Daniela Massiceti
Senior Researcher
About
I am a senior machine learning (ML) researcher in the Teachable AI Experiences (opens in new tab) team at Microsoft Research (opens in new tab) based in Sydney, Australia. I work at the intersection of ML and human-computer interaction and am primarily interested in the sociotechnical innovations we need to ensure AI systems work for those in the “tails” of the user distribution – from rethinking data collection and annotation pipelines, to ways of personalising large models to marginalised individuals and communities, to ensuring the evaluation frameworks we use drive research meaningfully forward.
Prior to joining MSR, I did my PhD in computer vision at the University of Oxford under Prof. Philip Torr, my Masters in Neuroscience also at the University of Oxford, and my Bachelors in Electrical and Computer Engineering at the University of Cape Town. I am passionate about diversity and inclusion in machine learning, and am an organiser of the Deep Learning Indaba. (opens in new tab)
Current work
Panel Discussion: Generative AI for Global Impact: Challenges and Opportunities
At the Microsoft Research Forum June 2024, Microsoft researchers discussed the challenges and opportunities of making AI more inclusive and impactful for everyone—from data that represents a broader range of communities and cultures to novel use cases for AI that are globally relevant.
Lightning Talk: Challenges and Opportunities of Large Multi-Modal Models for Blind and Low Vision Users: A Case Study of CLIP
At the Microsoft Research Forum June 2024, Daniela Massiceti delves into the transformative potential of multimodal models such as CLIP for assistive technologies. Specifically focusing on the blind/low-vision community, the talk explores the current distance from realizing this potential and the advancements needed to bridge this gap.
Announcing the ORBIT dataset: Advancing Real-World Few-Shot Learning using Teachable Object Recognition
In partnership with City, University of London, Microsoft Research introduces the ORBIT dataset - a dataset of videos captured by users who are blind and low vision of their personal objects. The dataset is accompanied by a new few-shot object recognition benchmark to drive progress on challenging real-world data.
Research Talk: Bucket of Me - Using Few-Shot Learning to Realize Teachable AI Systems
At the Microsoft Research Summit 2021, Daniela Massiceti explores how few-shot learning can realize a vision of teachable AI, giving users the agency to personalize their AI systems with their own “bucket of me.”
Panel Discussion: Pursuing a Resilient and Sustainable Global Society
To mark the 30th anniversary of Microsoft Research, Eric Horvitz, Chief Scientific Officer, hosts a panel discussion on the “Generations of Inspirational and Impactful Research” with scientists and engineers from Microsoft Research.