Papers › CODE-CL: Conceptor-Based Gradient Projection for Deep Continual Learning

CODE-CL: Conceptor-Based Gradient Projection for Deep Continual Learning

21 Nov 2024arXiv:2411.15235archive 2025-07-28

Marco Paul E. Apolinario, Sakshi Choudhary, Kaushik Roy

Continual learning (CL) - the ability to progressively acquire and integrate new concepts - is essential to intelligent systems to adapt to dynamic environments. However, deep neural networks struggle with catastrophic forgetting (CF) when learning tasks sequentially, as training for new tasks often overwrites previously learned knowledge. To address this, recent approaches constrain updates to orthogonal subspaces using gradient projection, effectively preserving important gradient directions for previous tasks. While effective in reducing forgetting, these approaches inadvertently hinder forward knowledge transfer (FWT), particularly when tasks are highly correlated. In this work, we propose Conceptor-based gradient projection for Deep Continual Learning (CODE-CL), a novel method that leverages conceptor matrix representations, a form of regularized reconstruction, to adaptively handle highly correlated tasks. CODE-CL mitigates CF by projecting gradients onto pseudo-orthogonal subspaces of previous task feature spaces while simultaneously promoting FWT. It achieves this by learning a linear combination of shared basis directions, allowing efficient balance between stability and plasticity and transfer of knowledge between overlapping input feature representations. Extensive experiments on continual learning benchmarks validate CODE-CL's efficacy, demonstrating superior performance, reduced forgetting, and improved FWT as compared to state-of-the-art methods.

PaperPDFCode

Code

mapolinario94/CODE-CL officialmentioned on GitHubpytorchMIT report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Continual LearningImage ClassificationTransfer Learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Continual Learning 5-Datasets CODE-CL Average Accuracy 93.32 #1 of 1 Archive leaderboard report
Continual Learning 5-Datasets CODE-CL BWT -0.25 #1 of 1 Archive leaderboard report
Continual Learning Permuted MNIST CODE-CL Average Accuracy 96.56 #3 of 3 Archive leaderboard report
Continual Learning Permuted MNIST CODE-CL BWT -0.24 #3 of 3 Archive leaderboard report
Continual Learning miniImagenet CODE-CL Average Accuracy 68.83 #1 of 1 Archive leaderboard report
Continual Learning miniImagenet CODE-CL BWT -1.1 #1 of 1 Archive leaderboard report
Continual Learning split CIFAR-100 CODE-CL Average Accuracy 77.21 #1 of 2 Archive leaderboard report
Continual Learning split CIFAR-100 CODE-CL BWT -1.1 #1 of 2 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections