{"total":19,"items":[{"citing_arxiv_id":"2606.00382","ref_index":35,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"CRMA: A Spectrally-Bounded Backbone for Modular Continual Fine-Tuning of LLMs","primary_cat":"cs.LG","submitted_at":"2026-05-29T21:50:23+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"CRMA adds a spectrally bounded residual adapter backbone to modular continual fine-tuning of LLMs, achieving near-zero loss drift and positive backward transfer on Mistral-7B across domains.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.31108","ref_index":34,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Remembering by Reconstructing: Domain Incremental Learning With Test-Time Training on Video Streams","primary_cat":"cs.CV","submitted_at":"2026-05-29T10:17:09+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Domain-incremental video learning that permits forgetting through per-domain LoRA adapters and recovers the matching adapter at inference via test-time training on a self-supervised MAE reconstruction head.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.11617","ref_index":62,"ref_count":2,"confidence":0.98,"is_internal_anchor":true,"paper_title":"MIST: Reliable Streaming Decision Trees for Online Class-Incremental Learning via McDiarmid Bound","primary_cat":"cs.LG","submitted_at":"2026-05-12T06:45:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"MIST fixes unreliable splits in streaming decision trees for class-incremental learning by replacing Hoeffding-style bounds with a K-independent McDiarmid radius on Gini, plus Bayesian parent-to-child inheritance and per-leaf quantile sketches.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Decomposed Attention-Based Prompting for Rehearsal-Free Continual Learning . In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11909- 11919, Los Alamitos, CA, USA, June 2023. IEEE Computer Society. [61] Gido van de Ven, Tinne Tuytelaars, and Andreas Tolias. Three types of incremental learning. Nature Machine Intelligence, 4:1-13, 12 2022. [62] Gido M. van de Ven and Andreas S. Tolias. Three scenarios for continual learning.CoRR, abs/1904.07734, 2019. [63] Yuandou Wang, Filip Gunnarsson, and Rihan Hai. An attention-based feature memory design for energy-efficient continual learning, 2025. 14 [64] Zifeng Wang, Zizhao Zhang, Sayna Ebrahimi, Ruoxi Sun, Han Zhang, Chen-Yu Lee, Xiaoqi Ren, Guolong Su, Vincent Perot, Jennifer Dy, and Tomas Pfister."},{"citing_arxiv_id":"2605.10529","ref_index":34,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"PrimeKG-CL: A Continual Graph Learning Benchmark on Evolving Biomedical Knowledge Graphs","primary_cat":"cs.AI","submitted_at":"2026-05-11T13:14:02+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"PrimeKG-CL supplies the first continual graph learning benchmark using authentic temporal snapshots from nine biomedical databases, showing strong interactions between embedding decoders and learning strategies plus limits of standard metrics on retention versus forgetting.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"32 when first learned (task 2) and degrades to ≈0.09 by task 10, a 3.4× collapse despite EWC and replay mitigations. mirrors how upstream databases push updates (GO functional annotations, HPO phenotype links, DrugBank interactions), creating a dual challenge oftask identity shiftsandtemporal adaptation (t0 →t 1 within each task), analogous to domain-incremental CL in vision [34]. 3.4 Evaluation Tasks PrimeKG-CL supports three evaluation tracks, chosen so that a continual-learning failure on each one corresponds to a concrete downstream cost: a missed drug-repurposing candidate (link prediction), a stale answer to a clinician's natural-language query (KGQA), or a newly catalogued entity assigned to the wrong type (node classification)."},{"citing_arxiv_id":"2605.07886","ref_index":7,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Characterizing and Correcting Effective Target Shift in Online Learning","primary_cat":"stat.ML","submitted_at":"2026-05-08T15:34:10+00:00","verdict":"UNVERDICTED","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Online kernel regression equals offline regression with shifted targets; correcting the targets lets online learning match offline performance and outperform true targets in continual image classification.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"[5] Cecilia S Lee and Aaron Y Lee. Clinical applications of continual learning machine learning.The Lancet Digital Health, 2(6):e279-e281, 2020. [6] Michael McCloskey and Neal J Cohen. Catastrophic interference in connectionist networks: The sequential learning problem. InPsychology of learning and motivation, volume 24, pages 109-165. Elsevier, 1989. [7] Gido M Van de Ven and Andreas S Tolias. Three scenarios for continual learning.arXiv preprint arXiv:1904.07734, 2019. [8] Raia Hadsell, Dushyant Rao, Andrei A Rusu, and Razvan Pascanu. Embracing change: Continual learning in deep neural networks.Trends in cognitive sciences, 24(12):1028-1040, 2020. [9] Robert M French. Using semi-distributed representations to overcome catastrophic forgetting in connec-"},{"citing_arxiv_id":"2605.05776","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning","primary_cat":"cs.AI","submitted_at":"2026-05-07T07:09:03+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"In machine learning, we analogize this to describe a sam- ple's classification probability. Given a model's output sim- ilarity H(x) for sample x, where H(x)[y] is the similarity for class y, the classification probability P(y|x) can be cal- culated via softmax, akin to Helmholtz free energy. This transforms into an energy formula: P(y|x) = eH(x)[y]/kT PU y′=1 eH(x)[y ′]/kT = e−E(x,y)/kT e−E(x)/kT ,(3) the numerator reflects the energy variation of sample x for class y, while the denominator represents the variation across all possible classes {yi}U i=1. The Helmholtz free energyE(x)for a sample is then defined as: E(x) =−kT·ln \" UX y=1 eH(x)[y]/kT # ,(4) which is termed energy in subsequent sections. 4. Methodology 4.1. Overall Architecture Fig."},{"citing_arxiv_id":"2604.27031","ref_index":22,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"NORACL: Neurogenesis for Oracle-free Resource-Adaptive Continual Learning","primary_cat":"cs.LG","submitted_at":"2026-04-29T15:04:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"NORACL dynamically grows network capacity via neurogenesis-inspired signals to achieve oracle-level continual learning performance without pre-specifying architecture size.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.24637","ref_index":39,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Cortex-Inspired Continual Learning: Unsupervised Instantiation and Recovery of Functional Task Networks","primary_cat":"cs.LG","submitted_at":"2026-04-27T16:06:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"FTN achieves near-zero forgetting on continual learning benchmarks by isolating task subnetworks via self-organizing binary masks generated through gradient descent, smoothing, and k-winner-take-all.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.21927","ref_index":18,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Fine-Tuning Regimes Define Distinct Continual Learning Problems","primary_cat":"cs.LG","submitted_at":"2026-04-23T17:59:34+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"The relative rankings of continual learning methods are not preserved across different fine-tuning regimes defined by trainable parameter depth.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.18270","ref_index":25,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Incremental learning for audio classification with Hebbian Deep Neural Networks","primary_cat":"eess.AS","submitted_at":"2026-04-20T13:45:31+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A kernel plasticity approach in Hebbian DNNs for incremental sound classification achieves 76.3% accuracy over five steps on ESC-50, outperforming the 68.7% baseline without plasticity.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.14259","ref_index":41,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Continual Learning for fMRI-Based Brain Disorder Diagnosis via Functional Connectivity Matrices Generative Replay","primary_cat":"q-bio.TO","submitted_at":"2026-04-15T16:08:52+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A structure-aware VAE generates realistic FC matrices for replay, combined with multi-level knowledge distillation and hierarchical contextual bandit sampling, to enable continual fMRI-based brain disorder diagnosis across sequentially arriving multi-site data without catastrophic forgetting.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2603.20410","ref_index":27,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"SLE-FNO: Single-Layer Extensions for Task-Agnostic Continual Learning in Fourier Neural Operators","primary_cat":"cs.LG","submitted_at":"2026-03-20T18:30:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"SLE-FNO achieves zero forgetting and strong plasticity-stability balance in continual learning for FNO surrogate models of pulsatile blood flow by adding minimal single-layer extensions across four out-of-distribution tasks.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"A continual learning survey: Defying forgetting in classification tasks.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3366-3385, 2021. [26] L. Wang, X. Zhang, H. Su, and J. Zhu. A comprehensive survey of continual learning: Theory, method and application.IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(8):5362-5383, 2024. [27] G. M. Van de Ven and A. S. Tolias. Three scenarios for continual learning.arXiv preprint arXiv:1904.07734, 2019. [28] Z. Li and D. Hoiem. Learning without forgetting.IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(12):2935-2947, 2017. [29] J. Kirkpatrick, R. Pascanu, N. Rabinowitz, J. Veness, G. Desjardins, A. A. Rusu, K. Milan,"},{"citing_arxiv_id":"2602.07940","ref_index":6,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"MePo: Meta Post-Refinement for Rehearsal-Free General Continual Learning","primary_cat":"cs.AI","submitted_at":"2026-02-08T12:15:35+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MePo refines pretrained backbones via meta-learning on constructed pseudo tasks and initializes a meta covariance matrix to enable robust second-order alignment, yielding 12-15% gains on CIFAR-100, ImageNet-R and CUB-200 in rehearsal-free GCL settings.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2601.16175","ref_index":74,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Learning to Discover at Test Time","primary_cat":"cs.LG","submitted_at":"2026-01-22T18:24:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"TTT-Discover applies test-time RL to set new state-of-the-art results on math inequalities, GPU kernels, algorithm contests, and single-cell denoising using an open model and public code.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2508.03941","ref_index":12,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Measuring the stability and plasticity of recommender systems","primary_cat":"cs.IR","submitted_at":"2025-08-05T22:15:43+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A new offline protocol to profile recommender algorithms by stability in retaining past patterns and plasticity in adapting to changes upon retraining, with preliminary results on the GoodReads dataset.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2505.18604","ref_index":55,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Exemplar-Free Continual Learning for State Space Models","primary_cat":"cs.LG","submitted_at":"2025-05-24T08:59:13+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Inf-SSM constrains the infinite-horizon evolution of SSMs via Grassmannian geometry and an efficient O(n^2) Sylvester solver to enable exemplar-free continual learning with reduced forgetting.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2305.17493","ref_index":10,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"The Curse of Recursion: Training on Generated Data Makes Models Forget","primary_cat":"cs.LG","submitted_at":"2023-05-27T15:10:41+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Use of model-generated content in training causes irreversible loss of distribution tails, termed model collapse, in VAEs, GMMs, and LLMs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2305.01507","ref_index":20,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"A Parameter-free Adaptive Resonance Theory-based Topological Clustering Algorithm Capable of Continual Learning","primary_cat":"cs.NE","submitted_at":"2023-05-01T01:04:07+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"An ART-based topological clustering algorithm estimates vigilance and edge deletion thresholds automatically using determinantal point process and edge age for parameter-free continual learning.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2209.14774","ref_index":6,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"RECALL: Rehearsal-free Continual Learning for Object Classification","primary_cat":"cs.CV","submitted_at":"2022-09-29T13:36:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"RECALL achieves rehearsal-free continual learning for object classification by logit recall before new training, regression regularization, Mahalanobis loss on known categories, and new heads per sequence, outperforming prior methods on CORe50, iCIFAR-100, and the introduced HOWS-CL-25 dataset.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}