Distant conversational speech recognition: Challenges and Opportunities
State-of-the-art ASR systems excel on close-talk benchmarks but struggle with far-field conversational speech, where error rates remain above 20%. Current benchmark datasets inadequately assess generalization across domains and real-world conditions, often relying on oracle segmentation…
BRAIN SIGNALS TO ACTION: Monitoring and Explaining User Cognitive Load with Foundation Models
Passive monitoring of cognitive load can enable personalized user experiences and even accelerate human learning by leveraging closed-loop adaptive training systems. Electroencephalography (EEG) provides a cost-effective, non-invasive window into brain activity, yet conventional methods struggle…
Study Information: Designing appropriate crisis management strategies for AI
Study information page We are a group of researchers at Microsoft Research, Northwestern University, University of Minnesota, Dartmouth College, Mental Health America and LivingWorks looking to improve AI chatbots (e.g., ChatGPT, Copilot, Claude) for users…
Applied Scientist 2 – M365 org
Our team of Applied Scientists 2 is chartered to tackle a wide variety of challenges, from building data ingestion to curating data to fine tuning prompts and models. The scientists focus on evaluation metrics to…
Evaluating the Cultural Relevance of AI Models and Products: Learnings on Maternal Health ASR, Data Augmentation and User Testing Methods
How do we ensure that AI systems are not only technologically advanced but also deeply resonant and beneficial for the specific communities they aim to serve? As AI becomes more integrated into critical sectors like…