13 papers
Aligning Shared and Routed Experts for Cross-Subject EEG Generalization
Zhi Zhang, Yan Liu, Zhejing Hu +7
Cross-subject EEG generalization is challenging due to substantial heterogeneity across subjects. Existing methods typically learn either a shared subject-invariant model or multip…
Structure-Guided Memory Consolidation for Mitigating Compounding Errors in Literature Review Generation
Zhi Zhang, Yan Liu, Zhejing Hu +4
Compounding errors pose a significant challenge in automatic literature review generation, as inaccuracies can cascade across multi-stage retrieval and generation workflows. Existi…
Continuous Vision-Language-Action Co-Learning with Semantic-Physical Alignment for Behavioral Cloning
Xiuxiu Qi, Yu Yang, Jiannong Cao +4
Language-conditioned manipulation facilitates human-robot interaction via behavioral cloning (BC), which learns control policies from human demonstrations and serves as a cornersto…
Evolving and Executing Research Plans via Double-Loop Multi-Agent Collaboration
Zhi Zhang, Yan Liu, Zhejing Hu +3
Automating the end-to-end scientific research process poses a fundamental challenge: it requires both evolving high-level plans that are novel and sound, and executing these plans…
Who Decides How Knowing Becomes Doing? Redistributing Authority in Human-AI Music Co-Creation
Zhejing Hu, Yan Liu, Zhi Zhang +3
In the era of human-AI co-creation, the maxim "knowing is easy, doing is hard" is redefined. AI has the potential to ease execution, yet the essence of "hard" lies in who governs t…
MusicScaffold: Bridging Machine Efficiency and Human Growth in Adolescent Creative Education through Generative AI
Zhejing Hu, Yan Liu, Zhi Zhang +4
Adolescence is marked by strong creative impulses but limited strategies for structured expression, often leading to frustration or disengagement. While generative AI lowers techni…