Constraining visual token budget per observation during VLM training forces genuine active perception and delivers 5% average relative improvement without auxiliary losses or architecture changes.
In: International Conference on Machine Learning (ICML) (2026), spotlight 21
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
-
Starve to Perceive: Taming Lazy Perception in VLMs with Constrained Visual Bandwidth
Constraining visual token budget per observation during VLM training forces genuine active perception and delivers 5% average relative improvement without auxiliary losses or architecture changes.