VISTA: Validation-Informed Trajectory Adaptation via Self-Distillation

• URL: http://arxiv.org/abs/2604.12044v1
• Date: Mon, 13 Apr 2026 20:36:29 GMT
• Title: VISTA: Validation-Informed Trajectory Adaptation via Self-Distillation
• Abstract summary: We introduce VISTA, an online self-distillation framework that enforces consistency along the optimization trajectory.<n> VISTA identifies expert anchors, which are earlier model states that retain specialized competence over distinct data regions.<n>A coverage-weighted ensemble of these anchors is integrated online during training, regularizing the loss landscape and preserving mastered knowledge.
• Score: 7.6344611819427035
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Deep learning models may converge to suboptimal solutions despite strong validation accuracy, masking an optimization failure we term Trajectory Deviation. This is because as training proceeds, models can abandon high generalization states for specific data sub-populations, thus discarding previously learned latent features without triggering classical overfitting signals. To address this problem we introduce VISTA, an online self-distillation framework that enforces consistency along the optimization trajectory. Using a validation-informed Marginal Coverage score, VISTA identifies expert anchors, which are earlier model states that retain specialized competence over distinct data regions. A coverage-weighted ensemble of these anchors is integrated online during training, regularizing the loss landscape and preserving mastered knowledge. When evaluated across multiple benchmarks, VISTA demonstrates improved robustness and generalization over standard training and prior self-distillation methods, while a lightweight implementation reduces storage overhead by 90% without performance loss.

Related papers

• On-the-go Forgetting without Explicit Unlearning via ERASE [2.6426981014417046]
  ERASE is a framework for on-the-go forgetting that suppresses the observable influence of private data without modifying model weights.<n>We rigorously characterize sufficient conditions when ERASE provably achieves functional forgetting of designated subclasses.<n> ERASE maintains the best observed balance between forgetting efficacy, computational efficiency, and retention fidelity over recent unlearning-based methods.
  arXiv  Detail & Related papers  (2026-09-05T08:09:27Z)
• Dual-Teacher Distillation with Subnetwork Rectification for Black-Box Domain Adaptation [69.7036779439312]
  Black box domain adaptation represents a highly practical yet extremely challenging setting.<n>We propose a dual teacher distillation with subnetwork rectification (DDSR) model that exploits specific knowledge embedded in black box source models.<n>Our approach demonstrates consistent improvements over state of the art methods, including those using source data or models.
  arXiv  Detail & Related papers  (2026-03-24T07:54:19Z)
• Reference-Guided Machine Unlearning [16.1797488074794]
  We argue that unlearning should instead prioritize distributional indistinguishability, aligning the model's behavior on forget data with its behavior on truly unseen data.<n>Motivated by this, we propose Reference-Guided Unlearning (ReGUn), a framework that leverages a disjoint held-out dataset to provide a principled, class-reference for distillation.
  arXiv  Detail & Related papers  (2026-03-11T18:27:05Z)
• CapTrack: Multifaceted Evaluation of Forgetting in LLM Post-Training [48.70704477452434]
  We introduce textbfCapTrack, a capability-centric framework for analyzing forgetting in large language models.<n>We conduct a large-scale empirical study across post-training algorithms, domains, and model families.<n>We find that forgetting extends beyond parametric knowledge, with pronounced drift in robustness and default behaviors.
  arXiv  Detail & Related papers  (2026-02-19T09:46:24Z)
• Unleashing the Power of Vision-Language Models for Long-Tailed Multi-Label Visual Recognition [55.189113121465816]
  We propose a novel correlation adaptation prompt network (CAPNET) for long-tailed multi-label visual recognition.<n>CAPNET explicitly models correlations from CLIP's textual encoder.<n>It improves generalization through test-time ensembling and realigns visual-textual modalities.
  arXiv  Detail & Related papers  (2025-11-25T18:57:28Z)
• Towards Real Unsupervised Anomaly Detection Via Confident Meta-Learning [12.690752260618522]
  Confident Meta-learning (CoMet) is a novel training strategy that enables deep anomaly detection models to learn from uncurated datasets.<n>Our approach integrates Soft Confident Learning, which assigns lower weights to low-confidence samples, and Meta-Learning, which stabilizes training by regularizing updates.
  arXiv  Detail & Related papers  (2025-08-04T11:03:12Z)
• The Achilles Heel of AI: Fundamentals of Risk-Aware Training Data for High-Consequence Models [0.0]
  AI systems in high-consequence domains must detect rare, high-impact events while operating under tight resource constraints.<n>Traditional annotation strategies that prioritize label volume over informational value introduce redundancy and noise.<n>This paper introduces smart-sizing, a training data strategy that emphasizes label diversity, model-guided selection, and marginal utility-based stopping.
  arXiv  Detail & Related papers  (2025-05-20T22:57:35Z)
• Continuous Visual Autoregressive Generation via Score Maximization [69.67438563485887]
  We introduce a Continuous VAR framework that enables direct visual autoregressive generation without vector quantization.<n>Within this framework, all we need is to select a strictly proper score and set it as the training objective to optimize.
  arXiv  Detail & Related papers  (2025-05-12T17:58:14Z)
• Enhancing Robustness of Vision-Language Models through Orthogonality Learning and Self-Regularization [77.62516752323207]
  We introduce an orthogonal fine-tuning method for efficiently fine-tuning pretrained weights and enabling enhanced robustness and generalization. A self-regularization strategy is further exploited to maintain the stability in terms of zero-shot generalization of VLMs, dubbed OrthSR. For the first time, we revisit the CLIP and CoOp with our method to effectively improve the model on few-shot image classficiation scenario.
  arXiv  Detail & Related papers  (2024-07-11T10:35:53Z)
• Posterior Consistency for Missing Data in Variational Autoencoders [11.18081298867526]
  We consider the problem of learning Variational Autoencoders (VAEs) from data with missing values. We propose an approach for regularizing an encoder's posterior distribution which promotes this consistency. This improved performance can be observed for many classes of VAEs including VAEs equipped with normalizing flows.
  arXiv  Detail & Related papers  (2023-10-25T13:56:02Z)
• Self-Supervised Training with Autoencoders for Visual Anomaly Detection [61.62861063776813]
  We focus on a specific use case in anomaly detection where the distribution of normal samples is supported by a lower-dimensional manifold. We adapt a self-supervised learning regime that exploits discriminative information during training but focuses on the submanifold of normal examples. We achieve a new state-of-the-art result on the MVTec AD dataset -- a challenging benchmark for visual anomaly detection in the manufacturing domain.
  arXiv  Detail & Related papers  (2022-06-23T14:16:30Z)
• Unsupervised Domain Adaptation for Speech Recognition via Uncertainty Driven Self-Training [55.824641135682725]
  Domain adaptation experiments using WSJ as a source domain and TED-LIUM 3 as well as SWITCHBOARD show that up to 80% of the performance of a system trained on ground-truth data can be recovered.
  arXiv  Detail & Related papers  (2020-11-26T18:51:26Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.