Copertina

Beyond Variance: Knowledge-Aware LLM Compression via Fisher-Aligned Subspace Diagnostics

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Shihab, Ibne Farabi, Akter, Sanjeda, Sharma, Anuj
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Anytime-Valid Answer Sufficiency Certificates for LLM Generation via Sequential Information Lift
di: Akter, Sanjeda, et al.
Pubblicazione: (2025)

CalPro: Prior-Aware Evidential--Conformal Prediction with Structure-Aware Guarantees for Protein Structures
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Differentiable Entropy Regularization: A Complexity-Aware Approach for Neural Optimization
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Minimax Optimality and Spectral Routing for Majority-Vote Ensembles under Markov Dependence
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Certificate-Guided Pruning for Stochastic Lipschitz Optimization
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

On the Sparsifiability of Correlation Clustering: Approximation Guarantees under Edge Sampling
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

What Fundamental Structure in Reward Functions Enables Efficient Sparse-Reward Learning?
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Beyond Accuracy: A Unified Random Matrix Theory Diagnostic Framework for Crash Classification Models
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Grassmannian Mixture-of-Experts: Concentration-Controlled Routing on Subspace Manifolds
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Selective Risk Certification for LLM Outputs via Information-Lift Statistics: PAC-Bayes, Robustness, and Skeleton Design
di: Akter, Sanjeda, et al.
Pubblicazione: (2025)

Detecting Proxy Gaming in RL and LLM Alignment via Evaluator Stress Tests
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Topology-Aware State Abstraction with Tangle Cores for Markov Decision Processes
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Continual Calibration: Coverage Can Collapse Before Accuracy in Lifelong LLM Fine-Tuning
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Cache-Efficient Posterior Sampling for Reinforcement Learning with LLM-Derived Priors Across Discrete and Continuous Domains
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Coverage-Based Calibration for Post-Training Quantization via Weighted Set Cover over Outlier Channels
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Auditing Near-Optimal Policies Can Be Exponentially Hard: Conditional Query Lower Bounds via Occupancy Rashomon Capacity
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Causal Consistency Regularization: Training Verifiably Sensitive Reasoning in Large Language Models
di: Akter, Sanjeda, et al.
Pubblicazione: (2025)

Why Colors Make Clustering Harder:Global Integrality Gaps, the Price of Fairness, and Color-Coupled Algorithms in Chromatic Correlation Clustering
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Calibrated Adaptation: Bayesian Stiefel Manifold Priors for Reliable Parameter-Efficient Fine-Tuning
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Learning to Forget Attention: Memory Consolidation for Adaptive Compute Reduction
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Enhancing Traffic Incident Response through Sub-Second Temporal Localization with HybridMamba
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Image Segmentation with Large Language Models: A Survey with Perspectives for Intelligent Transportation Systems
di: Akter, Sanjeda, et al.
Pubblicazione: (2025)

Large Language Models for Crash Detection in Video: A Survey of Methods, Datasets, and Challenges
di: Akter, Sanjeda, et al.
Pubblicazione: (2025)

HMAE: Self-Supervised Few-Shot Learning for Quantum Spin Systems
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Efficient Unstructured Pruning of Mamba State-Space Models for Resource-Constrained Environments
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

EST-PRM: Stress-Testing Process Reward Models Before They Become Load-Bearing
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

LLM-Guided Probabilistic Fusion for Label-Efficient Document Layout Analysis
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Adaptive Constraint Propagation: Scaling Structured Inference for Large Language Models via Meta-Reinforcement Learning
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Temporal Zoom Networks: Distance Regression and Continuous Depth for Efficient Action Localization
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

Accuracy is Not Agreement: Expert-Aligned Evaluation of Crash Narrative Classification Models
di: Bhagat, Sudesh Ramesh, et al.
Pubblicazione: (2025)

Unlocking Insights Addressing Alcohol Inference Mismatch through Database-Narrative Alignment
di: Bhagat, Sudesh, et al.
Pubblicazione: (2025)

Dynamic Proxy-Mixing: Transferring Replay Controllers from Small to Large Models for Continual Instruction Tuning
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Precise and Robust Sidewalk Detection: Leveraging Ensemble Learning to Surpass LLM Limitations in Urban Environments
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2024)

Heart rate and respiratory rate prediction from noisy real-world smartphone based on Deep Learning methods
di: Shihab, Ibne Farabi
Pubblicazione: (2025)

Calibrated and Resource-Aware Super-Resolution for Reliable Driver Behavior Analysis
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2025)

ClearVision: Leveraging CycleGAN and SigLIP-2 for Robust All-Weather Classification in Traffic Camera Imagery
di: Sivaraman, Anush Lakshman, et al.
Pubblicazione: (2025)

Canonicalized Stable-List Replay for Private Federated Continual Learning over Language-Model Embeddings
di: Shihab, Ibne Farabi, et al.
Pubblicazione: (2026)

Identification of Potentially Misclassified Crash Narratives using Machine Learning (ML) and Deep Learning (DL)
di: Bhagat, Sudesh, et al.
Pubblicazione: (2025)

Adaptive Variance-Penalized Continual Learning with Fisher Regularization
di: Sarkar, Krisanu
Pubblicazione: (2025)

Beyond Synthetic Benchmarks: Evaluating LLM Performance on Real-World Class-Level Code Generation
di: Rahman, Musfiqur, et al.
Pubblicazione: (2025)