Sen Li

Normalizing dual-encoder networks not only clarifies their interpretability but also reveals hidden structure in learned representations, challenging existing assumptions in the field.

Independent policy composition in multi-agent systems can lead to worse outcomes than any individual policy in the library, challenging conventional wisdom in reinforcement learning.

Aggregating rewards in the advantage while keeping likelihood ratios per-agent can significantly enhance cooperative multi-agent learning performance.

Achieving over 80% accuracy in cross-process welding penetration prediction could revolutionize intelligent monitoring across diverse welding systems.

Achieving 96.06% accuracy in laser welding penetration prediction with only 200 labeled images could revolutionize quality assurance in industrial applications.

WeldMamba achieves a remarkable 74.63% mIoU in predicting weld pool dynamics, setting a new benchmark for real-time welding applications.

Achieve state-of-the-art MARL performance by having agents reach a consensus in latent space before acting, effectively transforming the multi-agent problem into a hierarchical single-agent one.