We have 4 of 58 papers
We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.
Not the right person? Other researchers publish under this name.
Recent multimodal large language models (MLLMs) advance visual reasoning by strengthening both perception and reasoning, implicitly assuming a process that transitions seamlessly from perception to reasoning. However, we observe a counterintuitive phenomenon that challenges this assumption: holding the model, question,...
STD is proposed, a hierarchical token pruning framework that adapts token selection mechanisms to the functional role of each network stage, and introduces a Stability-Adaptive Trigger in deep layers to execute pruning only during semantically stable phases.
CoReFuse-Med is proposed, a Corruption-aware Rebalanced Fusion framework that suppresses corruption during feature transmission and rebalances modality contributions during high-level fusion, demonstrating improved accuracy and robustness under modality-quality discrepancies.
A compositional-learning-based method is proposed to learn two primitive sets (a common set and a discriminative set), which alleviates the regional shortcut by constraining the model to learn and utilize the common primitive set for base- and novel-class recognition.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
