In-Context Iterative Policy Improvement for Dynamic Manipulation
By: Mark Van der Merwe, Devesh Jha
Potential Business Impact:
Teaches robots to learn new tasks quickly.
Attention-based architectures trained on internet-scale language data have demonstrated state of the art reasoning ability for various language-based tasks, such as logic problems and textual reasoning. Additionally, these Large Language Models (LLMs) have exhibited the ability to perform few-shot prediction via in-context learning, in which input-output examples provided in the prompt are generalized to new inputs. This ability furthermore extends beyond standard language tasks, enabling few-shot learning for general patterns. In this work, we consider the application of in-context learning with pre-trained language models for dynamic manipulation. Dynamic manipulation introduces several crucial challenges, including increased dimensionality, complex dynamics, and partial observability. To address this, we take an iterative approach, and formulate our in-context learning problem to predict adjustments to a parametric policy based on previous interactions. We show across several tasks in simulation and on a physical robot that utilizing in-context learning outperforms alternative methods in the low data regime. Video summary of this work and experiments can be found https://youtu.be/2inxpdrq74U?si=dAdDYsUEr25nZvRn.
Similar Papers
RICL: Adding In-Context Adaptability to Pre-Trained Vision-Language-Action Models
Robotics
Teaches robots new tricks with few examples.
Leveraging In-Context Learning for Language Model Agents
Computation and Language
Helps AI agents learn by watching examples.
Robust Instant Policy: Leveraging Student's t-Regression Model for Robust In-context Imitation Learning of Robot Manipulation
Robotics
Helps robots learn tasks from watching humans.