Robot learns complex assembly tasks from simple success signals
AlphaNIST: Emergent Assembly Skills from Binary Task Rewards
Artificial IntelligenceRobotics
Summary
Assembly tasks involving small parts are hard for robots because they need precise movements and planning over many steps. The authors tested if a robot could figure out how to assemble parts using only a simple reward that says whether the task is done, without step-by-step hints. They trained a robot in simulation on 20 different assembly tasks and it got almost all of them right without any demonstrations or special instructions. The robot even learned to plan ahead and use clever strategies like reorienting parts and switching between tasks on its own.
What this means in practice
- •For industrial robotics engineers: Train robotic arms to assemble diverse parts without needing explicit instructions or demonstrations, reducing setup time for new products.
- •For robotics automation developers: Develop flexible assembly systems that discover efficient manipulation strategies from simple success feedback, allowing deployment across varied tasks.
Authors
Octi Zhang, Iretiayo Akinola, Rosario Scalise, Mateo Guaman Castro, Patrick Yin, Tobias Widmer, Miles Macklin, Chris Amevor, Dylan Turpin, Dirk Van Gelder, Jan Carius, Ankur Handa, Gavriel State, Yunrong Guo, Yashraj Narang
Abstract
The NIST Assembly Task Board #1 remains a challenging benchmark for learning-based methods in robotic manipulation. Over short horizons, it demands precise, contact-rich control of small, tight-fitting parts. Over long horizons, it requires recovery and coordination across extended action sequences. Demonstrations, task decomposition, and dense progress rewards have enabled research advances by providing guidance on how to assemble. But how much of the strategy must be specified? The progression from AlphaGo to AlphaGo Zero showed that RL at scale can discover superhuman strategies without prescribing how to win. AlphaNIST asks whether a sparse binary task reward and task diversity admitted by this task-agnostic reward can similarly let a robot discover how to assemble. In simulation, we train a single policy from scratch across all 20 assembly tasks on the NIST Assembly Task Board #1, reaching 98.2% overall training success without demonstrations, progress shaping, scripted primitives, or per-task tuning. The framework also trains policies to complete ten assembly pairs in a single episode without a prescribed order, with task success rewarded only when all ten pairs are simultaneously complete. The learned policies discover dexterous and planning-like behaviors: choosing grasps for later assembly, using surrounding geometry to reorient parts, and preparing one part, assembling another, then returning to finish the first without an explicit planner. For videos, see our website: https://alpha-nist.github.io/