5 papers
Action-Consequence Alignment for Reliable Planning and Self-Improving in Latent World Models
Jinping Wang1, Zhiqiang Gao, Xiantong Zhen +1
Latent world models learn to predict observed transitions, yet low prediction error alone does not guarantee reliable planning. Inspired by self tickling experiments in neuroscienc…
Fix the Loss, Not the Radius: Rethinking the Adversarial Perturbation of Sharpness-Aware Minimization
Jinping Wang, Qinhan Liu, Zhiwu Xie +1
Sharpness-Aware Minimization (SAM) improves generalization by minimizing the worst-case loss within a fixed parameter-space radius neighborhood. SAM and its variants mainly rely on…
Rethinking Loss Reweighting for Imbalance Learning as an Inverse Problem: A Neural Collapse Point of View
Jinping Wang, Zixin Tong, Zhiwu Xie +1
Loss reweighting is a widely used strategy for long-tailed classification, but existing reweighting strategies often rely on heuristics and rarely define a well-specified target. I…
Space Alignment Matters: The Missing Piece for Inducing Neural Collapse in Long-Tailed Learning
Jinping Wang, Zhiqiang Gao, Zhiwu Xie
Recent studies on Neural Collapse (NC) reveal that, under class-balanced conditions, the class feature means and classifier weights spontaneously align into a simplex equiangular t…
Escaping Optimization Stagnation: Taking Steps Beyond Task Arithmetic via Difference Vectors
Jinping Wang, Zhiqiang Gao, Dinggen Zhang +1
Current methods for editing pre-trained models face significant challenges, primarily high computational costs and limited scalability. Task arithmetic has recently emerged as a pr…