Feed-forward visual geometry models, exemplified by VGGT, have demonstrated strong multi-view reconstruction capabilities through dense global attention. However, the resulting all-to-all cross-view computation scales quadratically with the number of input views, making long-sequence inference particularly expensive du...
Shu-Jun Yao, Jun-Hui Lin, Cong-Ying Han et al.· 0 citations
A novel Preference-based Opponent Shaping (PBOS) method to enhance the strategy learning process by shaping agents’ preferences towards cooperation by incorporating the preference parameter into the agent’s loss function, thus allowing the agent to directly consider the opponent’s loss function when updating the strate...
Xinyu Qiao, Yudong Hu, Cong-Ying Han et al.· Machine-mediated learning· 1 citation
This paper establishes that every bounded sequence generated by Gi-PSADMM with the variance-reduced stochastic gradient estimator converges globally to a critical point in expectation and derives the convergence rate of Gi-PSADMM that depends on the KL exponent.
Jia Hu, Shi-Chen Liao, Cong-Ying Han et al.· Applied Mathematics and Opti...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.