Machine unlearning is a fundamental mechanism that enforces the right to be forgotten. Existing unlearning studies that rely on label manipulation or task-gradient reversal often deliver limited unlearning effectiveness. Moreover, they can undermine the original learning objective and typically do not guarantee equivalence to standard unlearning by retraining. In this paper, we propose \textbf{ManiF-SMC} (\textbf{Mani}fold \textbf{F}orgetting with \textbf{S}elf \textbf{M}ode \textbf{C}onnectivity), motivated by the observation that a model retrained on the remaining data tends to classify erased samples by their semantic similarity to the retained data. We begin with systematically recasting the approximate unlearning as pushing each erased sample away from its original learned manifold representation centroid toward its nearest semantic neighbors in the retained data. This reformulation aligns unlearning with retraining behavior and operates purely in representation space, reducing reliance on labels and task-specific gradients. To tackle the manifold representation-based unlearning problem, ManiF-SMC encapsulates the unlearning and representation preservation goals in a margin-based triplet loss. Because finding a suitable margin for unlearning is challenging, we propose a self-mode-connectivity module that rapidly reconstructs the local manifold to guide the adaptive margins generation for each unlearning case. Extensive experiments on four representative datasets show that ManiF-SMC achieves unlearning effectiveness comparable to state-of-the-art approximate methods while operating solely within the model's representation space.

Zero-Shot Neural Architecture Search with Weighted Response Correlation Kun Jing a,, Luoyu Chen b, Jungang Xu c,, Jianwei Tai a, Yiyu Wang d, Shuaimin Li e a School of Internet, Anhui University, Hefei, China b Alibaba Group Holding Limited, Hangzhou, China c School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing, China d Alibaba International Digital Commerce, Hangzhou, China e Shenzhen Key Laboratory for High Performance Data Mining, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, ChinaAbstract Neural architecture search (NAS) is a promising approach for automatically designing neural network architectures. However, the architecture estimation of NAS is computationally expensive and time-consuming because of training multiple architectures from scratch. Although existing zero-shot NAS methods use training-free proxies to accelerate the architecture estimation, their effectiveness, stability, and generality are still lacking. We present a novel training-free estimation proxy called weighted response correlation (WRCor). WRCor utilizes correlation coefficient matrices of responses across different input samples to calculate the proxy scores of estimated architectures, which can measure their expressivity and generalizability. Experimental results on proxy evaluation demonstrate that WRCor and its voting proxies are more efficient estimation strategies than existing proxies. We also apply them with different search strategies in architecture search. Experimental results on architecture search show that our zero-shot NAS algorithm outperforms most existing NAS algorithms in different search spaces. Our NAS algorithm can discover an architecture with a 22.1% test error on the ImageNet-1k dataset within 4 GPU hours. Introduction The success of deep learning in various fields [1], especially computer vision, causes a surge in demand for designing neural architectures. Designing neural architectures manually requires extensive expertise and time investment. Neural architecture search (NAS) [1, 2, 3, 4, 5, 6, 7] offers a potential solution for automatically designing neural architectures across various domains, eliminating the need for human involvement. The architecture estimation strategy is one of the core components of NAS.