OGBench
PulseAugur coverage of OGBench — every cluster mentioning OGBench across labs, papers, and developer communities, ranked by signal.
2 天有情绪数据
-
New CIG reward method enhances reinforcement learning exploration
Researchers have introduced Conditional Information Gain (CIG), a novel reward mechanism for reinforcement learning designed to improve exploration strategies. CIG addresses limitations of existing methods by providing …
-
新的流图策略加速机器人领域的生成式AI
研究人员开发了一类新的生成策略，称为流图策略，旨在加速复杂控制问题中的动作生成。这些策略学会了在生成动态中进行大跨步，与传统方法相比显著降低了推理成本。该方法，称为流图Q-引导（FMQ），优化了离线到在线强化学习的适应性，并在机器人操作和运动任务上展示了最先进的性能。
-
Refining Compositional Diffusion improves long-horizon planning by mitigating mode-averaging.
Researchers have developed Refining Compositional Diffusion (RCD), a new method to improve long-horizon trajectory planning for robots. RCD addresses the issue of mode-averaging in compositional diffusion planning, wher…
-
Gemma 4 31B weights show cross-modal transfer via thin trainable interface
Researchers have demonstrated that frozen weights from the Gemma 4 31B text-pretrained model can be effectively reused across different modalities, including robotics and associative recall tasks. By employing a thin, t…