Latent generative models typically follow a two-stage pipeline, training a variational autoencoder for reconstruction and then a generative model on the frozen latent space. Since reconstruction-optimized latents are not necessarily generation-friendly, jointly training both models is an appealing alternative. However,...
Guangting Zheng, Yiyuan Zhang, Tao Yang et al.· 2 citations· ⚡1
Adaptive Multi-Reward Projection is proposed, a lightweight online method that reallocates aggregation weights using three signals, relative shortfall, reward volatility, and recent progress, increasing pressure on lagging, unstable, or stagnant dimensions while relieving saturated ones.
RoMAN-Flow (Robotic Manipulation with Autoregressive Normalizing Flows), an offline reinforcement learning framework that makes AR-NF policies practical for robotic manipulation by addressing this sampling bottleneck in both stages.
Shao-Xuan Wang, Guangting Zheng, Rui Huang et al.· 0 citations
A role-aware signal that groups entities into agents, controlled objects, passive objects, and background, so that different physical roles can be modeled more clearly is introduced, and a modality-decoupled denoising strategy, where the visual and auxiliary channels are assigned independent noise levels, is proposed.
Guangting Zheng, Haojing Chen, Hao Li et al.· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
