PCGrad
PulseAugur coverage of PCGrad — every cluster mentioning PCGrad across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
Withdrawn research paper examines multi-task learning in materials science
A research paper exploring multi-task learning (MTL) in materials science has been withdrawn by its author, Sungwoo Kang. The study, which used a large dataset of metal alloys, found that MTL can degrade performance for…
-
Robotic agents gain steerable control via preference-conditioned reinforcement learning
Researchers have explored methods for steerable control in robotic agents, aiming to allow a single policy to adapt its behavior based on user preferences. They evaluated different aggregation levels for preference-cond…
-
New method improves training of physics-informed neural networks
Researchers have developed Norm-PCGrad, a novel method to improve the training of physics-informed neural networks (PINNs) and physics-informed Kolmogorov-Arnold Networks (PIKANs) when using domain decomposition. This t…
-
New Lantern model blends physics with diffusion for LHC calorimeter simulation
Researchers have developed Lantern, a novel physics-guided diffusion model designed to improve the simulation of calorimeter showers for the High-Luminosity LHC. Unlike previous methods, Lantern addresses the statistica…