Latent Action Models
PulseAugur coverage of Latent Action Models — every cluster mentioning Latent Action Models across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Research paper analyzes failures in Latent Action Models
This paper delves into the challenges of Latent Action Models (LAMs) in learning action representations from unlabeled videos. The research identifies that exogenous state, such as background clutter, interferes with re…
-
New ACT-LAM framework improves latent action modeling in videos
Researchers have introduced ACT-LAM, a novel framework designed to improve latent action modeling in videos. This approach addresses a core issue where lower reconstruction error doesn't always translate to better dynam…
-
Robot learning study clarifies impact of latent action models
Researchers have conducted a comprehensive empirical study on Latent Action Models (LAMs) for robot learning, aiming to clarify which design choices significantly impact downstream robotic manipulation performance. The …
-
FlexLAM introduces variable-length latent actions to improve video-based decision-making
Researchers have introduced FlexLAM, a novel approach to latent action learning that addresses the bottleneck trade-off in existing models. Unlike previous methods that use a fixed-capacity bottleneck, FlexLAM employs v…
-
New MaskLAM Method Enhances Embodied Agent Training
Researchers have developed a new method called MaskLAM to improve the training of embodied agents using latent action models. This technique addresses the issue of action-correlated visual distractors in videos, which c…
-
DiLA model advances self-supervised world models with disentangled learning
Researchers have developed DiLA, a novel Disentangled Latent Action world model designed to improve video generation and action abstraction. DiLA addresses the trade-off between action abstraction and generation fidelit…