Imagen de Portada

Spectral Unforgetting: Post-Hoc Recovery of Damaged Capabilities Without Retraining

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Abro, Aarash, Tahir, Muhammad
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Post Hoc Regression Refinement via Pairwise Rankings
por: Wijaya, Kevin Tirta, et al.
Publicado: (2025)

Post-Hoc Reversal: Are We Selecting Models Prematurely?
por: Ranjan, Rishabh, et al.
Publicado: (2024)

Retrieval Augmented Anomaly Detection (RAAD): Nimble Model Adjustment Without Retraining
por: Pastoriza, Sam, et al.
Publicado: (2025)

Sustainable Machine Learning Retraining: Optimizing Energy Efficiency Without Compromising Accuracy
por: Poenaru-Olaru, Lorena, et al.
Publicado: (2025)

Agent-Based Post-Hoc Correction of Agricultural Yield Forecasts
por: Beddows, Matthew, et al.
Publicado: (2026)

Smooth InfoMax -- Towards Easier Post-Hoc Interpretability
por: Denoodt, Fabian, et al.
Publicado: (2024)

Informative Post-Hoc Explanations Only Exist for Simple Functions
por: Günther, Eric, et al.
Publicado: (2025)

CalArena: A Large-Scale Post-Hoc Calibration Benchmark
por: Berta, Eugène, et al.
Publicado: (2026)

Operationalizing Fairness: Post-Hoc Threshold Optimization Under Hard Resource Limits
por: Singh, Moirangthem Tiken, et al.
Publicado: (2026)

Principled Input-Output-Conditioned Post-Hoc Uncertainty Estimation for Regression Networks
por: Bramlage, Lennart, et al.
Publicado: (2025)

Are We Merely Justifying Results ex Post Facto? Quantifying Explanatory Inversion in Post-Hoc Model Explanations
por: Tan, Zhen, et al.
Publicado: (2025)

When Policies Cannot Be Retrained: A Unified Closed-Form View of Post-Training Steering in Offline Reinforcement Learning
por: Hossain, Elias, et al.
Publicado: (2026)

ShapeX: Shapelet-Driven Post Hoc Explanations for Time Series Classification Models
por: Huang, Bosong, et al.
Publicado: (2025)

Improving Policy Optimization via $\varepsilon$-Retrain
por: Marzari, Luca, et al.
Publicado: (2024)

On the Learn-to-Optimize Capabilities of Transformers in In-Context Sparse Recovery
por: Liu, Renpu, et al.
Publicado: (2024)

PERP: Rethinking the Prune-Retrain Paradigm in the Era of LLMs
por: Zimmer, Max, et al.
Publicado: (2023)

Comparing Post-Hoc Explainable AI Methods for Interpreting Black-Box EEG Models in Depression Detection
por: Šarčević, Antonia, et al.
Publicado: (2026)

TaylorPODA: A Taylor Expansion-Based Method to Improve Post-Hoc Attributions for Opaque Models
por: Tang, Yuchi, et al.
Publicado: (2025)

ChaosMining: A Benchmark to Evaluate Post-Hoc Local Attribution Methods in Low SNR Environments
por: Shi, Ge, et al.
Publicado: (2024)

Assessing Per-Sample Membership Inference Vulnerability without Retraining
por: Dorseuil, Valentin, et al.
Publicado: (2026)

Evaluating LLMs Capabilities Towards Understanding Social Dynamics
por: Tahir, Anique, et al.
Publicado: (2024)

Position: Rethinking Post-Hoc Search-Based Neural Approaches for Solving Large-Scale Traveling Salesman Problems
por: Xia, Yifan, et al.
Publicado: (2024)

SHapley Estimated Explanation (SHEP): A Fast Post-Hoc Attribution Method for Interpreting Intelligent Fault Diagnosis
por: Chen, Qian, et al.
Publicado: (2025)

Inner Loop Inference for Pretrained Transformers: Unlocking Latent Capabilities Without Training
por: Lys, Jonathan, et al.
Publicado: (2026)

A Free Lunch in LLM Compression: Revisiting Retraining after Pruning
por: Wagner, Moritz, et al.
Publicado: (2025)

xai_evals : A Framework for Evaluating Post-Hoc Local Explanation Methods
por: Seth, Pratinav, et al.
Publicado: (2025)

AI Scientists Fail Without Strong Implementation Capability
por: Zhu, Minjun, et al.
Publicado: (2025)

Is Retraining-Free Enough? The Necessity of Router Calibration for Efficient MoE Compression
por: Hyeon, Sieun, et al.
Publicado: (2026)

Estimating the Effects of Sample Training Orders for Large Language Models without Retraining
por: Yang, Hao, et al.
Publicado: (2025)

EvalxNLP: A Framework for Benchmarking Post-Hoc Explainability Methods on NLP Models
por: Dhaini, Mahdi, et al.
Publicado: (2025)

Safe Fairness Guarantees Without Demographics in Classification: Spectral Uncertainty Set Perspective
por: Barrainkua, Ainhize, et al.
Publicado: (2026)

Don't Retrain, Align: Adapting Autoregressive LMs to Diffusion LMs via Representation Alignment
por: Peng, Fred Zhangzhi, et al.
Publicado: (2026)

Post-Pruning Accuracy Recovery via Data-Free Knowledge Distillation
por: Tripurwar, Chinmay, et al.
Publicado: (2025)

Symmetry-Breaking Augmentations for Ad Hoc Teamwork
por: Hammond, Ravi, et al.
Publicado: (2024)

Pruning Foundation Models for High Accuracy without Retraining
por: Zhao, Pu, et al.
Publicado: (2024)

Multi-Level Safety Continual Projection for Fine-Tuned Large Language Models without Retraining
por: Han, Bing, et al.
Publicado: (2025)

Feature Attribution Stability Suite: How Stable Are Post-Hoc Attributions?
por: Subramaniakuppusamy, Kamalasankari, et al.
Publicado: (2026)

Post-Hoc Concept Disentanglement: From Correlated to Isolated Concept Representations
por: Erogullari, Eren, et al.
Publicado: (2025)

On Background Bias of Post-Hoc Concept Embeddings in Computer Vision DNNs
por: Schwalbe, Gesina, et al.
Publicado: (2025)

GPS-SSL: Guided Positive Sampling to Inject Prior Into Self-Supervised Learning
por: Feizi, Aarash, et al.
Publicado: (2024)