4 papers
MHWA: Multi-timescale Hierarchical World-Action Model
Pengcheng Pan, Guoqing Ma, Yuhan Zhang +4
Autoregressive world models have emerged as a powerful paradigm for sample efficient decision making, yet their open loop rollouts often exhibit increasing drift that leads to erro…
EVA: Bridging Performance and Human Alignment in Hard-Attention Vision Models for Image Classification
Pengcheng Pan, Yonekura Shogo, Kuniyoshi Yasuo
Optimizing vision models purely for classification accuracy can impose an alignment tax, degrading human-like scanpaths and limiting interpretability. We introduce EVA, a neuroscie…
Debiasing Central Fixation Confounds Reveals a Peripheral "Sweet Spot" for Human-like Scanpaths in Hard-Attention Vision
Pengcheng Pan, Yonekura Shogo, Yasuo Kuniyosh
Human eye movements in visual recognition reflect a balance between foveal sampling and peripheral context. Task-driven hard-attention models for vision are often evaluated by how…
Emergence of Fixational and Saccadic Movements in a Multi-Level Recurrent Attention Model for Vision
Pengcheng Pan, Yonekura Shogo, Yasuo Kuniyoshi
Inspired by foveal vision, hard attention models promise interpretability and parameter economy. However, existing models like the Recurrent Model of Visual Attention (RAM) and Dee…