Catastrophic Forgetting

Follow this topic Ask about this topic

Catastrophic forgetting occurs when learning new tasks or adapting to shifting data degrades performance on earlier tasks, a central challenge for systems expected to learn over time. Common approaches include replaying examples, distilling prior outputs, regularizing important weights or gradients, and isolating updates with adapters or modular experts. Researchers study these methods in sequential vision and language benchmarks as well as foundation-model fine-tuning, multimodal systems, and robotics, measuring both new-task performance and retention. No single approach works best everywhere: stronger protection can limit adaptation, while memory, compute, and benchmark choices affect results; apparent forgetting may also reflect knowledge becoming inaccessible rather than erased.

Latest papers

351 papers59 papers in the last four weeks, up 119% on the four weeks before. 0.3% of all new papers.

Jun 23, 2026

Curvature-Guided Mixing for MLLM Adaptation

Fine-tuning Multimodal Large Language Models (MLLMs) on specialized tasks often leads to catastrophic forgetting of their general capabilities. Existing model merging methods to combat this are often heuristic or use sub-optimal objectives. We propose CurvatureGuided Mixing (CGM), a theoretically grounded framework that merges pre-trained and fine-tuned models. CGM formulates a joint optimization objective and uses a second-order (Hessian) approximation of the loss landscapes to analytically derive an optimal, closed-form "soft mixing" ratio. This ratio intelligently blends parameters based on their relative task-specific curvatures. We also introduce CGM

\dagger

, a robust "hard mixing" variant that performs sparse parameter selection guided by a novel, curvature-aware score. Experiments on LLaVA-1.5 and Qwen2.5VL across multiple downstream tasks show that CGM and CGM

\dagger

consistently improve the trade-off between task specialization and general knowledge retention over existing methods. Code is available at github.com/zzsyjl/CGM-ECCV-2026.

Jinglong Yang, Jiaxuan He, Wenjian Huang +2

Research Institute of Trustworthy Autonomous Systems and Department of Computer Science and Engineering, Southern University of Science and Technology · Department of Computer Science, City University of Hong Kong

VLM Adaptation Multimodal Large Language Models Language Model Merging Model Merging Catastrophic Forgetting

Jun 23, 2026

Towards Continuous Power Forecasting: Practical Continual Learning for Real-World Energy Systems in Nonstationary Time Series

Power forecasting models deployed in real-world energy markets must operate under nonstationary conditions, where data distributions continually evolve due to weather variability, infrastructure upgrades, and changing consumption behaviors. In practice, these models face strict operational constraints: historical data may be limited or unavailable for repeated retraining, and uninterrupted long-term service is often required. This paper addresses these challenges by proposing the paradigm of Continuous Power Forecasting, which views power forecasting as a continual learning problem rather than a static offline task. Based on an adaptive continual learning framework for regression, we systematically investigate the practical effectiveness of six representative continual learning approaches from three methodological categories. These approaches are evaluated under different realistic assumptions regarding data accessibility and update policies. Experimental validation on real-world power datasets demonstrates that continual learning enables forecasting models to self-adapt to distributional drift, accumulate knowledge over time, and mitigate catastrophic forgetting without relying on large-scale historical data storage. Beyond performance gains, our study provides practical insights into the stability and adaptation behaviors of different continual learning approaches under realistic operational constraints. Overall, this work illustrates how continual learning can be pragmatically integrated into industrial power forecasting pipelines, offering a scalable and sustainable solution for long-term deployment in dynamic environments.

Yujiang He, Frederic Uhrweiller, Bernhard Sick

Intelligent Embedded Systems, University of Kassel, 34121 Kassel, Germany

Continual Learning Time Series Forecasting Non-Stationary Time Series Forecasting Catastrophic Forgetting

Jun 22, 2026

RECALL: Recovery Experience Collection for Active Lifelong Learning in Vision-Language-Action Models

Vision-Language-Action (VLA) models are commonly fine-tuned through passive imitation learning, where additional demonstrations are collected for tasks where the policy performs poorly. This approach incurs several downsides: it requires the robot to fail before data collection is triggered, provides little guidance about which states require supervision, and wastes demonstrator effort on redundant parts of the task where the policy already performs well. In this paper, we propose an active, continual learning paradigm for VLAs. We demonstrate that active, uncertainty-guided data collection leads to more efficient fine-tuning than when using passively-collected demonstrations. However, we also find that fine-tuning only on actively-collected recovery data leads to catastrophic forgetting. We evaluate techniques for continual learning, including replay-based data mixing and elastic weight consolidation, and identify tradeoffs between plasticity to uncertainty-guided recovery data and retention of previously learned behaviors. Overall, our work contributes an empirical study of active continual learning for autoregressive VLAs, establishing that uncertainty-guided recovery demonstrations can improve adaptation efficiency while also revealing open challenges when targeted new data is incorporated into large robot policies.

Ulas Berk Karli, Tesca Fitzgerald

Department of Computer Science Yale University United States

Continual Robot Learning Robot Skill Learning Elastic Weight Consolidation Imitation Learning Continual Learning for VLA Models Data-Efficient Robot Learning+3

Jun 20, 2026

SamatNext v0.2-B: An Exploratory Study of RMS-Normalized Hybrid Decoders for Curriculum Retention in Small Code Models

Standard autoregressive Transformer decoders can often exhibit substantial forgetting under sequential fine-tuning on shifting curriculum distributions. This technical report evaluates SamatNext v0.2-B, an experimental 356M-parameter hybrid sequence decoder that alternates Differential-Attention-style layers with DeltaNet-inspired simplified linear-state mixer layers using RMS normalization and output scale calibration. We study the model under a controlled staged Python code curriculum and compare it with a parameter-matched Transformer baseline. In this setting, SamatNext v0.2-B achieves a 100.0% pass rate on the controlled Stage 5 holdout while retaining 98.8% of adjacent Stage 3 semantic behavior and reaching 12.0% on the Stage 2E early syntax holdout. The strongest Transformer baseline reaches 97.6% on Stage 5 but retains only 6.0% of Stage 3 behavior. Both architectures remain weak on long-horizon early-stage retention, so the result should be interpreted as evidence of an altered retention/plasticity tradeoff in this controlled setting, not as a general solution to catastrophic forgetting. Code, model specifications, evaluation scripts, and result tables are provided for independent verification.

Samat Zharassov

Continual Learning for LLMs Decoder-Only Language Models Code Generation Catastrophic Forgetting

Jun 19, 2026

Task-Differentiated Atomic Skill Expansion and Routing for Continual Learning Across Highly Heterogeneous Tasks

Continual learning (CL) is commonly studied under the assumption that sequential tasks are semantically related or structurally similar. However, in highly heterogeneous settings, where tasks differ substantially in reasoning patterns and input-output formats, existing methods often suffer from catastrophic forgetting and inefficient capacity allocation. To address this challenge, we propose Task-differentiated Atomic Skill Expansion and Routing (TASER), a CL framework that jointly determines how many new atomic skills to introduce for each task and which skills to activate. The framework first uses atomic skill incremental learning to dynamically expand capacity based on task divergence and model uncertainty. It then applies orthogonality-enhanced skill detection to ensure these skills remain semantically distinct and independently reusable. Finally, a skill dynamic routing mechanism composes task-relevant skills through lightweight task-conditioned gating. We further introduce HeteroCLBench, a highly heterogeneous benchmark for CL, comprising 19 diverse tasks across 9 cognitive dimensions under a standardized sequential protocol. Experiments on HeteroCLBench show that TASER consistently outperforms strong baselines by improving plasticity and reducing catastrophic forgetting.

Jiacheng Wang, Xinjia He, Qi Ding +5

East China Normal University · Fudan University · Ocean University of China

Dynamic Neural Networks Continual Learning Catastrophic Forgetting

Jun 18, 2026

Sparsity, Superposition, and Forgetting: A Mechanistic Study of Representation Retention in Continual Learning

Continual learning (CL) systems often forget previously acquired knowledge, yet the mechanisms driving forgetting remain hard to isolate in practice because real datasets entangle many factors. We present a controlled, toy-world framework that makes these mechanisms observable and testable. Using a synthetic generator-separator pipeline, we define ground-truth latent features, build tasks with tunable sparsity and overlap, and introduce measurable quantities for representation strength and superposition (directional overlap among features). We then study retention dynamics-the temporal change of representation strength by fitting sparse dynamical relations (via SINDy) between retention, superposition, and exposure history. A complementary task-level analysis based on effective rank characterizes how representational capacity is allocated across tasks. Our controlled experiments yield three takeaways. (1) Superposition tends to increase over time with transient dips at task boundaries, suggesting boundary-specific interference rather than steady drift. (2) Higher feature sparsity induces more superposition yet does not inevitably cause forgetting; when representations remain strong, forgetting can be reduced despite overlap. (3) Task-level effective rank grows with sparsity, indicating broader capacity usage under sparse regimes. Together, these results nuance the common intuition that more superposition leads to more forgetting by showing that overlap interacts with representation strength and capacity allocation. Our toy analysis provides falsifiable hypotheses and diagnostic tools for CL.

Jan Wasilewski, Jędrzej Kozal, Michał Woźniak +1

Rochester Institute of Technology · Rochester NY, USA · Wrocław University of Science and Technology +1

Feature Superposition Representation Learning Continual Learning Catastrophic Forgetting

Jun 18, 2026

Beyond Uniform Forgetting: A Study of Sequential Direct Preference Optimization Across Preference Settings

Aligning language models with human preferences often requires optimising multiple behavioural objectives. A practical approach is to apply these objectives sequentially using preference optimisation methods such as Direct Preference Optimisation (DPO), but it remains unclear whether later training uniformly degrades preferences learned earlier or whether the effect depends on the relationship between objectives. We study sequential DPO across four preference settings covering distributional conflict, multi-attribute interaction, strong safety signal, and compatible response-quality objectives. Using Llama-3.1-8B-Instruct with LoRA adapters, we evaluate all objectives after every stage with a fixed base-model reference. We find that sequential DPO does not produce a single forgetting pattern; preference change ranges from partial degradation to stability, pair-level redistribution, or positive transfer depending on objective relationship, signal strength, and training order. Pair-level analysis using length-normalised policy margins shows that aggregate metrics can mask heterogeneous changes across preference pairs, whereas quartile decomposition reveals that high-confidence pairs can either degrade or improve depending on the setting. Mechanistic diagnostics show that Stage~2 gradients and adapter updates are near-orthogonal to the previous objective across all settings, providing little evidence that direct gradient opposition is the primary driver. These findings suggest that future sequential alignment pipelines should account for objective compatibility and signal strength, rather than assuming that later objectives affect earlier preferences uniformly.

Pranav Bhandari, Nicolas Fay, Amitava Datta +2

1Network Analysis and Social Influence Modelling (NASIM) Lab · School of Physics Maths and Computing, The University of Western Australia · School of Psychological Science, The University of Western Australia +1

Continual Learning for LLMs Direct Preference Optimization Preference Optimization Multi-Objective Language Model Alignment Catastrophic Forgetting

Jun 18, 2026

LOKI: Memory-Free Null-Space Constrained Lifelong Knowledge Editing

Lifelong knowledge editing aims to efficiently and sequentially update language models over time, as new knowledge becomes available or when the model makes mistakes, while preserving acceptable performance on past knowledge. One unresolved challenge is that existing methods modify a fixed set of layers for all new knowledge samples, reducing flexibility and increasing catastrophic forgetting. Another is requiring access to previous knowledge and extensive pre-processing to obtain data statistics. To address these challenges, we introduce LOKI, a novel approach that uses dynamic layer selection based on the Hilbert-Schmidt Independence Criterion and projects gradient updates onto the null-space of the model weights, bypassing the requirement for previous knowledge access. We show that LOKI achieves superior performance to existing approaches across a wide variety of experiments, achieving up to a 14% improvement in average accuracy.

Masih Eskandar, Miquel Sirera Perelló, Stratis Ioannidis +1

Department of Electrical and Computer Engineering Northeastern University Boston, MA, 02115

Knowledge Editing Continual Learning for LLMs Model Editing Language Model Editing Catastrophic Forgetting

Jun 16, 2026

Catastrophic Forgetting is Low-Rank: A Function-Space Theory for Continual Adaptation

Catastrophic forgetting in continual adaptation is usually studied through parameter drift, replay, or distillation, but these views do not identify which output-space directions are vulnerable. We give a function-space account in the NTK regime: new-task training induces old-task prediction drift through the cross-task kernel, yielding a closed-form predictor for the forgetting vector before any new-task gradient step. In frozen-backbone linear-head PEFT-CL, where the model is linear in the trainable parameters, the predictor is exact up to numerical precision; for nonlinear adapters/full fine-tuning, it is a local NTK approximation. The same expression reveals that forgetting concentrates in a small number of old-task NTK eigenmodes and under frozen linear heads gives a Kronecker scaling rule for the vulnerable rank. These results clarify the relation to prior NTK-overlap theory, explain why parameter-space regularizers can miss output-space interference, and motivate a targeted spectral regularizer.

Ido Nitzan Hidekel, Dan Raviv

Tel Aviv University, Tel Aviv, Israel.

Continual Learning Catastrophic Forgetting

Jun 13, 2026

When the Same Musical Knowledge Forgets Differently: A Clean Probe of Pathway-Dependent Forgetting

A model can learn that the piano piece Für Elise is calm and reflective by listening to the audio or by reading a text description, but does it matter which route that knowledge took when it is later at risk of being forgotten? Forgetting research in multimodal models measures what knowledge is lost under adaptation, yet has not asked whether acquisition route affects how easily that knowledge is forgotten. We call this untested premise the Pathway-Invariant Assumption. Music understanding enables a clean test because a music clip and a canonical text description can be aligned to the same perceptual content, allowing the same knowledge unit to enter a model through listening or reading while the target remains fixed. Across multiple architecturally distinct audio-language models, we observe a consistent asymmetry: text-pathway knowledge is forgotten more than matched audio-pathway knowledge under identical adaptation pressure. To attribute this effect to route rather than confounds, we introduce the Paired Pathway Controlled Protocol (PPCP), a three-phase design that establishes matched pathway baselines, activates both pathways under symmetric supervision on the same knowledge pool, and applies identical forgetting pressure to both pathways. The gap is stable across models and gain-controlled analyses, persists when contradictory overwrite is replaced by correct-label cross-domain learning, remains under single-modality pressure, and is not removed by lightweight replay. Two independent routing-depth controls confirm that the effect is not explained by architectural depth, pointing to input representation as the dominant factor. Under PPCP, our results demonstrate that forgetting is highly route-dependent, establishing acquisition route as a new analytical dimension for forgetting research and multimodal system design.

Yu Liu, Zhiwei Yang, Wenxiao Zhang +8

Institute of Information Engineering, CAS · School of Cyber Security, UCAS · The University of Western Australia +1

Music Understanding Audio-Language Models Multimodal Learning Catastrophic Forgetting

Jun 11, 2026

The Stable Recovery Manifold: Geometric Principles Governing Recoverability in Continual Learning

Catastrophic forgetting is often viewed as the destruction of previously learned knowledge during sequential learning. Building on the Accessibility Collapse framework, we investigate the geometric structure of recoverability in continual learning. Using Split CIFAR-100 and a sequentially trained ResNet-18, we analyze recoverability, representational drift, and recovery complexity across ten tasks. We introduce Recovery Subspace Dimensionality (k_t), a measure of the minimum number of singular directions required to preserve 90 percent of full probe performance. Contrary to our Recoverability Diffusion hypothesis, recovery dimensionality remains stable throughout training (mean k_t = 8.0) despite substantial representational drift. Principal-angle drift strongly predicts recoverability (r = -0.862), and a simple geometric model explains 82.2 percent of recoverability variance. These findings support the Stable Recovery Manifold hypothesis, suggesting that forgotten knowledge remains compactly decodable despite representational reorganization. The results indicate that catastrophic forgetting is primarily an accessibility and manifold-alignment problem rather than information destruction.

Ayushman Trivedi, Bhavika Melwani

Continual Learning Geometric Representation Learning Representation Drift Catastrophic Forgetting

Jun 10, 2026

ECA: Efficient Continual Alignment for Open-Ended Image-to-Text Generation

Incremental Learning (IL) for Open-ended Image-to-Text Generation (OpenITG) enables models to continuously generate accurate, contextually relevant text for new images while preserving previously acquired knowledge. Unlike prior studies, this paper addresses a more practical scenario in which the predominant category of visual data shifts over time as environments evolve. In this context, we introduce a new notion of continual alignment, which incrementally adapts the alignment module within pre-trained VLMs to preserve high-quality cross-modal representations. Based on this idea, we propose Efficient Continual Alignment (ECA), a novel exemplar-free IL approach for OpenITG. The key challenge is enabling the model to acquire new, task-specific features while minimizing interference with the established alignment without accessing raw data from previous tasks. To address this, ECA employs three core mechanisms: a Mixture of Query (MoQ) module that adapts task-specific query tokens, a Fisher Dynamic Expansion (FeDEx) that dynamically expands model structure based on a Fisher Information Matrix (FIM)-based metric, and an embedding dictionary with Dictionary Replay (DR) to retain past knowledge. To evaluate ECA's performance, we construct four new IL OpenITG benchmarks that better reflect real-world scenarios. Experimental results demonstrate that ECA significantly mitigates catastrophic forgetting and improves IL performance compared to baseline methods. Code and benchmarks are available at https://github.com/Snowball0823/ECA.

Jiangtao Kong, Peijun Zhao, Chun-Fu Chen +4

Department of Computer Science, William & Mary, Virginia, USA · Global Technology Applied Research, JPMorgan-Chase, USA · Mohamed bin Zayed University of Artificial Intelligence, Masdar City, Abu Dhabi

Large Vision-Language Models Incremental Learning Vision-Language Alignment Continual Learning for VLMs Catastrophic Forgetting

Jun 9, 2026

FOGO: Forgetting-aware Orthogonalization Optimizer

We argue that forgetting is not confined to continual learning but is a general optimization phenomenon: during standard training, dominant mini-batch gradients suppress rare but useful update directions, causing short-term forgetting at every step. When such knowledge is never revisited, these losses compound into long-term forgetting-the classical failure mode of continual learning. We introduce FOGO, a scalable optimizer that continuously detects and resolves gradient interference across both regimes. FOGO spectrally orthogonalizes momentum updates to prevent dominant directions from monopolizing optimization, then stores representative past directions in a compact codebook memory built on random projection, where pairwise distances are provably preserved in low-dimensional space. At each step, conflicts between the current update and stored directions are resolved via lightweight orthogonal correction and lifted back through a proximal step, with minimal overhead and no data storage. Across class-imbalanced classification, continual visual learning under domain and class shifts, continual fine-tuning of LLaVA-7B, and GPT-2 pretraining, FOGO consistently improves convergence and knowledge retention, outperforming Adam and Muon.

Toan Nguyen, Yang Liu, Trung Le +2

School of Computer Science and Engineering University of New South Wales Sydney, Australia · Department of Data Science & AI Monash University Melbourne, Australia · DEVCOM Army Research Laboratory USA

Deep Learning Optimization Continual Learning Catastrophic Forgetting

Jun 7, 2026

Theoretical Foundations of Continual Learning via Drift-Plus-Penalty

In many real-world settings, data streams are nonstationary and arrive sequentially, requiring learning systems to adapt continuously without retraining from scratch. Continual learning (CL) addresses this challenge by incorporating new tasks while mitigating catastrophic forgetting, where learning new information degrades performance on previously acquired knowledge. We introduce a control-theoretic perspective on CL that explicitly regulates the evolution of forgetting, framing adaptation as a controlled process subject to long-term stability constraints. We focus on replay-based CL, where a finite memory buffer stores representative samples from prior tasks. We propose COntinual Learning with Drift-Plus-Penalty (COLD), a continual learning framework based on the Drift-Plus-Penalty (DPP) principle from stochastic optimization. To facilitate analysis, we also consider an oracle variant, COLD-ORACLE, as a reference benchmark. At each task, both methods minimize the current task loss while maintaining a virtual queue that tracks deviations from long-term stability on previously learned tasks, capturing the stability-plasticity trade-off as a regulated dynamical process. We establish stability and convergence guarantees that characterize this trade-off through a tunable control parameter. Experiments on standard benchmarks demonstrate that COLD consistently outperforms a broad range of state-of-the-art CL methods while providing competitive and controllable forgetting behavior through explicit regulation of stability and plasticity.

Nazreen Shah, Govinda Arya, Bharath B. N. +1

IIIT Delhi · IIT Dharwad

Continual Learning Replay-Based Continual Learning Catastrophic Forgetting

Jun 7, 2026

Not Just After One: Sleep-Inspired Replay Prevents Catastrophic Forgetting After Sequential Tasks

One of the critical limitations of artificial neural networks is their lack of ability to continually learn: training on new tasks often leads to interference and forgetting of the previous ones. While several algorithms have been proposed to protect old memories from interference, they are typically applied during or immediately after each new episode of training. In contrast, humans and animals can learn continuously, acquiring multiple new memories during active learning before consolidating all of them into long-term storage. Here we show that multiple new tasks can be trained sequentially before an unsupervised sleep-like replay phase is applied to partially restore performance across all previously learned tasks. Our study further suggests that task-specific information remains resilient to new training but decays gradually as network is trained on new tasks. These findings point to novel principles for developing a broad range of continual learning AI solutions.

Anthony Bazhenov, Jean Erik Delanois, Giri P. Krishnan

Khoury College of Computer Sciences, Northeastern University, Boston, MA · Department of Medicine, University of California, San Diego, La Jolla, CA, USA · ARTISAN, Georgia Institute of Technology, Atlanta, GA USA

Replay-Based Continual Learning Memory Consolidation Catastrophic Forgetting

Jun 6, 2026

Evaluating the Impact of Task Granularity on Catastrophic Forgetting in Continual Learning

Catastrophic forgetting, the abrupt loss of previously acquired knowledge upon learning new information, remains the central challenge in Continual Learning. This project investigates whether the order in which a model learns information affects how well it retains knowledge. Specifically, we ask: does learning general categories first (like "animals" vs "vehicles") before learning specific classes (like "dog" vs "cat") reduce forgetting compared to learning all classes at once? We test three approaches on CIFAR-100: (1) Coarse-to-Fine: train on 2 super-classes, then expand to 10 specific sub-classes, (2) Fine-to-Coarse: train on 10 sub-classes, then group into 2 super-classes, and (3) Flat: train on all 10 classes from the start. We use Elastic Weight Consolidation (EWC) to prevent forgetting during transitions. Our hypothesis is that learning general patterns first creates a stable foundation that helps the model retain knowledge when learning more detailed distinctions. We evaluate using standard metrics (accuracy, precision, recall, F1) plus continual learning metrics like backward transfer and forgetting rates. This work could inform how we design learning sequences for real-world systems that need to learn incrementally.

Emre Alyamac, Himanshu Janmeda, Shashwat Krishna +1

College of Engineering · College of Natural Science

Class-Incremental Learning Continual Learning Catastrophic Forgetting

Jun 5, 2026

Sparse Subspace-to-Expert Sharing for Task-Agnostic Continual Learning

Continual learning in Large Language Models (LLMs) is hindered by the plasticity-stability dilemma, where acquiring new capabilities often leads to catastrophic forgetting of previous knowledge. Existing methods typically treat parameters uniformly, failing to distinguish between specific task knowledge and shared capabilities. We introduce Mixture of Sparse Experts for Task Agnostic Continual Learning (SETA), a framework that resolves the plasticity-stability conflict through adaptive sparse subspace decomposition into task-specific expert modules. Unlike standard updates, where tasks compete for the same parameters, SETA separates knowledge into unique experts, designed to isolate task-specific patterns, and shared experts, responsible for capturing common features. This structure is maintained through adaptive elastic anchoring and a routing-aware regularization that jointly protect shared knowledge at both the weight and routing levels and enable a unified gating network to automatically retrieve the correct expert combination during inference. Extensive experiments across diverse domain-specific benchmarks demonstrate that SETA achieves competitive or superior overall performance relative to state-of-the-art continual learning baselines, with particularly strong retention of early-task knowledge and improved backward transfer on LLaMA-2 7B and Qwen3-4B.

Fatema Siddika, Md Anwar Hossen, Tanwi Mallick +1

Iowa State University, Ames, USA · Argonne National Laboratory, USA

Continual Learning for LLMs Mixture-of-Experts Language Models Task-Free Continual Learning Catastrophic Forgetting

Jun 5, 2026

CoMetaPNS: Continually Meta-learning Personalized Neural Surrogates for Cardiac Electrophysiology Simulations

Personalized virtual heart simulations face challenges in model personalization and computational cost. While neural surrogates offer state-of-the-art solutions, they typically address either efficient personalization or training generalizable models. Recent work reframes this by learning the process of personalizing a surrogate using limited subject-specific context data, through few-shot generative modeling with set-conditioned surrogates and meta-learned amortized inference. These methods, however, assume a static and diverse training distribution with known task identifiers. When new data becomes available, they require costly retraining with all prior data to avoid catastrophic forgetting - a phenomena where the model forgets earlier tasks when trained on new ones. This is a major limitation in clinical settings where often unlabeled data arrives sequentially and full retraining is infeasible. This paper presents a new continual meta-learning framework to achieve personalized neural surrogates able to not only continually integrate information but also identify whether incoming data stems from a known or unknown dynamics source. By leveraging a continual Bayesian Gaussian Mixture Model over a memory buffer, our framework can infer the identifiers and relationships of data over time - required for effective meta-learning. Empirical results on synthetic cardiac data demonstrate superior simulation forecasting, computational scalability, and resilience to catastrophic forgetting compared to existing baselines.

Ryan Missel, Xiajun Jiang, Linwei Wang

Neural Surrogate Modeling Surrogate Modeling Meta-Learning Physics-Based Simulation Catastrophic Forgetting

Jun 5, 2026

Unsupervised Continual Clustering via Forward-Backward Knowledge Distillation

Unsupervised Continual Learning (UCL) aims to enable neural networks to learn sequential tasks without labels or access to past data. A major challenge in this setting is Catastrophic Forgetting, where models forget previously learned tasks upon learning new ones. This challenge is amplified in UCL due to the absence of labels to guide learning and memory retention. Existing mitigation strategies, such as knowledge distillation and replay buffers, often raise memory and privacy concerns. Moreover, current UCL methods largely overlook clustering-specific objectives. To fill this gap, we introduce Unsupervised Continual Clustering (UCC) and propose Forward-Backward Knowledge Distillation for Continual Clustering (FBCC). FBCC employs a continual teacher network with a clustering projector and lightweight task-specific students. Through a dual-phase forward-backward distillation process, the teacher learns new clusters while preserving previously discovered cluster structure without storing past data. FBCC represents a pioneering approach to UCC, demonstrating improved clustering performance across sequential tasks. Experiments on four benchmark datasets demonstrate that FBCC consistently outperforms existing continual learning baselines in clustering accuracy while significantly reducing catastrophic forgetting.

Mohammadreza Sadeghi, Sareh Soleimani, Zihan Wang +1

Department of Electrical and Computer Engineering, McGill University, Montreal, QC H3A 0E9, Canada · Mila – Quebec AI Institute, Montreal, QC H2S 3H1, Canada

Clustering Unsupervised Clustering Knowledge Distillation Catastrophic Forgetting

Jun 4, 2026

Catastrophic Forgetting as Accessibility Collapse: A Three-Level Framework for Knowledge Persistence in Continual Learning

Catastrophic forgetting is commonly interpreted as the irreversible erasure of previously acquired knowledge during sequential learning. In this work, we investigate an alternative perspective: that forgetting may arise not from complete destruction of task representations but from a loss of accessibility to preserved information. We introduce a three-level framework separating knowledge storage, representation, and accessibility, and evaluate each component through a series of continual-learning experiments on sequential CIFAR-100 classification using ResNet-18. Our analysis combines checkpoint persistence, linear probing, representation geometry, classifier-reset recovery, and layer-wise recoverability experiments. We observe complete behavioral forgetting of earlier tasks, with task accuracy collapsing from 54.8% to 0%, while linear probe performance retains approximately 76% of the original representational information. Furthermore, retraining only the final classifier restores 75.7% of the original task performance without modifying the backbone network. Layer-wise analysis reveals that early and intermediate layers preserve highly recoverable task information despite severe degradation at later stages. Projection-energy and principal-angle analyses indicate that retained knowledge persists as distributed high-dimensional representations rather than through preservation of a small dominant subspace. These findings suggest that catastrophic forgetting is better characterized as an accessibility failure than complete representational erasure, and that substantial task-relevant information remains embedded within neural representations even after functional forgetting has occurred.

Ayushman Trivedi, Bhavika Melwani

Independent Researchers

Continual Learning Catastrophic Forgetting

Jun 4, 2026

Two-Way Is Better Than One: Bidirectional Alignment with Cycle Consistency for Exemplar-Free Class-Incremental Learning

Continual learning (CL) seeks models that acquire new skills without erasing prior knowledge. In exemplar-free class-incremental learning (EFCIL), this challenge is amplified because past data cannot be stored, making representation drift for old classes particularly harmful. Prototype-based EFCIL is attractive for its efficiency, yet prototypes drift as the embedding space evolves; therefore, projection-based drift compensation has become a popular remedy. We show, however, that existing one-directional projections introduce systematic bias: they either retroactively distort the current feature geometry or align past classes only locally, leaving cycle inconsistencies that accumulate across tasks. We introduce BiCyc, a bidirectional projector alignment approach with a cycle-consistency objective. BiCyc jointly optimizes two maps, old-to-new and new-to-old, with stop-gradient gating so that transport and representation co-evolve. Analytically, we show that the cycle loss contracts the singular spectrum toward unity in whitened space, and that improved transport of class means and covariances yields smaller perturbations of classification log-odds, preserving old-class decisions and mitigating catastrophic forgetting. Empirically, across standard EFCIL benchmarks, BiCyc substantially reduces forgetting and improves accuracy in from-scratch settings, while remaining competitive in the pretrained fine-grained regime.

Hongye Xu, Bartosz Krawczyk

Chester F. Carlson Center for Imaging Science · Rochester Institute of Technology

Class-Incremental Learning Prototype-Based Classification Incremental Learning Cycle-Consistent Learning Consistency Regularization Catastrophic Forgetting

Jun 2, 2026

BBR-Net: Boundary-Balanced Replay for Continual Medical Image Segmentation

Continual learning for medical image segmentation remains challenging under domain shift because replay-based methods often preserve appearance information without explicitly modeling anatomical structure. This study investigates whether structural consistency governs knowledge retention in continual cardiac ultrasound segmentation. We propose the Boundary-Balanced Replay Network (BBR-Net), which selects replay samples using boundary-aware priority and class balance to preserve anatomically informative regions. The method is evaluated on CAMUS and CardiacNet under forward (CAMUS to CardiacNet) and reverse (CardiacNet to CAMUS) task orders. In the forward setting, BBR-Net retains source-task performance close to an offline joint-training reference, while markedly reducing catastrophic forgetting and preserving competitive target-task adaptation. Ablation results show that boundary-aware prioritization contributes to retention and improves the balance between source-task preservation and target-task adaptation when combined with class-aware sampling. In contrast, the reverse setting reveals that structure-aware replay fails when initial representations are learned from noisy and structurally inconsistent data. To isolate this effect, we conduct a controlled structural perturbation analysis by progressively corrupting source-task boundaries while keeping the dataset, architecture, and training protocol fixed. Forgetting increases consistently as structural reliability decreases, suggesting that replay effectiveness is strongly influenced by the quality of stored structural information, rather than by memory capacity alone. These findings indicate that preserving anatomical structure under domain shift is a central factor in continual medical image segmentation, and that replay mechanisms should account for structural reliability to support robust knowledge retention.

Zahid Ullah, Sieun Choi, Jihie Kim

Department of Computer Science and Artificial Intelligence, Dongguk University, Seoul, 04620, Republic of Korea

Interactive Medical Image Segmentation Ultrasound Image Segmentation Replay-Based Continual Learning Medical Image Segmentation Catastrophic Forgetting

Jun 1, 2026

Forgetting is Not Erasure: Recovering Latent Knowledge via Transport Keys

Catastrophic forgetting is often framed as a representational problem: after sequential training, a model appears to lose the features that supported performance on earlier tasks. We challenge the stronger form of this view. Across controlled continual-learning settings, we find that a significant portion of apparent forgetting can be attributed to interface drift between internal stages rather than permanent erasure of task-relevant computation. We study this phenomenon through a stitched evaluation protocol that combines early computation from a post-update network with late computation from its predecessor, optionally mediated by a compact, task-specific transport key. We describe transport keys at a systems level as compact interface-alignment operators estimated from a small set of paired anchor activations and evaluated through model stitching. On split CIFAR-100 with a ResNet-style network, transport keys recover most of the original Task A performance after sequential training on Task B. On a compact vision transformer, we observe a similar recovery pattern. These results suggest that continual learning may require better mechanisms for indexing and re-accessing latent computations, not only methods that prevent weight change.

Archie Chaudhury

Axionic Labs

Continual Learning Representation Alignment Catastrophic Forgetting

Jun 1, 2026

Continual Learning as a Multiphase Moving-Boundary Problem

Continual learning struggles to balance retaining past knowledge with absorbing new tasks. Stefan-CL elegantly resolves this stability-plasticity dilemma through the physics of melting. It frames consolidated knowledge as a protected "solid" and unused capacity as an adaptable "liquid." As the network learns, this boundary expands, governed by a "latent heat" tuning dial. By mathematically freezing the learned interior, Stefan-CL cuts forgetting to near zero, matching memory-heavy baselines without storing raw data, forging a beautiful, physics-grounded path for AI.

Snigdha Chandan Khilar

Independent Researcher

Continual Learning Catastrophic Forgetting

Jun 1, 2026

AlphaToken: Decoupling Adaptation and Stability for Path-Aware Response Token Valuation in LLM Post-Training

Token selection is pivotal for effective LLM post-training. However, existing methods mostly rely on local heuristics and rarely formulate token selection as a principled valuation of individual response tokens. We introduce

\textbf{AlphaToken}

, a response token valuation framework that decouples valuation into

\textbf{adaptation}

(promoting target-task learning) and

\textbf{stability}

(preserving pre-trained capabilities), and makes each objective

\textbf{path-aware}

by combining the direct-path signal from local token gradients with the downstream causal-path signal in autoregressive generation. Since retention data are typically unavailable, AlphaToken approximates stability via a

\textbf{Fisher-drift proxy}

anchored at the pre-trained reference model. For efficient computation, we extend Ghost Dot-Product to token-level valuation. AlphaToken masks low-value response tokens during fine-tuning and preference optimization, concentrating training signals on more valuable positions. Experiments show that AlphaToken improves post-training performance and mitigates catastrophic forgetting.

Liu Qing, Ou Wu, Yi Du

Hangzhou Institute for Advanced Study

LLM Fine-Tuning Language Model Post-Training Catastrophic Forgetting

May 31, 2026

Turning Back Without Forgetting: Selective Backward Refinement for Parameter-Efficient Continual Learning

While prompt-based parameter-efficient continual learning mitigates catastrophic forgetting by isolating task-specific prompts, this isolation also limits later tasks from improving earlier ones, leaving backward knowledge transfer underexplored. We address this limitation by proposing Selective bAckward refinement for positive Backward knowledge transfER (SABER), a replay-free framework that enables controlled backward transfer in prompt-based continual learning. SABER determines when backward refinement is beneficial using complementary task-correlation criteria based on prompt-gradient geometry and loss-distribution similarity, and how to perform refinement safely by restricting updates to non-interfering directions in the prompt parameter space. Extensive experiments across multiple continual learning benchmarks and diverse pretrained backbones, including T5-Large, LLaMA, and Qwen, demonstrate that SABER consistently achieves positive backward transfer while maintaining strong overall average performance. Code is available at https://github.com/OptMN-Lab/SABER-ICML-2026/.

Anushka Tiwari, Kaiyi Ji

Institute for Artificial Intelligence and Data Science, University at Buffalo, Buffalo, NY, USA · Department of Computer Science and Engineering, University at Buffalo, Buffalo, NY, USA

Storage-Efficient Continual Learning Parameter-Efficient Fine-Tuning Catastrophic Forgetting

May 30, 2026

Revisiting Parameter-Based Knowledge Editing in Large Language Models: Theoretical Limits and Empirical Evidence

Parameter-based knowledge editing updates the internal knowledge of large language models (LLMs) via localized weight modifications and has attracted significant attention. However, most existing methods overlook fundamental theoretical limitations and are rarely evaluated under realistic, practice-oriented settings. In this paper, we first present a theoretical analysis based on the dimensional Collapse Hypothesis, explaining how localized parameter edits can propagate along fragile directions in the representation space, inducing global interference and ultimately causing reasoning collapse. Building on this insight, we conduct a comprehensive empirical evaluation by systematically varying knowledge complexity, number of edits, evaluation dimensions, and baseline methods. Our results show that parameter-based editing methods consistently damage core LLM capabilities. In contrast, a simple retrieval-based baseline achieves consistently stronger performance than all parameter-editing methods across all evaluated conditions. These findings highlight that preserving the fundamental capabilities of LLMs after knowledge editing should be a central concern for future research.

Wanying Ren, Xin Song, Futing Wang +2

East China Normal University, Shanghai, China · Westlake University, Hangzhou, China · Nanyang Technological University, Singapore

Model Editing Language Model Editing Catastrophic Forgetting

May 29, 2026

Multi-Objective Reference-Aligned Machine Unlearning

Machine unlearning aims to remove the influence of specific training samples while preserving the model's utility. Existing single-objective approaches, such as gradient ascent or random relabeling, often induce catastrophic forgetting due to conflicting optimization dynamics and unbounded forgetting objectives that cause the model to drift from its pre-trained knowledge. We propose Reference-Aligned UnLearning (RAUL), a multi-objective framework that jointly optimizes forgetting and retention by replacing unbounded loss maximization with a bounded KL alignment of predictions on forgotten samples toward a reference distribution representing unseen data, instantiated either as a uniform distribution or an empirical distribution from a held-out reference set, which constrains the forgetting objective and reduces gradient conflict with retention. The resulting multi-objective optimization (MOO) problem is solved via Jacobian descent, which aggregates multiple gradients into a direction that does not conflict. Our results demonstrate that RAUL achieves the closest gap compared to full retraining.

Rasa Khosrowshahli, Stephen Asobiela, Beatrice Ombuki-Berman +1

Department of Computer Science, Brock University, St. Catharines, ON, Canada · Yousef Haj-Ahmad Department of Engineering, Brock University, St. Catharines, ON, Canada

Machine Unlearning Catastrophic Forgetting

May 29, 2026

CRMA: A Spectrally-Bounded Backbone for Modular Continual Fine-Tuning of LLMs

Sequential fine-tuning of large language models forces a choice: let the shared substrate keep learning and accept catastrophic forgetting, or freeze it after task one and foreclose cross-task refinement. Per-task adapter methods (LoRAHub, AdapterFusion, PackNet, Progressive Networks) take the second path. We introduce CRMA (Constrained Residual Mixing Adapter), a residual adapter whose internal mixing matrix M is doubly-stochastic at every forward pass via Sinkhorn normalization, so by Birkhoff's theorem ||M||_2 <= 1 holds by construction -- a structural bound, not a penalty. CRMA's spectrally bounded backbone provides a continuously trained shared substrate that earlier modular methods could not, while preserving their forgetting guarantees. On Mistral-7B across 5 sequential domains and 3 seeds, modular per-task LoRA on a CRMA backbone reduces loss-relative drift from +42.96% +/- 5.5 (naive sequential fine-tuning) to -0.17% +/- 0.17, with disjoint per-seed ranges, and improves prior-task holdout loss by 1.99% +/- 0.54 over a matched frozen-substrate baseline. Three independent experimental setups (Mistral-7B 4-domain controlled ablation, TinyLlama 3-domain contamination-controlled replication, Mistral-7B cross-domain probes at 7B) all show positive backward transfer -- without replay buffers, without growing per-task memory, and without distillation. An inference-time ablation on Gemma-2-9B confirms CRMA mediates access to sequentially trained knowledge: 98/100 vs. 38/100 on the same weights and same questions with only CRMA injection toggled. 867 logged training steps verify ||M||_2 = 1.0 within float32 precision (max deviation 1.2 x 10^-7). The forgetting-prevention effect holds across 1.1B-9.2B parameters and four architecture families.

Kiran Nayudu, Aswini Nutakki, Sai Vinay Naidu +1

ModelBrew AI

Continual Learning for LLMs Fine-Tuning Adapter Tuning Parameter-Efficient Fine-Tuning Catastrophic Forgetting

May 29, 2026

Parameter Alignment Mitigates Catastrophic Forgetting in Multilingual Expert Language Models

While continual pretraining~(CPT) is a practical way to extend large language models to new languages, naïve finetuning on targeted data erodes existing capabilities through catastrophic forgetting. Organizing training around language families reduces cross-language interference but cannot alone prevent forgetting of the general knowledge needed for downstream tasks. We link this forgetting to parameter drift in multilingual CPT and present a suite of five layer-aware parameter alignment strategies: hard layer freezing, soft regularization, post-hoc weight reversion, and model merging. We systematically compare our alignment strategies against two unregularized CPT baselines on benchmarks spanning 32 training languages from five language families, plus held-out languages, across four evaluation axes: perplexity, reading comprehension, physical reasoning, and translation. Parameter alignment substantially reduces forgetting at minimal cost to language acquisition: layer freezing and regularization best preserve comprehension, whereas post-hoc reversion yields the strongest translation gains. Together, these results map the acquisition--forgetting frontier for family-expert CPT and offer practical deployment guidelines pairing each strategy to the tasks it best serves.

Sanchit Ahuja, Terra Blevins

Northeastern University

Multilingual Language Models Continual Learning for LLMs LLM Alignment Catastrophic Forgetting