再帰的割り当てによる適応的可変長画像トークン化
Adaptive Length Image Tokenization via Recurrent Allocation
画像の情報量に応じて32〜256個のトークン数を再帰的に増減させる可変長トークナイザを提案し、再構成誤差とFIDで有効性を検証した。
著者: Shivam Duggal, Phillip Isola, Antonio Torralba, William T. Freeman
分類: cs.CV, cs.AI, cs.LG, cs.RO
原文アブストラクト
Current vision systems typically assign fixed-length representations to images, regardless of the information content. This contrasts with human intelligence - and even large language models - which allocate varying representational capacities based on entropy, context and familiarity. Inspired by this, we propose an approach to learn variable-length token representations for 2D images. Our encoder-decoder architecture recursively processes 2D image tokens, distilling them into 1D latent tokens over multiple iterations of recurrent rollouts. Each iteration refines the 2D tokens, updates the existing 1D latent tokens, and adaptively increases representational capacity by adding new tokens. This enables compression of images into a variable number of tokens, ranging from 32 to 256. We validate our tokenizer using reconstruction loss and FID metrics, demonstrating that token count aligns with image entropy, familiarity and downstream task requirements. Recurrent token processing with increasing representational capacity in each iteration shows signs of token specialization, revealing potential for object / part discovery.
