Linguistic Properties and Model Scale in Brain Encoding: From Small to Compressed Language Models
has been accepted as a spotlight paper at ICML 2026 (A*) (Top 2.2% of 23,918 submissions)
News
Our paper featured by EqualyzAI on Twitter (opens in new tab)
'How much AI do we really need to mirror the human brain? 3B-parameter models can predict brain activity about as well as much larger 14B models — brain alignment saturates at modest model sizes.' - EqualyzAI
New Preprint: How Does Longer Temporal Context Enhance Multimodal Narrative Video Processing in the Brain? (opens in new tab)
Our latest work exploring the impact of temporal context on multimodal narrative video processing in the brain is now available on arXiv.
New Preprint: Linguistic Properties and Model Scale in Brain Encoding: From Small to Compressed Language Models (opens in new tab)
Rejected at ICLR 2026 (A*). Meta-review: 'The paper would likely have landed just below or around the acceptance threshold.' [OpenReview]
Instruction-Tuned Video-Audio Models Elucidate Functional Specialization in the Brain (opens in new tab)
Decision: Rejected from ICLR 2026 (A*) with scores of 6, 6, 4. This is the second rejection despite positive reviews — after NeurIPS 2025 rejection (scores: 5, 5, 5, 3).
Joined ADIA (Abu Dhabi Investment Authority) Lab as a Research Scientist (opens in new tab)
I'm excited to announce that I've joined ADIA Lab (Abu Dhabi Investment Authority) as a Research Scientist, where I will be working on building AI agents for healthcare.
Our NeurIPS 2025 paper featured in Instagram video (opens in new tab)
'Brain-Informed Fine-Tuning for Improved Multilingual Understanding in Language Models' highlighted as one of the good papers.
Tutorial: Encoding and Decoding Language in the Brain with Language Models
has been accepted to EACL 2026! (A)
Aligning Text/Speech Representations from Multimodal Models with MEG Brain Activity During Listening
has been selected for SAC Highlight Award (Top 2% of accepted papers)
Invited to serve as a reviewer for Transactions on Machine Learning Research (TMLR). (opens in new tab)
I'm honored to join the reviewer pool for Transactions on Machine Learning Research (TMLR), a top-tier journal in the field.
Invited to serve as a reviewer for ACM Computing Surverys. (opens in new tab)
I'm honored to join the reviewer pool for ACM Computing Surveys, a top-tier journal in the field.
IndicSentEval: How Effectively do Multilingual Transformer Models encode Linguistic Properties for Indic Languages? (opens in new tab)
has been accepted to AACL-IJCNLP Findings 2025 (Findings, Long)
Invited to serve as a reviewer for IEEE Transactions on Neural Networks and Learning Systems (TNNLS). (opens in new tab)
I'm honored to join the reviewer pool for IEEE Transactions on Neural Networks and Learning Systems (TNNLS), a top-tier journal in the field.
Received 'Great Review' Recognition from ACL Rolling Review (opens in new tab)
Recognized by Area Chairs for providing an exceptional review in the ACL Rolling Review, noted for its clarity, helpfulness, and decisiveness.
I am serving as an Area Chair for the October 2025 ACL ARR cycle (EACL 2026) (opens in new tab)
I'm honored to serve as an Area Chair for ACL Rolling Review starting October 2025, which will handle submissions for EACL-2026.
Brain-Informed Fine-Tuning for Improved Multilingual Understanding in Language Models (opens in new tab)
has been accepted to NeurIPS Conference 2025 (Main Track, A*): Scores: 5,5,4,4,3
Instruction-Tuned Video-Audio Models Elucidate Functional Specialization in the Brain (opens in new tab)
Decision: Rejected from NeurIPS conference 2025 (Main Track, A*)-even with mostly positive reviews (scores: 5 (accept), 5 (accept), 5 (accept), 3 (border-line reject)).
Tutorial: Brain-Inspired AI 2.0: Aligning Language Models Across Languages and Modalities
has been accepted to AAAI 2026! (A*)
Received 'Great Review' Recognition from ACL Rolling Review (opens in new tab)
Review marked as 'Great Review' by Area Chairs in the ACL Rolling Review for its quality and constructive feedback.
Aligning Text/Speech Representations from Multimodal Models with MEG Brain Activity During Listening
has been accepted to EMNLP Conference 2025 (Main, Long, A*) (Nominated for the Paper Award)
Invited to serve as a reviewer for Nature Human Behavior. (opens in new tab)
I'm honored to join the reviewer pool for Nature Human Behavior, a top-tier journal in cognitive and behavioral sciences.
Appointed as Demos & Exhibits Co-Chair for AI-ML Systems 2025 (opens in new tab)
I'm excited to join the organizing committee of AI-ML Systems 2025 as the Demos & Exhibits Co-Chair.
New Preprint: Brain-Informed Fine-Tuning for Improved Multilingual Understanding in Language Models (opens in new tab)
Our latest work on leveraging brain data to improve multilingual capabilities in language models is now available on bioRxiv.
Two Extended Abstracts are Accepted as Posters at Cognitive Computational Neuroscience (CCN) 2025 (opens in new tab)
Both of our recent works — one on comparing variance partitioning and the residual method for interpreting brain recordings, and another on inducing bias from bilingual brains into language models — have been accepted as poster presentations at CCN 2025.
Instruction-Tuned Video-Audio Models Elucidate Functional Specialization in the Brain (opens in new tab)
Our new preprint explores how instruction-tuned multimodal models can provide insights into functional specialization in the brain.
USDC A Dataset of User Stance and Dogmatism in Long Conversations (opens in new tab)
has been accepted to ACL Findings 2025 (Findings, Long, A*)
Multi-modal brain encoding models for multi-modal stimuli (opens in new tab)
has been accepted to ICLR 2025! (Main Track, A*)
Correlating instruction-tuning (in multimodal models) with vision-language processing (in the brain) (opens in new tab)
has been accepted to ICLR 2025! (Main Track, A*)