Introduces Image Reconstruction Game benchmark showing describer model dominates reconstruction quality in multi-turn VLM-generator dialogue, with math images hardest and token budget affecting convergence.
Devon Hjelm, Layla El Asri, Samira Ebrahimi Kahou, Yoshua Bengio, and Graham W
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
-
The Image Reconstruction Game: Drawing Common Ground Through Iterative Multimodal Dialogue
Introduces Image Reconstruction Game benchmark showing describer model dominates reconstruction quality in multi-turn VLM-generator dialogue, with math images hardest and token budget affecting convergence.