today local_bar

Kanthashree Mysore Sathyendra

Researcher, Amazon

1 paper at NeurIPS 2025

OpenReview· Semantic Scholar· Google Scholar

Poster Session 3

Thursday, December 4, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

Compress, Gather, and Recompute: REFORMing Long-Context Processing in Transformers

#3511 · Woomin Song, Sai Muralidhar Jayanthi, Srikanth Ronanki, Kanthashree Mysore Sathyendra, Jinwoo Shin, Aram Galstyan, Shubham Katiyar, Sravan Babu Bodapati

We propose a compute- and memory-efficient inference framework for handling extremely long inputs with pre-trained Transformers.