AI & ML
Grana: Compositional Reward Modeling for Long-Horizon Robot Manipulation
arXiv cs.RO0 views
arXiv:2611.24926v1 Announce Type: new Abstract: Robot reward modeling faces two challenges: generalizing across tasks and accurately tracking progress in long-horizon manipulation. General-purpose reward models demonstrate broad task transfer, but transfer from short-task pretraining to long-horizon tasks remains underexplored. Long-horizon progress models often rely on task-specific subtask classifiers or fixed action vocabularies, limiting the reuse of learned progress knowledge when adapting to new tasks. To address these challenges, we introduce Grana, a vision-language robot reward model that learns to identify unfinished subtasks and estimate their continuous progress using open-vocabulary subtask descriptions. We train a language-conditioned model to estimate progress in short-horizon tasks and individual subtasks of long-horizon tasks. For long-horizon tasks, we combine the selected subtask's predicted progress with credit for preceding completed subtasks to estimate whole-task progress. General progress prediction requires training the model across diverse tasks with progress supervision that accounts for imperfect execution. Elapsed-time proxies can misrepresent progress when relative subtask durations vary, or failures reverse local progress. We therefore introduce an annotation protocol that converts sparse subtask boundaries and action-quality annotations into dense progress targets, without requiring annotators to assign numerical progress scores. Pretraining with this supervision on 200 diverse, mostly short-horizon manipulation tasks enables zero-shot reward prediction on unseen tasks and improves progress prediction after adaptation to multi-minute, long-horizon tasks. Experiments demonstrate (i) generalizable reward prediction, including adaptation to five long-horizon tasks, (ii) benefits from pretraining and composing subtask progress estimates, and (iii) improved downstream policy performance.
Source:https://arxiv.org/abs/2611.24926PreviousRender the Robot, Generate the World: One Video World Model Across EmbodimentsNextPhoenix: Learning Scene Structure Representations for Generalizable World-Action Models
Related posts
P_SAPEA Exam Practice Questions: The Comparisons That Decide SAP Enterprise Architect Scenarios
0dev.to
Olala 7A1B says 8 GB, but the BF16 file is 13.59 GB: what a 1B-active hybrid MoE needs to run
0dev.to
A human liver has been kept alive outside the body for a record seven weeks
0MIT Technology Review

The Download: a living liver, a billion-dollar battery deal, and AI hype
0MIT Technology Review
Anthropic’s Record-Breaking IPO Highlights the AI Giant’s Worst Enemy—Itself
0Wired
21 Best Home Entertainment Gifts (2026): TVs, Lights, and More
0Wired