HeLM: Hierarchical Explicit Language Memory for Non-Markovian Manipulation
HeLM gives robot policies an explicit language memory for counts, sequences, and cross-episode context, improving long-horizon manipulation while exposing key limits.
ls ./papers | grep "#vision-language-action models"
Trending AI research papers from Hugging Face, summarized in plain language: what each paper does, why it matters and where to read more.
8 papers · #vision-language-action models
HeLM gives robot policies an explicit language memory for counts, sequences, and cross-episode context, improving long-horizon manipulation while exposing key limits.
SLA adds spatial supervision to VLA continual pre-training, helping robots generalize across scenes, embodiments, perturbations, and limited downstream data.
H2RB helps robot-pretrained vision-language-action policies learn from unpaired human demonstrations by aligning hand geometry and sharing latent actions.
RoboColosseum tests robot policies across real-world sites, embodiments, and tasks, showing why partial progress, adaptation, and large-scale evaluation matter.
TinyAstral shows how evolving a robot agent’s harness before distillation can produce a compact 4B student that retains strong manipulation performance while using far less inference compute.
HiFOR links short-term motion, task milestones, and final goals in a VLA policy, then retrieves only the future information most relevant to each action.
ThinkWorld introduces Chain-of-Dynamics reasoning reinforced by world imagination, enabling robots to learn from human videos and generalize to unseen tasks and environments.
PDRVLA introduces compact physical tokens to vision-language-action models, improving robot manipulation by anticipating scene dynamics without costly video generation.