Complexity Experts are Task-Discriminative Learners for Any Image Restoration
This work addresses inefficiencies in MoE-based image restoration models for researchers and practitioners, though it appears incremental as it builds on existing MoE frameworks with a novel architectural tweak.
The paper tackles the problem of inconsistent behavior in mixture-of-experts (MoE) architectures for all-in-one image restoration by introducing 'complexity experts' with varying computational complexity and receptive fields, resulting in a model that outperforms state-of-the-art methods while efficiently bypassing irrelevant experts during inference.
Recent advancements in all-in-one image restoration models have revolutionized the ability to address diverse degradations through a unified framework. However, parameters tied to specific tasks often remain inactive for other tasks, making mixture-of-experts (MoE) architectures a natural extension. Despite this, MoEs often show inconsistent behavior, with some experts unexpectedly generalizing across tasks while others struggle within their intended scope. This hinders leveraging MoEs' computational benefits by bypassing irrelevant experts during inference. We attribute this undesired behavior to the uniform and rigid architecture of traditional MoEs. To address this, we introduce ``complexity experts" -- flexible expert blocks with varying computational complexity and receptive fields. A key challenge is assigning tasks to each expert, as degradation complexity is unknown in advance. Thus, we execute tasks with a simple bias toward lower complexity. To our surprise, this preference effectively drives task-specific allocation, assigning tasks to experts with the appropriate complexity. Extensive experiments validate our approach, demonstrating the ability to bypass irrelevant experts during inference while maintaining superior performance. The proposed MoCE-IR model outperforms state-of-the-art methods, affirming its efficiency and practical applicability. The source code and models are publicly available at \href{https://eduardzamfir.github.io/moceir/}{\texttt{eduardzamfir.github.io/MoCE-IR/}}