Distillation aligns compression mechanisms between full-history and recurrent transformers, enabling linear-time recurrent memory that narrows the performance gap for streaming vision and robotics tasks.
S- MUSt3R: Sliding multi-view 3d reconstruction.arXiv preprint arXiv:2602.04517, 2026
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Compressing Observation History into Agent Memory: Distilling Transformers into Recurrent Transformers
Distillation aligns compression mechanisms between full-history and recurrent transformers, enabling linear-time recurrent memory that narrows the performance gap for streaming vision and robotics tasks.