Advanced AI may make it easier for bad actors to deceive others online. A multidisciplinary research team is exploring one solution: a credential that allows people to show they’re not bots without sharing identifying information. Shrey Jain and Zoë Hitzig explain.
A lack of appropriate data, decreased model performance, and other obstacles have made it difficult to expand the input language models can receive. Li Lyna Zhang introduces LongRoPE, a method capable of extending content windows to more than 2 million tokens.
Senior Researcher Arindam Mitra introduces AgentInstruct. Using raw data sources, the automated multi-agent framework can create diverse, high-quality synthetic data at scale for the post-training of small and large language models.
Andrey Kolobov discusses WindSeer, a small CNN capable of estimating the wind field around an sUAV in flight more finely and with less compute and data than traditional models. The advancement can help support longer and safer autonomous flights.
Researcher Michel Galley explores how he and fellow researchers combined new and existing data to create MathVista, an open-source benchmark for measuring the mathematical reasoning capabilities of foundation models in scenarios that involve text and images.
Tusher Chakraborty talks about the paper “Spectrumize: Spectrum-efficient Satellite Networks for the Internet of Things,” including a method for supporting communication between a large IoT-satellite constellation and devices on Earth within a limited spectrum.
Senior Researcher Chang Liu discusses M-OFDFT, a variation of orbital-free density functional theory (OFDFT) that leverages deep learning to help identify molecular properties in a way that minimizes the tradeoff between accuracy and efficiency.
Can how we think about our thinking help us better incorporate generative AI in our lives & work? Explore metacognition’s potential to improve the tech’s usability on “Abstracts,” then sign up for Microsoft Research Forum for more on this & other AI work.
On “Abstracts,” Jordan Ash & Dipendra Misra discuss the parameter reduction method LASER. Tune in to learn how selective removal of stored data alone can boost LLM performance, then sign up for Microsoft Research Forum for more on LASER & related topics.
Members of the research community at Microsoft work continuously to advance their respective fields. Abstracts brings its audience to the cutting edge with them through short, compelling conversations about new and noteworthy achievements. In this episode, Senior Principal Research Manager Tao Qin and Senior Researcher…
By treating language models as layers in a network and prompts as learnable parameters, researchers aim for more adaptable, reusable LLM architectures. Check out the work in the “Abstracts” podcast series with guest Alessandro Sordoni and at #NeurIPS2023:
"Abstracts”—your source for world-class research in brief—welcomes Senior Principal Research Manager Xing Xie to the podcast series to discuss his paper on evaluating general-purpose AI with psychometrics.
