd5\u5df2\u7ecf\u53d6\u5f97\u4e86\u5de8\u5927\u7684\u7a81\u7834\uff0c\u4ee5\u7269\u4f53\u68c0\u6d4b\u4e3a\u4f8b\uff0c\u4f20\u7edf\u68c0\u6d4b\u4e3b\u6d41\u7b97\u6cd5DPM\u5728VOC2007\u4e0a\u7684mAP\u53ea\u80fd\u8fbe\u523043%\uff0c\u4e0e\u4e4b\u76f8\u5bf9\u5e94\u7684\u662f\u57fa\u4e8eCNN\u7684faster R-CNN\u5728\u540c\u6837\u7684\u6570\u636e\u96c6\u4e0amAP\u8fbe\u5230\u4e8673%\uff0c\u6574\u6574\u63d0\u9ad8\u63d0\u9ad8\u4e8630\u4e2a\u767e\u5206\u70b9\u3002\u4f46\u662f\u7b97\u6cd5\u9ad8\u7cbe\u5ea6\u5f80\u5f80\u610f\u5473\u7740\u66f4\u5927\u7684\u8ba1\u7b97\u91cf\uff0c\u66f4\u5927\u7684\u53c2\u6570\u91cf\uff0c\u4f8b\u5982\uff0cfaster R-CNN\u7684\u8ba1\u7b97\u91cf\u8fbe\u5230\u4e86100G\uff0c\u53c2\u6570\u91cf\u8fbe\u5230\u4e86600M\uff08\u4f30\u8ba1\uff09\uff0c\u53d7\u9650\u4e8e\u7247\u4e0a\u6709\u9650\u7684\u8ba1\u7b97\u4e0e\u5b58\u50a8\u8d44\u6e90\uff0c\u6211\u4eec\u65e0\u6cd5\u5c06faster R-CNN\u76f4\u63a5\u79fb\u690d\u5230\u7247\u4e0a\uff0c\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e24\u4e2a\u4e3b\u8981\u7684\u51b2\u7a81\uff0c\u6211\u4eec\u5f00\u53d1DPU\u63d0\u9ad8\u7247\u4e0a\u7684\u8ba1\u7b97\u901f\u7387\uff0c\u91c7\u7528\u5b9a\u70b9\u538b\u7f29\u6280\u672f\u51cf\u5c11\u8ba1\u7b97\u91cf\uff0c\u6700\u7ec8\u5728\u7247\u4e0a\u6210\u529f\u90e8\u7f72\u4e863fps\u7684faster R-CNN\uff0c\u7cbe\u5ea6\u4e5f\u8fbe\u5230\u4e86\u4e16\u754c\u9886\u5148\u6c34\u5e73\u3002\u5728\u7269\u4f53\u8ffd\u8e2a\u65b9\u9762\uff0c\u6211\u4eec\u91c7\u7528\u7684\u662f\u76ee\u524d\u6d41\u884c\u7684KCF\u7b97\u6cd5\uff0c\u5e76\u4e3a\u4e86\u7247\u4e0a\u7684\u9ad8\u6548\u5b9e\u73b0\u5bf9\u7b97\u6cd5\u505a\u4e86\u4e00\u90e8\u5206\u6539\u5584\uff0c\u4f7f\u4e4b\u80fd\u591f\u5145\u5206\u5229\u7528\u7247\u4e0a\u7684\u8ba1\u7b97\u5b58\u50a8\u8d44\u6e90\uff0c\u901a\u8fc7\u8fd9\u79cd\u65b9\u5f0f\uff0c\u6211\u4eec\u5c06KCF\u5728ARM\u4e0a\u7684\u5355\u68465fps\u5b8c\u5584\u6210\u7247\u4e0a\u76845\u6846100fps\u3002\u4ee5\u6b64\u540c\u65f6\uff0c\u4e3a\u4e86\u66f4\u52a0\u9ad8\u6548\u5730\u8054\u5408\u7269\u4f53\u68c0\u6d4b\u548c\u8ffd\u8e2a\uff0c\u6211\u4eec\u8bbe\u8ba1\u4e86\u4e00\u79cd\u7b56\u7565\u4f7f\u5f97\u7269\u4f53\u68c0\u6d4b\u548c\u8ffd\u8e2a\u80fd\u591f\u66f4\u52a0\u9ad8\u6548\uff0c\u534f\u4f5c\u5730\u8fd0\u884c\uff0c\u6700\u7ec8\u5b9e\u73b0\u7684\u7247\u4e0a\u7684\u5b9e\u65f6\u7269\u4f53\u68c0\u6d4b\u4e0e\u8ffd\u8e2a\u7cfb\u7edf\u3002\r\n[\/panel]\r\n\r\n[panel header=\"Inference on Syntactic and Semantic Structures for Machine Comprehension\"]\r\n\r\nSpeaker<\/strong>: Yuanbin Wu, East China Normal University\r\n\r\nOpen domain machine comprehension is one of the major tasks in natural language processing. Being of great practical use, it attracts long lasting research interest. Both world knowledge and linguistic analysis are important for the task. In this talk, I will focus on answer reasoning with limited world knowledge following the setting of MCTest task. I will share our idea on using syntactic and semantic structures for answer inference, which helps us to better utilize prior linguistic structures and achieve competitive performances against popular deep learning models.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Cloud-Scale Multitenant IPsec Gateway \"]\r\n\r\nSpeaker<\/strong>: Yongqiang Xiong, Microsoft Research\r\n\r\nVirtual cloud network services let users have their own private networks in the public cloud. IPsec gateways are growing in importance accordingly since they provide VPN connections for customers to remotely access those private networks. Major cloud service providers (CSPs) offer IPsec gateway functions to tenants using virtual machines (VMs) running a software IPsec gateway inside. Those virtualized IPsec gateways enable CSPs to deploy a scalable and flexible VPN gateway service. However, dedicating individual IPsec gateway VMs to each tenant results in significant resource waste due to the strong isolation mechanism of VMs. We design Protego, a distributed IPsec gateway service designed for multitenancy. By separating the control plane and the data plane of an IPsec gateway, Protego achieves high availability with active redundancy. Furthermore, Protego can seamlessly migrate IPsec tunnels between the data nodes without compromising the throughput of them. Hence, it elastically scales in and out to adopt to the service traffic changes while guaranteeing an expected maximum throughput. Our evaluation, and simulation based on production data show that Protego together with a simple resource provisioning algorithm can save 84% of resources compared with allocating independent VMs.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"\u77e5\u8bc6\u6316\u6398\u53ca\u667a\u80fd\u5e94\u7528\"]\r\n\r\nSpeaker<\/strong>: Jun Yan, Microsoft Research Asia\r\n\r\n\u4eba\u5de5\u667a\u80fd\u7684\u5b66\u672f\u7814\u7a76\u4e0e\u5de5\u4e1a\u5e94\u7528\u6b63\u5f15\u8d77\u8d8a\u6765\u8d8a\u5e7f\u6cdb\u7684\u5173\u6ce8\u3002\u8ba1\u7b97\u673a\u5c31\u50cf\u4eba\u4e00\u6837\uff0c\u65e2\u9700\u8981\u806a\u660e\u7684\u5927\u8111\uff0c\u53c8\u9700\u8981\u535a\u5b66\u7684\u77e5\u8bc6\uff0c\u624d\u80fd\u4e3a\u4eba\u7c7b\u63d0\u4f9b\u771f\u6b63\u667a\u80fd\u7684\u670d\u52a1\u3002\u672c\u6b21\u62a5\u544a\u5c06\u91cd\u70b9\u5173\u6ce8\u5982\u4f55\u901a\u8fc7\u6570\u636e\u6316\u6398\u4e0e\u81ea\u7136\u8bed\u8a00\u5904\u7406\u6280\u672f\u8ba9\u8ba1\u7b97\u673a\u638c\u63e1\u77e5\u8bc6\u5e76\u5728\u5e94\u7528\u573a\u666f\u4e2d\u4f7f\u7528\u77e5\u8bc6\uff0c\u4ee5\u6b64\u6765\u89e3\u51b3\u5b9e\u9645\u5e94\u7528\u4e2d\u7684\u95ee\u9898\u3002\u6211\u4eec\u5c06\u4ece\u77e5\u8bc6\u7684\u5b9a\u4e49\uff0c\u62bd\u53d6\u65b9\u6cd5\uff0c\u8868\u793a\u65b9\u6cd5\u5f00\u59cb\uff0c\u8fdb\u800c\u7b80\u5355\u4ecb\u7ecd\u4e00\u4e9b\u77e5\u8bc6\u63a8\u7406\u4e0e\u8bed\u4e49\u8ba1\u7b97\u7684\u57fa\u672c\u60f3\u6cd5\uff0c\u6700\u540e\u6269\u5c55\u5230\u5982\u4f55\u901a\u8fc7\u77e5\u8bc6\u8ba1\u7b97\u8d4b\u4e88\u8ba1\u7b97\u673a\u4e00\u5b9a\u7684\u8054\u60f3\u4e0e\u521b\u9020\u80fd\u529b\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"4-7\u5c81\u513f\u7ae5\u4eb2\u5b50\u8bed\u8a00\u4e92\u52a8\u6a21\u5f0f\u7684\u53d1\u5c55\u7279\u70b9\u53ca\u5176\u5f71\u54cd\u56e0\u7d20\"]\r\n\r\nSpeaker<\/strong>: Yinghe Chen and Yongxia Shi, Beijing Normal University\r\n\r\n\u672c\u7814\u7a76\u62df\u8003\u5bdf4-7\u5c81\u513f\u7ae5\u7684\u4eb2\u5b50\u8a00\u8bed\u4e92\u52a8\u6a21\u5f0f\u7684\u53d1\u5c55\u7279\u70b9\u53ca\u5176\u5f71\u54cd\u56e0\u7d20\u3002\u6211\u4eec\u9080\u8bf7\u4e8650\u4e2a\u6765\u81ea\u5168\u56fd\u4e0d\u540c\u5730\u533a\u7684\u5bb6\u5ead\u53c2\u52a0\u4eb2\u5b50\u8a00\u8bed\u4e92\u52a8\u7684\u97f3\u9891\u5f55\u5236\uff0c\u8c08\u8bdd\u6d89\u53ca\u751f\u65e5\u6d3e\u5bf9\u3001\u90ca\u6e38\u3001\u4e0a\u5b66\u3001\u770b\u52a8\u753b\u7247\u7b49\u516b\u4e2a\u4e3b\u9898\uff1b\u6b64\u5916\u8fd8\u5bf9\u7236\u6bcd\u6559\u517b\u65b9\u5f0f\u548c\u513f\u7ae5\u884c\u4e3a\u7279\u70b9\u8fdb\u884c\u4e86\u5728\u7ebf\u6d4b\u8bc4\u3002\u5c4a\u65f6\u5c06\u901a\u8fc7\u7f16\u7801\u5206\u6790\u53ca\u4e92\u52a8\u97f3\u9891\u548c\u5f71\u50cf\u7684\u73b0\u573a\u5c55\u793a\uff0c\u5c55\u793a\u513f\u7ae5\u7684\u5e74\u9f84\u3001\u6027\u522b\u3001\u5bb6\u5ead\u80cc\u666f\u7b49\u5bf9\u4ea4\u6d41\u7684\u4e3b\u9898\u4fa7\u91cd\u70b9\u3001\u8bcd\u8bed\u60c5\u611f\u5c5e\u6027\u7b49\u65b9\u9762\u7684\u5f71\u54cd\uff0c\u5e76\u63a2\u8ba8\u7236\u6bcd\u6559\u517b\u65b9\u5f0f\u3001\u513f\u7ae5\u884c\u4e3a\u7279\u70b9\u5bf9\u4eb2\u5b50\u8a00\u8bed\u4e92\u52a8\u6a21\u5f0f\u7684\u5f71\u54cd\u3002\u5e0c\u671b\u57fa\u4e8e\u6211\u4eec\u7684\u524d\u671f\u7814\u7a76\u80fd\u4e3a\u540e\u671f\u7684\u6570\u5b57\u5316\u5de5\u4f5c\u63d0\u4f9b\u7d20\u6750\u4ee5\u53ca\u63d0\u793a\u540e\u671f\u7ed9\u4e88\u4e0d\u540c\u5e74\u9f84\u6bb5\u513f\u7ae5\u5728\u4eb2\u5b50\u5728\u4e92\u52a8\u4e3b\u9898\u7b49\u65b9\u9762\u5dee\u5f02\u7684\u5173\u6ce8\uff0c\u5e76\u57fa\u4e8e\u7236\u6bcd\u6559\u517b\u65b9\u5f0f\u548c\u513f\u7ae5\u884c\u4e3a\u7279\u70b9\u7b49\u8003\u8651\u6570\u5b57\u4ea7\u54c1\u7684\u5206\u7c7b\u6027\u548c\u591a\u5143\u6027\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"A RGBD Fusion Deep Neural Network Framework for Object Segmentation\"]\r\n\r\nSpeaker<\/strong>: Yong Hu, Beihang University\r\n\r\nStudy the problem of semantic object segmentation for RGBD cluttered scenes with a fusion deep CNN framework, which is composed of two different neural networks which is extended and fused from RGB to RGBD.The first neural network solves the problem of category-level semantic segmentation. The second region proposal network (RPN) solves the problem of object-level semantic segmentation.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Derivative-free Algorithms for Non-Convex Machine Learning\"]\r\n\r\nSpeaker<\/strong>: Yang Yu, Nanjing University\r\n\r\nMany machine learning tasks involve non-convex optimization problems, which can be non-differentiable, non-smooth and have many local optima. Derivative-free methods are suitable for these difficult problems, but were weak in theoretical foundation and practical scalability. This talk will introduce our recent progress in making theoretical-grounded derivative-free methods towards practical size non-convex machine learning tasks.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Predicting Spatial-temporal Job Mobility Leveraging Heterogeneous Social Media Data\"]\r\n\r\nSpeaker<\/strong>: Zhiwen Yu, Northwestern Polytechnical University\r\n\r\nDiscovering the underlying motivation and regularity of job mobility is one of the primary goals in human resource research area. Traditional researches mainly rely on limited surveys to gain the resumes and personal data to drive their investigation, which makes it difficult while expanding the scale and time scope. Recently, the universalized Internet, especially online social and professional networks, have made that information digitized and publicly available. Online professional networks (OPNs) like LinkedIn maintain huge resume warehouses which are dynamically spanning career records from hundreds of industries and companies. Meanwhile, location-based social networks (LBSNs) like Foursquare traces the human trajectories from all over the world, carrying rich sentiment information about human daily activities including geographical, textual and social interaction data. The growing clues carried on heterogeneous social media (like OPNs and LBSNs) provide unprecedented opportunities to achieve spatial-temporal job mobility prediction in a meticulous way.\r\n\r\nIn this talk, I'll give an introduction of our research work on predicting spatial-temporal job mobility. Specifically, I will introduce the designed job change prediction framework for predicting the job change occasion. Besides, I'll talk about the proposed talent circle detection method for mining group level job transition patterns.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"High Performance Human Action Recognition from Pose\"]\r\n\r\nSpeaker<\/strong>: Wenjun Zeng, Microsoft Research Asia\r\n\r\nRecently computer vision and deep learning technologies have been significantly leveraged to turn raw video data into insights to facilitate various applications and services. Since human is the main subject in many videos, understanding human becomes a critical step in video understanding. In this talk, I will introduce Microsoft Research Asia\u2019s recent efforts on skeleton-based human action recognition.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"A Lightweight Performance Anomaly Online Detection Tool\"]\r\n\r\nSpeaker<\/strong>: Jidong Zhai, Tsinghua University\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Understanding and Protecting Big Personal Data\"]\r\n\r\nSpeaker<\/strong>: Lan Zhang, University of Science and Technology of China\r\n\r\nRecently, we have witnessed the rapid growth of personal data, which contains huge amounts of valuable information and also a lot of privacy. Zhang\u9225\u6a9a talk focuses on deep understanding and privacy protection of multi-source multi-modality personal data collected by mobile devices.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"KV-Direct: High-Performance In-Memory Key-Value Store with Programmable NIC\"]\r\n\r\nSpeaker: Lintao Zhang, Microsoft Research Asia\r\n\r\nPerformance of in-memory key-value store (KVS) continues to be of great importance as modern KVS goes beyond the traditional object-caching workload and becomes a key infrastructure to support distributed main-memory computation in data centers. In this talk, I\u2019ll introduce KV-Direct, a high performance KVS that leverages programmable NIC to extend RDMA primitives and enable remote direct key-value access to the main host memory. A single NIC KV-Direct is able to achieve up to 180 M key-value operations per second, equivalent to the throughput of tens of CPU cores. Compared with CPU based KVS implementation, KV-Direct improves power efficiency by 10\u223c20x, while keeping tail latency below 10 \u00b5s. Moreover, KV-Direct can achieve near linear scalability with multiple NICs. With 8 programmable NIC cards in a server we achieve over one billion KV operations per second per server node, which is almost an order-of-magnitude improvement over existing systems, setting a new milestone for a general distributed key-value store.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Bayesian Tensor Decomposition for multiway data completion\"]\r\n\r\nSpeaker<\/strong>: Liqing Zhang, Shanghai Jiao Tong University\r\n\r\nTensor is a generalized data representation of vectors and matrices to higher dimensions based on multilinear algebra. It enables one to effectively capture the multilinear structures of the data, which is usually available as a priori information about the data. We present a generative model for robust tensor factorization in the presence of both missing data and outliers. The objective is to explicitly infer the underlying low-CP-rank tensor capturing the global information and a sparse tensor capturing the outliers, thus providing the robust predictive distribution over missing entries. To identify the model, we develop an efficient variational inference under a fully Bayesian treatment, which can effectively prevent the overfitting problem and scales linearly with data size. The extensive experiments and comparisons with many state-of-the-art algorithms on both synthetic and real-world datasets demonstrate the superiorities of our method from several perspectives.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Hand gesture recognition with path signature feature\"]\r\n\r\nSpeaker<\/strong>: Xin Zhang, South China University of Technology\r\n\r\nHand gesture recognition in videos has been an important research topic due to its potential wide applications in human-computer interaction. One of the challenges is to design and extract discriminative features to represent spatial position and temporal dynamics. Path signature is a compact representation of any open loop trajectory, and it has been successfully applied to financial trend prediction, voice signal compression and hand-written character recognition, etc. Here, we further introduce path signature feature to encode trajectory information of hand gestures and incorporate it into the deep learning framework. Experiments on various dataset demonstrate better performance in comparison with state-of-art methods.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Brain Inspired Convolutional Neural Network\"]\r\n\r\nSpeaker<\/strong>: Zhaoxiang Zhang, Institute of Automation, Chinese Academy of Sciences\r\n\r\nComparing to the biological neural networks, current deep neural networks show significant limitations on over-simplified neuron models, rigid structures, as well as poor adaptabilities. Bio-inspired neural networks, trying to combine the academic advantages of the biological evidences and machine Learning, show a promising prospect of further improvements on effectiveness, robustness and autonomy in various vision applications.\r\n\r\nRecently, we carried out the preliminary exploration on combining biological evidences and deep neural networks. In this talk, I will introduce some achievements on this topic, includes \"diverse neuron type selection for convolutional neural networks\"; dynamic multi-task learning\" and \"random-shifting for effective receptive field selection\".\r\n\r\nFrom the improvements achieved by these investigations, we try to demonstrate the significant potential of developing visual computing models and methods by seeking inspirations from the human visual system.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"\u5927\u89c4\u6a21\u4f18\u5316\u95ee\u9898\u7684\u8fd1\u4f3c\u725b\u987f\u65b9\u6cd5\"]\r\n\r\nSpeaker<\/strong>: Zhihua Zhang, Peking University\r\n\r\n\u8bb8\u591a\u673a\u5668\u5b66\u4e60\u95ee\u9898\u90fd\u53ef\u4ee5\u88ab\u5b9a\u4e49\u4e3a\u4e00\u4e2a\u4f18\u5316\u95ee\u9898\uff0c\u56e0\u6b64\u6c42\u89e3\u5927\u89c4\u6a21\u7684\u4f18\u5316\u95ee\u9898\u662f\u673a\u5668\u5b66\u4e60\u4e00\u4e2a\u975e\u5e38\u5177\u6709\u6311\u6218\u7684\u65b9\u5411\u3002\u8fd9\u4e2a\u62a5\u544a\u8ba8\u8bba\u4e00\u7c7b\u8fd1\u4f3c\u4e8c\u9636\u7b97\u6cd5\uff0c\u5b83\u5305\u542b\u5b50\u91c7\u6837\u725b\u987f\u65b9\u6cd5\uff0c\u6982\u7565\u725b\u987f\u65b9\u6cd5\u4ee5\u53ca\u975e\u7cbe\u786e\u725b\u987f\u65b9\u6cd5\u7b49\u3002\u8fd9\u7c7b\u65b9\u6cd5\u5177\u6709\u6807\u51c6\u725b\u987f\u65b9\u6cd5\u7684\u8d85\u7ebf\u6027\u6536\u655b\u7387\uff0c\u4f46\u540c\u65f6\u5b83\u7684\u8ba1\u7b97\u590d\u6742\u5ea6\u5219\u76f8\u5bf9\u6bd4\u8f83\u4f4e\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Intelligent Question Answering\"]\r\n\r\nSpeaker<\/strong>: Zhou Zhao, Zhejiang University\r\n\r\nQuestion Answering is a challenging task in natural language processing, computer vision and machine learning, which provides the accurate answer to the reference contents according to the given question. In this talk, I will briefly introduce the recent progress in textual question answering, visual question answering and dialogue learning.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Urban Computing: Enabling Intelligent Cities with Big Data and AI Technology\"]\r\n\r\nSpeaker<\/strong>: Yu Zheng, Microsoft Research Asia\r\n\r\nIn this talk I will introduce an urban big data platform that can empower people to manage and mine knowledge from big data using AI technology. Two recent examples will be presented. One is about using Mobike\u2019s trajectory data to plan bike lanes in a city more effectively. The other is to predict the flow of crowds based on deep learning techniques.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Learning with Deep Generative Models\"]\r\n\r\nSpeaker<\/strong>: Jun Zhu, Tsinghua University\r\n\r\nDeep generative models are effective for unsupervised and semi-supervised learning. In this talk, I will present some recent work on max-margin learning of deep generative models and triple generative adversarial networks, which provide a game-theoretical framework for semi-supervised learning, with state-of-the-art performance.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"The Research on Recommender System based on Deep Learning\"]\r\n\r\nSpeaker<\/strong>: Fuzhen Zhuang, Institute of Computing Technology, Chinese Academy of Sciences\r\n\r\nRecommendation has provoked vast amount of attention and research in recent decades. Most previous works employ matrix factorization techniques to learn the latent factors of users and items. However, matrix factorization methods may not make full use of the limited information from rating or check-in matrices, and achieve unsatisfying results. Recently, deep learning has proven able to learn good representation, and thus we try to exploit deep learning for recommendation. Along this line, we propose a new representation learning framework called Recommendation via Dual-Autoencoder (ReDa). In which we simultaneously learn the new hidden representations of users and items using autoencoders, and minimize the deviations of training data by the learnt representations of users and items. Furthermore, we propose a collaborative ranking framework via REpresentAtion learning with Pair-wise constraints (REAP for short), in which autoencoder is used to simultaneously learn the latent factors of both users and items and pair-wise ranked loss defined by (user, item) pairs is considered. Extensive experiments demonstrate the effectiveness of the proposed models.\r\n\r\n[\/panel]\r\n\r\n[\/accordion]"},{"id":3,"name":"Speakers","content":"[accordion]\r\n\r\n[panel header=\"Hailong Cao, Harbin Institute of Technology\"]\r\n\r\nHailong Cao received his PhD from Harbin Institute of Technology (HIT) on 2006. Now he is a lecturer in HIT, the Machine Intelligence and Translation Lab (MITLAB). He is focusing on the teaching and research about natural language processing (NLP). He is interested in machine translation and syntactic parsing and other areas as well. His papers appeared on ACL, COLING and EMNLP etc.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Liang Jeff Chen, Microsoft Research Asia\"]\r\n\r\nLiang Jeff Chen is commonly referred to as Jeff in his professional career. Jeff is a member of\u00a0Cloud & Mobile Research Group. He is generally interested in data management and also worked\u00a0on text search and ranking before. Most recently, He\u00a0becomes a system researcher and build\u00a0NoSQL systems that use SQL databases as core engines. The long term\u00a0goal is to\u00a0bridge the fundamental gap between NoSQL challenges and fast-evolving SQL technologies, so that the efforts we invest today are as long lived as SQL for many years to come. Jeff obtained BS and MS from Tsinghua University, and PhD from UC San Diego.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Hong Cheng, University of Electronic Science and Technology of China\"]\r\n\r\nHong Cheng is a full professor of University of Electronic Science and Technology of China (UESTC), school of Automation and Engineering. He serves as an executive director of the Center for Robotics since 2014. He was a visiting scholar at School of Computer Science, Carnegie Mellon University, USA from 2006 to 2009. Before this, he received his Ph.D degree in Pattern Recognition and Intelligent Systems from Xi\u2019 an Jiaotong University in 2003 and became an associate Professor of Xi\u2019 an Jiaotong University since 2005. He joined UESTC since 2010. His current research interests include machine learning in human robot hybrid systems. Prof. Cheng has over 100 academic publications including three books- \u201cDigital Signal Processing (Tsinghua University Press, Sep. 2007)\u201d, \u201cAutonomous Intelligent Vehicles: Theory, Algorithms and Implementation (Springer, Dec. He served\/is serving as a General Chair of VALSE 2015, Program Chair of CCPR 2016, and a General Chair for CCSR 2016. Now, he is a senior member of IEEE.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Jun Du, University of Science and Technology of China\"]\r\n\r\nJun Du received the B.Eng. and Ph.D. degrees from University of Science and Technology of China (USTC), in 2004 and 2009, respectively. From July 2009 to June 2010, he worked with iFlytek Research on speech recognition. From July 2010 to January 2013, he joined MSRA as an Associate Researcher, working on handwriting recognition, OCR, and speech recognition. Since February 2013, he has been with the National Engineering Laboratory for Speech and Language Information Processing of USTC as an Associate Professor. His research interests include speech signal processing and pattern recognition.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Minyi Guo, Shanghai Jiao Tong University\"]\r\n\r\nMinyi Guo is currently Zhiyuan Chair professor and chair of the Department of Computer Science and Engineering, Shanghai Jiao Tong University (SJTU), China. Before joined SJTU, Dr. Guo had been a professor of the school of computer science and engineering, University of Aizu, Japan. Dr. Guo received the national science fund for distinguished young scholars from NSFC in 2007, and was supported by \u201c1000 recruitment program of China\u201d in 2010. His present research interests include parallel\/distributed computing, compiler optimizations, embedded systems, pervasive computing, and cloud computing. He has more than 300 publications in major journals and international conferences in these areas. He is now on the editorial board of IEEE Transactions on Parallel and Distributed Systems and Journal of Parallel and Distributed Computing. Dr. Guo is a senior member of IEEE.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Chengchen Hu, Xi'an Jiaotong University\"]\r\n\r\nChengchen Hu received the Ph.D. degree from Tsinghua University, Beijing, China, in 2008. Now, he is a professor and the head of Department of Computer Science and Technology in XJTU. He is recipient of a fellowship from the European Research Consortium for Informatics and Mathematics (ERCIM), Microsoft \u201cStar-Track\u201d Young Faculty Program, New Century Excellent Talents in University awarded by Ministry of Education, China. Chengchen Hu\u2019s main research interests include network measurement, cloud data center networking, software defined networking.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yihua Huang, Nanjing University\"]\r\n\r\nMy name is Haipeng Zhang. I\u2019m currently a master student in PASA Big Data Lab at Nanjing University led by Prof. Yihua Huang. My research interests focus on big data parallel processing and recommendation algorithms. I have researched on deep learning for recommendation algorithms and system during my internship at Microsoft Research Asia from 07\/2016 to 01\/2017.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yu-Gang Jiang, Fudan University\"]\r\n\r\nYu-Gang Jiang is a Professor in School of Computer Science and Vice Director of Shanghai Engineering Research Center for Video Technology and System at Fudan University, China. His Lab for Big Video Data Analytics conducts research on all aspects of extracting high-level information from big video data, such as video event recognition, object\/scene recognition and large-scale visual search. He is the lead architect of a few best-performing video analytic systems in worldwide competitions such as the annual U.S. NIST TRECVID evaluation. His visual concept detector library (VIREO-374) and video datasets (e.g., CCV and FCVID) are widely used resources in the research community. His work has led to many awards, including \"emerging leader in multimedia\" award from IBM T.J. Watson Research in 2009, early career faculty award from Intel and China Computer Federation in 2013, the 2014 ACM China Rising Star Award, and the 2015 ACM SIGMM Rising Star Award.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Weiyao Lin, Shanghai Jiao Tong University\"]\r\n\r\nWeiyao Lin received the B.E. degree from Shanghai Jiao Tong University, China, in 2003, the M.E. degree from Shanghai Jiao Tong University, China, in 2005, and the Ph.D degree from the University of Washington, Seattle, USA, in 2010, all in electrical engineering. He is currently an Associate Professor with the Department of Electronic Engineering, Shanghai Jiao Tong University, China. He has authored or coauthored 90+ technical papers on top journals\/conferences including TPAMI, TIP, TCSVT, IJCV, CVPR, ACM MM, and ICCV. He holds 10 patents and has 10+ under reviewing patents. His research interests include video surveillance, video-based motion analysis, video compression & coding, and image\/video processing applications.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Zhouchen Lin, Peking University\"]\r\n\r\nZhouchen Lin received the Ph.D. degree in applied mathematics from Peking University in 2000. He is currently a Professor with the Key Laboratory of Machine Perception, School of Electronics Engineering and Computer Science, Peking University. His research areas include computer vision, image processing, machine learning, pattern recognition, and numerical optimization. He is an Area Chair of the CVPR 2014\/2016, the ICCV 2015, and the NIPS 2015, and a Senior Program Committee Member of the AAAI 2016\/2017 and the IJCAI 2016. He is an Associate Editor of the IEEE Transactions on Pattern Analysis and Machine Intelligence and the International Journal of Computer Vision. He is an IAPR Fellow.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Xuanzhe Liu, Peking University\"]\r\n\r\nXuanzhe Liu is now an associate professor of the Institute of Software, Peking University. He was a visiting researcher with Microsoft Research (with \"Star-Track Young Faculty Program\") from 2013-2014 and the winner of Microsoft Fellowship in 2007. He is now directing the SAAS (Systems, Applications, Analytics, and Services) research group Peking University. His recent research interests are focused on the software systems and engineering approaches for mobility and the Web, mostly from a data-driven perspective. My current projects cover the topics from measurement and performance evaluation of mobile systems\/web browsers, data-driven modeling and machine-learning based analytics of user behavior\/interactions, and so on. He has published over 60 referred papers at premier conferences such as WWW\/ICSE\/ OOPSLA\/UbiComp\/IMC and high-impact journals such as ACM TOIS\/TOIT and IEEE TSE\/TMC\/TSC.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Baoliang Lu, Shanghai Jiao Tong University\"]\r\n\r\nBao-Liang Lu received the Ph.D. degree in electrical engineering from Kyoto University, Kyoto, Japan, in 1994. From April 1994 to March 1999, He was a Frontier Researcher at the Bio-Mimetic Control Research Center, the Institute of Physical and Chemical Research (RIKEN), Japan. From April 1999 to August 2002, he joined the RIKEN Brain Science Institute, Japan, as a research scientist. Since August 2002, he has been a full professor at the Department of Computer Science and Engineering, Shanghai Jiao Tong University, China. He has been an adjunct professor of the Laboratory for Computational Biology, Shanghai Center for Systems Biomedicine since 2005. His research interests include brain-like computing, neural network, machine learning, brain-computer interface and affect computing. He was the past President of the Asia Pacific Neural Network Assembly (APNNA) and the general Chair of the 18th International Conference on Neural Information Processing(ICONIP2011). He is Associate Editors of the Neural Networks and IEEE Transactions on Cognitive and Developmental Systems, and a senior member of the IEEE. He is the directors of the Center for Brain-Like Computing and Machine Intelligence and the Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering, Shanghai Jiao Tong University.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Jiwen Lu, Tsinghua University\"]\r\n\r\nDr. Jiwen Lu is currently an associate professor with the Department of Automation, Tsinghua University, China. His current research interests include computer vision, pattern recognition, and machine learning. He has authored\/co-authored over 150 scientific papers in these areas, including 41 IEEE Transactions papers and 22 ICCV\/CVPR\/ECCV papers. He serves\/has served as an Associate Editor of Pattern Recognition Letters, Neurocomputing, the IEEE Access, a Guest Editor of five journals including Pattern Recognition, Computer Vision and Image Understanding, and Image and Vision Computing. He is\/was a Workshop Chair, Special Session Chair, or Area Chair for more than 10 international conferences.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Qi Liu, University of Science and Technology of China\"]\r\n\r\nShort bio Qi Liu is an Associate Professor in University of Science and Technology of China (USTC). His general area of research is data mining and knowledge discovery. He has published prolifically in refereed journals and conference proceedings, e.g., TKDE, TOIS, TKDD, TIST, KDD, IJCAI, AAAI, ICDM, SDM and CIKM. Dr. Liu is the recipient of the ICDM 2011 Best Research Paper Award, the Best of SDM 2015 Award, the Special Prize of President Scholarship for Postgraduate Students, Chinese Academy of Sciences (CAS) and the Distinguished Doctoral Dissertation Award of CAS.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Xueming Qian, Xi'an Jiaotong University\"]\r\n\r\nXueming Qian (M\u201910) received the B.S. and M.S. degrees in Xi\u2019an University of Technology, Xi\u2019an, China, in 1999 and 2004, respectively, and the Ph.D. degree in the School of Electronics and Information Engineering, Xi\u2019an Jiaotong University, Xi\u2019an, China, in 2008, after that he was an assistant professor. He was an associate professor from Nov. 2011 to March 2014, and now he was a full professor. He was awarded Microsoft fellowship in 2006. He was awarded outstanding doctoral dissertations of Xi\u2019an Jiaotong University and Shaanxi Province in 2010 and 2011 respectively. He is the director of SMILES LAB. He was a visit scholar at Microsoft research Asia from Aug. 2010 to March 2011. His research interests include social media big data mining and search. His research is supported by NSFC, Microsoft Research, and MOST.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Shuai Ma, Beihang University\"]\r\n\r\nShuai Ma is a full professor in the School of Computer Science and Engineering, Beihang University, China. He obtained two PhD degrees: University of Edinburgh in 2010 and Peking University in 2004, respectively. His research interests include database theory and systems, and he has published a number of papers in top conferences (SIMGOD, VLDB, ICDE, WWW, ICDM, MobiCom, USENIX ATC, WSDM) and journals (TODS, VLDB J, TKDE, TCS). He is a recipient of the best paper award of VLDB 2010, the best challenge paper award of WISE 2013, the National Science Fund of China for Excellent Young Scholars in 2013, and , and the second place in the final ranking of WSDM CUP 2016.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Guangyu Sun, Tsinghua University\"]\r\n\r\nDr. Guangyu Sun received his B.S. and M.S degrees from Tsinghua University, Beijing, in 2003 and 2006, respectively. He received his Ph.D. degree in Computer Science from the Pennsylvania State University in 2011. He joined the faculty of Center for Energy-Efficient Computing and Applications (CECA), School of EECS at Peking University from August 2011. His research interests include computer architecture, storage systems, and application-specific accelerator design.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Jian Sun, Xi'an Jiaotong University\"]\r\n\r\nJian Sun received Ph.D. in applied mathematics from Xi\u2019an Jiaotong University in 2009. Now he worked as a professor in School of Mathematics and Statistics, Xi\u2019an Jiaotong University. He focuses on the research of mathematical modeling in natural and medical image analysis. He worked as a visiting student in MSRA from 2015 to 2018, a postdoctoral researcher in INRIA and ENS from 2012 to 2014. He is a recipient of \u201cthe National Science Fund for Excellent Young Scholars\u201d (\u4f18\u9752).\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Haisheng Tan, University of Science and Technology of China\"]\r\n\r\nHaisheng Tan is currently an associate professor in School of Computer Science and Technology at University of Science and Technology of China (USTC), Hefei, China. He received his B.E. degree in Software Engineering (with the highest honor) and B.S. degree in Management both from USTC. Then, he got his Ph.D. degree in computer science at the University of Hong Kong (HKU) in 2011. After that, he was a postdoctoral fellow in Prof. Andrew Yao\u2019s group at Tsinghua University, Beijing. His research interests include algorithms and networking, mainly in the areas of wireless networking, data center networks and cloud computing. Dr. Tan has published over 30 papers in prestigious journals and conferences including theoretical journals as TCS, JOCO, IPL, and networking conferences as ACM MobiHoc and IEEE INFOCOM. For more information, please visit his webpage http:\/\/staff.ustc.edu.cn\/~hstan.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Jie Tang, Tsinghua University\"]\r\n\r\nAMiner is the second generation of the ArnetMiner system. We focus on developing author-centric analytic and mining tools for gaining a deep understanding of the large and heterogeneous networks formed by authors, papers, venues, and knowledge concepts. One fundamental goal is how to extract and integrate semantics from different sources. We have developed algorithms to automatically extract researchers\u2019 profiles from the Web and resolve the name ambiguity problem, and connect different professional networks. We also developed methodologies to incorporate knowledge from the Wikipedia and other sources into the system to bridge the gap between network science and the web mining research. In this talk, I will focus on answering two fundamental questions for author-centric network analysis: who is who? and who are similar to each other? The system has been in operation since 2006 and has collected more than 100,000,000 author profiles, 200,000,000 publication papers, and 7,800,000 knowledge concepts. It has been widely used for collaboration recommendation, similarity analysis, and community evolution.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Feng Xiong, Harbin Institute of Technology\"]\r\n\r\nFeng Xiong is a first year doctoral candidate at Harbin Institute of Technology. He also received both bachelor and master degree in computer science at Harbin Institute of Technology. His current research interests include big data computation, data quality management and big data analysis.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Jiao Wang, Northeastern University\"]\r\n\r\nJiao Wang received the Ph.D. degree in Pattern Recognition and Intelligent System from Northeastern University in 2006. He is now a professor in College of Information Science and Engineering, Northeastern university. His main re-search focuses on hardware computing and computer games.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yu Wang, Tsinghua University\"]\r\n\r\nYu Wang received his B.S. degree in 2002 and Ph.D. degree (with honor) in 2007 from Tsinghua University, Beijing. He is currently an associate professor with the Department of Electronic Engineering, Tsinghua University, Beijing. His research interests include parallel circuit analysis, application specific hardware computing (especially on the brain related problems), and power\/reliability aware system design methodology. Dr. Wang has authored and coauthored over 150 papers in refereed journals and conferences. He is the recipient of IBM X10 Faculty Award in 2010, the Best Paper Award in ISFPGA 2017, ISVLSI 2012, and 8 Best Paper Nominations in ASPDAC\/CODES\/ISLPED. He serves as the associate editor for IEEE Trans. CAD, Journal of Circuits, Systems, and Computers. He is the TPC Co-Chair of ICFPT 2011, finance chair of ISLPED 2012\u223c2015, and serves as TPC member in many important conferences (DAC, FPGA, DATE, ASPDAC, ISLPED, ISQED, ICFPT, ISVLSI, etc.).\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Liwei Wang, Peking University\"]\r\n\r\nLiwei Wang is a professor of School of Electronics Engineering and Computer Sciences, Peking University. His research interest is machine learning. He was named among \u201cAI\u2019s 10 to Watch\u201d in 2010.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Xinbing Wang, Shanghai Jiao Tong University\"]\r\n\r\nXinbing Wang received the B.S. degree (with hons.) in Automation from Shanghai Jiao Tong University, Shanghai, China, in 1998, the M.S. degree in computer science and technology from Tsinghua University, Beijing, China, in 2001, and the Ph.D. degree with a major in electrical and computer engineering and minor in mathematics from North Carolina State University, Raleigh, in 2006. Currently, he is a Professor in the Department of Electronic Engineering, and Department of Computer Science, Shanghai Jiao Tong University, Shanghai, China. Dr. Wang has been an Associate Editor for IEEE\/ACM Transactions on Networking,IEEE Transactions on Mobile Computing, and ACM Transactions on Sensor Networks. He has also been the Technical Program Committees of several conferences including ACM MobiCom 2012,2014, ACM MobiHoc 2012-2017, IEEE INFOCOM 2009-2017.\r\n\r\n[\/panel]\r\n\r\n \r\n\r\n[panel header=\"Dongdong Weng, Beijing Institute of Technology\"]\r\n\r\nDongdong Weng achieved his bachelor degree at Beijing Institute of Technology in 2001, and he achieved his Ph.D. degree at Beijing Institute of Technology in 2006. He worked in School of Optoelectronics at Beijing Institute of Technology from 2006. Until now, he is the associate research fellow and doctoral tutor of School of Optoelectronics in Beijing Institute of Technology. His research interest is focused on augmented reality, virtual reality and human-computer interaction. He has published more than 50 papers and instructed more than 30 graduates.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yingcai Wu, Zhejiang University\"]\r\n\r\nYingcai Wu is a National Youth-1000 scholar and a ZJU100 Young Professor at the State Key Lab of CAD & CG, Zhejiang University. He obtained his Ph.D. degree in Computer Science from the Hong Kong University of Science and Technology (HKUST). Prior to his current position, Yingcai Wu was a researcher in the Microsoft Research Asia, Beijing, China from 2012 to 2015, and a postdoctoral researcher in the University of California, Davis from 2010 to 2012.\r\nHis main research interests are in visual analytics, information visualization, and human computer interaction, with focuses on urban computing, social media analysis, text visualization, and behavior analysis. He has published more than 50 refereed papers and his three papers have been awarded Honorable Mention at IEEE VIS (SciVis) 2009, IEEE VIS (VAST) 2014, and IEEE PacificVis 2016.\r\nFor more information, visit www.ycwu.org\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yuanbin Wu, East China Normal University\"]\r\n\r\nYuanbin Wu is currently an assistant professor at Computer Science Department, East China Normal University. Before joining ECNU, he got both B.S. (2007) and Ph.D (2012) from Fudan University, and worked as a research fellow at National University of Singapore (2013). Yuanbin\u2019s research interests include question answering, information extraction and structured prediction algorithms in NLP.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yinghe Chen, Beijing Normal University\"]\r\n\r\n\u9648\u82f1\u548c\u535a\u58eb\uff0c\u6559\u6388\uff0c\u535a\u58eb\u751f\u5bfc\u5e08\uff0c\u5317\u4eac\u5e08\u8303\u5927\u5b66\u5fc3\u7406\u5b66\u90e8\u53d1\u5c55\u5fc3\u7406\u7814\u7a76\u9662\u526f\u9662\u957f\u3002\u6240\u5728\u5b66\u79d1\u4e3a\u53d1\u5c55\u4e0e\u6559\u80b2\u5fc3\u7406\u5b66\uff0c\u5177\u4f53\u7814\u7a76\u9886\u57df\u4e3a\u513f\u7ae5\u8ba4\u77e5\u5c55\uff0c\u8fd1\u5e74\u6765\u7684\u7814\u7a76\u5174\u8da3\uff1a\u513f\u7ae5\u8ba4\u77e5\u7b56\u7565\u3001\u6570\u8ba4\u77e5\u3001\u5143\u8ba4\u77e5\u3001\u5de5\u4f5c\u8bb0\u5fc6\u3001\u8868\u5f81\u3001\u60c5\u7eea\u8ba4\u77e5\u548c\u5fc3\u7406\u7406\u8bba\uff08TOM\uff09\u7b49\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yong Hu, Beihang University\"]\r\n\r\nAn Associate Professor from the State Key Lab of Virtual Reality Technology and Systems and the School of New Media Art & Design, Beihang University. From 2014 to 2015, he was a Visiting Researcher with Institute for Creative Technologies, University of Southern California. Over the past years, his research interests include appearance modeling, hand pose estimation, face modeling etc. He has authored and co-authored more than 25 scienti\ufb01c articles, including Graphics Interface, SGP, ACCV, VRST etc.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yongxia Shi, Beijing Normal University\"]\r\n\r\n\u65f6\u6c38\u971e\uff0c\u7814\u7a76\u5458\uff0c\u4fe1\u606f\u7f51\u7edc\u4e2d\u5fc3\uff0c\u4e3b\u8981\u7814\u7a76\u9886\u57df\uff1a\u6559\u5b66\u6570\u5b57\u5316\u3001\u8fdc\u7a0b\u6559\u80b2\u3001\u8ba1\u7b97\u673a\u6559\u5b66\u53ca\u7f51\u7edc\u5e73\u53f0\u7ba1\u7406.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Yang Yu, Nanjing University\"]\r\n\r\nDr. Yang Yu is an Associate Professor in the Department of Computer Science, Nanjing University, China. His research interest is in artificial intelligence, mainly on reinforcement learning, ensemble learning, and evolutionary computation for learning. His work has been published in Artificial Intelligence, IJCAI, AAAI, KDD, NIPS, ICDM, etc. He has been granted several awards such as the National Outstanding Doctoral Dissertation Award of China, and the best paper award of IDEAL\u201916, GECCO'11, PAKDD'08. He is\/was a Senior PC member of IJCAI\u201915\/17, a Publicity Chair of IJCAI\u201916\/17 and IEEE ICDM\u201916, a Workshop Chair of ACML'16.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Zhiwen Yu, Northwestern Polytechnical University\"]\r\n\r\nDr. Zhiwen Yu is currently a professor of the School of Computer Science, Northwestern Polytechnical University, P. R. China. He also serves as the director of the Department of Discipline Construction. He received his B.Eng, M.Eng and Ph.D. degree of Engineering in computer science and technology in 2000, 2003 and 2005 respectively from the Northwestern Polytechnical University. He has worked as a research fellow at the Academic Center for Computing and Media Studies, Kyoto University, Japan from Feb. 2007 to Jan. 2009, and a post-doctoral researcher at the Information Technology Center, Nagoya University, Japan in 2006-2007. He has been a visiting researcher at the Context-Aware Systems Department, Institute for Infocomm Research (I2R), Singapore from Sep. 2004 to May 2005. He has been an Alexander von Humboldt Fellow at the Mannheim University, Germany from Nov. 2009 to Oct. 2010.\r\n\r\nHe is the associate editor or editorial board of IEEE Transactions on Human-Machine Systems, IEEE Communications Magazine, ACM\/Springer Personal and Ubiquitous Computing (PUC), Entertainment Computing (Elsevier), and International Journal of Social Network Mining (IJSNM, Inderscience). He serves as the guest editor of ACM Transactions on Intelligent Systems and Technology, ACM Multimedia Systems Journal, Multimedia Tools and Applications (Springer), Pervasive and Mobile Computing (Elsevier), and Cybernetics and Systems (Taylor & Francis). He is the General Chair of the 8th IEEE International Conference on Cyber, Physical, and Social Computing (CPSCom 2015), and the 11th IEEE International Conference on Ubiquitous Intelligence and Computing (UIC 2014). He is the Program Chair of the 11th IEEE\/IFIP International Conference on Embedded and Ubiquitous Computing (EUC 2013), the 5th International Conference on Human-Centric Computing (HumanCom 2012), and the 7th International Conference on Ubiquitous Intelligence and Computing (UIC 2010). He serves as Vice Program Chair of the 13th IEEE International Conference on Pervasive Computing and Communications (PerCom 2015), the Workshop Chair of the 13th ACM International Conference on Ubiquitous Computing (UbiComp 2011), and the Publicity Chair of the Annual IEEE International Conference on Pervasive Computing and Communications (PerCom 2013\/2010). He has also served as PC member for a number of conferences, e.g., ACM Multimedia, IJCAI, IEEE PerCom, DEXA, GLOBECOM, ICC, ICME, IEEE CCNC, Euro-Par, MobiQuitous, etc.\r\n\r\nDr. Yu has published around 130 scientific papers in refereed journals and conferences, e.g., ACM Computing Surveys, IEEE TKDE, IEEE TMC, IEEE THMS, ACM TKDD, INFOCOM, UbiComp, PerCom, etc. His research interests cover pervasive computing, context-aware systems, human-computer interaction, mobile social networks, and personalization. Zhiwen Yu is a senior member of IEEE, a member of ACM, a distinguished member of CCF (China Computer Federation) and a senior member of CCF Pervasive Computing Technical Committee. He received the Young Teacher Award founded by Fok Ying Tong Education Foundation in 2014, the CCF Young Scientist Award in 2011, the CPSCom'13\/GPC'12\/AMT'12\/UIC'09 Best Paper Award, the Humboldt Fellowship in 2008, and the CCF Excellent Doctoral Dissertation Award in 2006.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Lan Zhang, University of Science and Technology of China\"]\r\n\r\nLan Zhang received her Bachelor degree (2007) in School of Software at Tsinghua University, China, and her Ph.D. degree (2014) in the department of Computer Science and Technology, Tsinghua University, China. She received 2015 ACM China Doctoral Dissertation Award (1\/2 nationally) and CCF Outstanding Doctoral Dissertation Award (1\/10 nationally). She is currently a researcher at the School of Computer Science and Technology, at University of Science and Technology of China. Her research interests span mobile computing, privacy protection, and data understanding. She has published 29 conference and journal papers, including 4 ACM MobiCom papers, 5 IEEE INFOCOM papers, etc. She has applied 3 United States patents and 17 Chinese patents, and 9 of them have been granted. She will be or has been TPC member of IEEE INFCOM 2018\uff0cIEEE ICC 2017, IEEE MASS 2017, MSN 2016, IEEE IPCCC 2016, IEEE DCOSS 2015, etc.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Liqing Zhang, Shanghai Jiao Tong University\"]\r\n\r\nHe received his PhD from Sun Yat-sen University in 1988. He was promoted a full professor of South China University of Technology. He joined RIKEN Brain Science Institute, Japan as research scientist in 1997. He now is a tenured professor of Shanghai Jiao Tong University. His research interests include computational theory for cortical networks, brain\u2013computer interface, statistical learning and inference.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Xin Zhang, South China University of Technology\"]\r\n\r\nDr. Xin Zhang received her B.S. degree in automatic engineering from Northwestern Polytechnical University, and the M.S. and Ph. D. degree in electrical engineering from Oklahoma State University, U.S. She is the Associate Professor in the School of Electronic and Information Engineering, South China University of Technology (SCUT). Her research interests include computer vision, image processing and machine learning. Dr. Zhang has published over 20 articles in journals, books, and conferences. She has also severed as the regular reviewer for many international conferences and journals.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Zhaoxiang Zhang, Chinese Academy of Sciences\"]\r\n\r\nZhaoxiang Zhang received the B.S. degree in electronic science and technology from the University of Science and Technology of China, in 2004, and the Ph.D. degree in pattern recognition and intelligent systems from the Institute of Automation, Chinese Academy of Sciences, in 2009. In 2009, he joined the School of Computer Science and Engineering, Beihang University, as an Assistant Professor and then an Associate Professor. In 2015, he joined the Institute of Automation, Chinese Academy of Sciences, as a Full Professor. Specifically, he is recently focusing on brain-inspired vision and humanlike learning. He has published around 80 papers in reputable journals and conferences. His research interests include computer vision, pattern recognition, and machine learning. He is the Associate Editor of Neurocomputing, involved on the Editorial Board of the Frontiers of Computer Science, the Program Committee Member of over ten international conferences, and the Reviewer of over 20 international journals. He has been granted several awards, including the MOE New Century Excellent Talents and the Beijing Youth Talents.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Zhihua Zhang, Peking University\"]\r\n\r\n\u5f20\u5fd7\u534e \u5317\u4eac\u5927\u5b66\u6570\u5b66\u79d1\u5b66\u5b66\u9662\u6559\u6388\uff0c\u5317\u4eac\u5927\u6570\u636e\u7814\u7a76\u9662\u6559\u6388\u3002\u4e4b\u524d\u66fe\u7ecf\u5148\u540e\u4efb\u6559\u4e8e\u4e0a\u6d77\u4ea4\u901a\u5927\u5b66\u548c\u6d59\u6c5f\u5927\u5b66\u3002\u4e3b\u8981\u4ece\u4e8b\u4e8e\u7edf\u8ba1\u673a\u5668\u5b66\u4e60\u4e0e\u4eba\u5de5\u667a\u80fd\u9886\u57df\u7684\u7814\u7a76\u548c\u6559\u5b66\u3002\u662f\u56fd\u9645\u673a\u5668\u5b66\u4e60\u520a\u7269JMLR\u7684\u6267\u884c\u7f16\u59d4\uff0c\u5e76\u591a\u6b21\u53d7\u9080\u62c5\u4efb\u56fd\u9645\u4eba\u5de5\u667a\u80fd\u9876\u7ea7\u5b66\u672f\u4f1a\u8bae\u7684\u7a0b\u5e8f\u59d4\u5458\u6216\u9ad8\u7ea7\u7a0b\u5e8f\u59d4\u5458\u3002\u5176\u7f51\u7edc\u516c\u5f00\u8bfe\u201c\u7edf\u8ba1\u673a\u5668\u5b66\u4e60\u201d\u548c\u201c\u673a\u5668\u5b66\u4e60\u5bfc\u8bba\u201d\u53d7\u5230\u5e7f\u6cdb\u5173\u6ce8\uff0c\u8fc4\u4eca\u4e3a\u6b62\u8bbf\u95ee\u91cf\u5df2\u8d85\u8fc750\u4e07\u6b21\u3002\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Zhou Zhao, Zhejiang University\"]\r\n\r\nZhou Zhao received the Ph.D. degrees in computer science from the Hong Kong University of Science and Technology (HKUST)in 2015. He is currently an associate professor with the College of Computer Science, Zhejiang University. His research interests include machine learning, data mining and natural language processing.\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Jun Zhu, Tsinhua University\"]\r\n\r\nJun Zhu is an Associate Professor at Department of Computer Science and Technology, Tsinghua University, and an Adjunct Faculty at Machine Learning Department, Carnegie Mellon University. His research interests lie in developing scalable machine learning methods to understand complex scientific and engineering data. Dr. Zhu has published over 80 peer-reviewed papers in the prestigious conferences and journals. He is an Associate Editor for IEEE Trans. on PAMI and Artificial Intelligence. He served as Area Chair for ICML, NIPS, UAI, IJCAI and AAAI. He was a local chair of ICML 2014. He is a recipient of the IEEE Intelligent Systems \"AI's 10 to Watch\" Award, NSFC Excellent Young Scholar Award, CCF Young Scientist Award, and CVIC SE Talents Award. His work is supported by the National \u201cTen Thousands Talents\u201d Program for Outstanding Young Scholars and Tsinghua \"221 Basic Research Plan for Young Talents\".\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Fuzhen Zhuang, Chinese Academy of Sciences\"]\r\n\r\nFuzhen Zhuang is an associate professor in the Institute of Computing Technology, Chinese Academy of Sciences. His research interests include transfer learning, multi-task learning, recommender systems, data mining, parallel classification and clustering. He has published more than 70 papers in some prestigious refereed journals and conference proceedings, such as IEEE TKDE, IEEE TOC, Information Sciences, IJCAI, AAAI, IEEE ICDE, WWW, ACM CIKM, ACM WSDM, SIAM SDM and IEEE ICDM. His papers about transfer learning have been selected as the best paper candidates in SDM 2010 and CIKM 2010. He is the recipient of the Doctoral Dissertation Award, Chinese Association for Artificial Intelligence; he wins the champion of data mining competition in IJCAI 2015.\r\n\r\n[\/panel]\r\n\r\n[\/accordion]"},{"id":4,"name":"Technology Showcases","content":"[accordion]\r\n\r\n[panel header=\"Exhibit 1: Chatting Robot With Behavior Learning\"]\r\n\r\nContact<\/strong>: Zhaoyuan Ma and Katsu Ikeuchi, MSRA Robotics and MS Strategic Prototyping Team\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 2: Microsoft Conversation Hub\"]\r\n\r\nContact<\/strong>: Lei Cui and Furu Wei, MSRA Natural Language Computing Group\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 3: Distant Interaction by Finger Pointing\"]\r\n\r\nContact<\/strong>: Minglei Li, Lei Sun and Qiang Huo, MSRA\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 4: \u77e5\u8bc6\u6316\u6398\u5728\u533b\u7597\u5927\u6570\u636e\u4e2d\u7684\u5e94\u7528\"]\r\n\r\nContact<\/strong>: DMEI@Microsoft\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 5: Fine-Grained Flower Recognition\"]\r\n\r\nContact<\/strong>: Jianlong Fu and Mei Tao, MSRA\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 6: Near-Field 3D Interaction System with Optical See-Through Head-Mounted Displays\"]\r\n\r\nContact<\/strong>: Dongdong Weng, Beijing Institute of Technology\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 7: Improving the Quality of User Experiences of Mobile Web Browsing\"]\r\n\r\nContact<\/strong>: Xuanzhe Liu and Yun Ma, Peking University; Yunxin Liu, MSRA\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 8: \u89c6\u9891\u6458\u8981\u53ca\u5f02\u5e38\u4e8b\u4ef6\u68c0\u6d4b\u7cfb\u7edf\"]\r\n\r\nContact<\/strong>: Weiyao Lin and Shihao Zhang, Shanghai Jiao Tong University\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 9: \u57fa\u4e8e\u865a\u62df\u673a\u5668\u4eba\u7684\u8fdc\u7a0b\u4eba\u673a\u4ea4\u4e92\u7cfb\u7edf\"]\r\n\r\nContact<\/strong>: Hong Cheng, UESTC\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 10: The Deep Learning Framework for Distant Speech Recognition\"]\r\n\r\nContact<\/strong>: Jun Du, University of Science and Technology of China; Qiang Huo, MSRA\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 11: SmartAdP - Visual Analytics of Large-scale Taxi Trajectories for Selecting Billboard Location \"]\r\n\r\nContact<\/strong>: Yingcai Wu, Zhejiang University\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 12: Leveraging Child Psychology to Improve Emotional Chat Engine for Kids\"]\r\n\r\nContact<\/strong>: Yinghe Chen and Yongxia Shi, Beijing Normal University Children Development Lab\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 13: Real-Time Optical Flow Based VR\/AR System\"]\r\n\r\nContact<\/strong>: Jiangtao Wen\u00a0and Yu Zhang, Tsinghua\u00a0University\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 14: FP-DNN (Field Programmable-DNN)\"]\r\n\r\nContact<\/strong>: Guangyu Sun and Yijin Guan, System Group, Peking University\r\n\r\n[\/panel]\r\n\r\n[panel header=\"Exhibit 15: Integrated Object Analysis in Surveillance Video\"]\r\n\r\nContact<\/strong>: Yuchan Liu and Xueming Qian, SMILES Lab, Xi'an Jiaotong University\r\n\r\n[\/panel]\r\n\r\n[\/accordion]"}],"msr_startdate":"2017-05-18","msr_enddate":"2017-05-18","msr_event_time":"","msr_location":"Beijing, China","msr_event_link":"","msr_event_recording_link":"","msr_startdate_formatted":"May 18, 2017","msr_register_text":"Watch now","msr_cta_link":"","msr_cta_text":"","msr_cta_bi_name":"","featured_image_thumbnail":"\"Artificial","event_excerpt":"Welcome to 2017 Microsoft Research Asia Symposium on Collaborative Research. This symposium is organized by Microsoft Research Asia to not only share\u00a0the latest scientific achievements from cooperation with Academia in China, but also explore opportunities with great potential in the near future. Over the years, Microsoft Research Asia has been proactively collaborating with the academic community on joint research projects and programs in China. Relying on a world-class research teams, Microsoft Research Asia has built…","msr_research_lab":[],"related-researchers":[],"msr_impact_theme":[],"related-academic-programs":[],"related-groups":[],"related-projects":[],"related-opportunities":[],"related-publications":[],"related-videos":[],"related-posts":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event\/365522"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-event"}],"version-history":[{"count":2,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event\/365522\/revisions"}],"predecessor-version":[{"id":386330,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event\/365522\/revisions\/386330"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/330716"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=365522"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=365522"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=365522"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=365522"},{"taxonomy":"msr-video-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video-type?post=365522"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=365522"},{"taxonomy":"msr-program-audience","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-program-audience?post=365522"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=365522"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=365522"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}