CORAL: Scalable Multi-Task Robot Learning via LoRA Experts

TL;DR

CORAL introduces a scalable multi-task robot learning framework that uses a frozen backbone and lightweight task-specific experts with dynamic routing, effectively mitigating task interference and enabling continual learning in real-world robotics.

Contribution

The paper proposes CORAL, a novel framework that employs LoRA experts and a dynamic routing system to address multi-task interference and facilitate lifelong learning in robotic systems.

Findings

CORAL outperforms joint training in real-world and simulation benchmarks.

The framework prevents catastrophic forgetting during sequential task learning.

CORAL achieves zero inference overhead with expert swapping.

Abstract

Deploying Vision-Language-Action (VLA) models in real-world robotics exposes a core multi-task learning challenge: reconciling task interference in multi-task robotic learning. When multiple tasks are jointly fine-tuned in a single stage, gradients from different tasks can conflict, causing negative transfer and reducing per-task performance. Yet maintaining a separate full checkpoint per task is often storage- and deployment-prohibitive. To address this dilemma, we present CORAL, a backbone- and embodiment-agnostic framework designed primarily to mitigate multi-task interference while remaining naturally extensible to a continuous stream of new tasks. CORAL freezes a single pre-trained VLA backbone and attaches one lightweight Low-Rank Adaptation (LoRA) expert per task; at runtime, a dynamic inference engine (the CORAL Manager) routes language instructions to the appropriate expert and swaps experts on the fly with zero inference overhead. This strict parameter isolation avoids complex gating networks and prevents parameter-level cross-task interference by construction; as an added capability, it also enables sequentially introducing new tasks without parameter overwriting caused by catastrophic forgetting. We validate CORAL on a real-world Galaxea R1 dual-arm mobile manipulator and three simulation benchmarks (LIBERO, WidowX, Google Robot), where CORAL overcomes fine-grained instructional ambiguity and substantially outperforms joint training, yielding a practical and scalable system for lifelong multi-task robot learning. Website: https://frontierrobo.github.io/CORAL