Mamba-CL: Optimizing Selective State Space Model in Null Space for Continual Learning

TL;DR

Mamba-CL introduces a novel continual learning framework that fine-tunes the Mamba state space model by orthogonal parameter updates, effectively preventing forgetting and improving performance on class-incremental tasks.

Contribution

This work presents a new method for continual learning using null-space optimization of the Mamba model's parameters, ensuring task-specific consistency and reducing catastrophic forgetting.

Findings

Outperforms state-of-the-art continual learning methods on four benchmarks.

Effectively maintains task-specific knowledge through null-space parameter updates.

Demonstrates robustness and efficiency in class-incremental learning scenarios.

Abstract

Continual Learning (CL) aims to equip AI models with the ability to learn a sequence of tasks over time, without forgetting previously learned knowledge. Recently, State Space Models (SSMs), particularly the Mamba model, have achieved notable success in computer vision. Building on the strengths of SSMs, this study explores leveraging the Mamba model for CL. Therefore, we introduce Mamba-CL, a framework that continuously fine-tunes the core SSMs of the large-scale Mamba foundation model by updating parameters orthogonal to the feature subspace of previous tasks. This approach theoretically guarantees the consistency objective aiming to preserves consistent output for each SSM module across both previous and current tasks, so as to overcome catastrophic forgetting issue. Specifically, we achieve this goal by deducing the overall consistency constraints on four key time-invariant parameters in the Mamba model, streamlining its recurrent state-space structure and non-linear discretization process in SSM. In practice, we apply the null-space projection to efficiently implement the orthogonality within Mamba model. Extensive experiments on four class-incremental benchmarks demonstrate the effectiveness of Mamba-CL for anti-forgetting, achieving superior performances to state-of-the-art methods. Code is available in the supplementary materials.