World-action models use predicted visual futures to condition robot actions, yet execution feedback can invalidate parts of a prediction while leaving its task structure useful. We propose Revisable Temporal Planning (RTP), which maintains the visual future as a persistent action condition and revises it after feedback...
Peng Liu, Jun-Bo Niu, Wen-Hao Zheng et al.· 0 citations
This work proposes TEVI, a framework that uses captions as a signal for what to retain from image embeddings and uses sparse autoencoders to disentangle image embeddings and train a masking module to selectively reconstruct the embedding based on a given caption.
S. Mahajan, Sukrut Rao, Jiahao Xie et al.· arXiv.org· 1 citation
GAS is presented, a generation-guided training framework that reinterprets visual generation as auxiliary supervision for representation learning and improves aggregate multimodal understanding, with its most reliable gains on perception and spatial comprehension.
Zhongbin Guo, Jiahao Xie, Dongling Xiao et al.· 1 citation
The DecoupleMix framework delivers two critical capabilities: guiding what data to collect next and rendering dataset validation a controlled, attributable experiment, and is competitive with strong open-source models trained with substantially larger multimodal budgets.
Jiahao Xie, Zhongbin Guo, Qianle Wang et al.· arXiv.org· 0 citations
This work builds an agent harness that separates reasoning from perception and improves by evolving context rather than optimizing weights, and proposes a gradient-free, reward-gated Heuristic Skill Distillation loop that mines the agent's own low-scoring traces and keeps a candidate skill only when it raises a validat...
Yingying Fan, Penghui Du, Leyan Zhu et al.· 1 citation
This work proposes PRISM, a four-stage data synthesis framework that produces persona--task pairs, prefix-guided rule sets, quality-filtered rubrics, and structured verification traces, and introduces PRISM-Eval, whose Loose and Strict metrics use deterministic matching against fixed labels and therefore require no inf...
Xiaomin He, Dongling Xiao, Jiahao Xie et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
