Copertina

eOptShrinkQ: Near-Lossless KV Cache Compression Through Optimal Spectral Denoising and Quantization

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autore principale:	Su, Pei-Chun
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Information Theory
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Lossless and Near-Lossless Compression for Foundation Models
di: Hershcovitch, Moshik, et al.
Pubblicazione: (2024)

FibQuant: Universal Vector Quantization for Random-Access KV-Cache Compression
di: Lee, Namyoon, et al.
Pubblicazione: (2026)

RateQuant: Optimal Mixed-Precision KV Cache Quantization via Rate-Distortion Theory
di: Zuo, Fei, et al.
Pubblicazione: (2026)

SQuat: Subspace-orthogonal KV Cache Quantization
di: Wang, Hao, et al.
Pubblicazione: (2025)

TurboAngle: Near-Lossless KV Cache Compression via Uniform Angle Quantization
di: Patel, Dipkumar
Pubblicazione: (2026)

ZipNN: Lossless Compression for AI Models
di: Hershcovitch, Moshik, et al.
Pubblicazione: (2024)

Quad Length Codes for Lossless Compression of e4m3
di: Agrawal, Aditya, et al.
Pubblicazione: (2026)

Lossless Compression via Chained Lightweight Neural Predictors with Information Inheritance
di: Kim, Yuriy, et al.
Pubblicazione: (2026)

Polynomial Context-Truncation Sensitivity in Autoregressive Language Models: Sequential Wyner-Ziv Bounds for KV Cache Compression
di: Kim, Munsik
Pubblicazione: (2026)

Matrix Denoising with Doubly Heteroscedastic Noise: Fundamental Limits and Optimal Spectral Methods
di: Zhang, Yihan, et al.
Pubblicazione: (2024)

Bottlenecked Transformers: Periodic KV Cache Consolidation for Generalised Reasoning
di: Oomerjee, Adnan, et al.
Pubblicazione: (2025)

AlphaZip: Neural Network-Enhanced Lossless Text Compression
di: Narashiman, Swathi Shree, et al.
Pubblicazione: (2024)

Statistical Inference and Quality Measures of KV Cache Quantisations Inspired by TurboQuant
di: D'Alberto, Paolo
Pubblicazione: (2026)

Rethinking KV Cache Eviction via a Unified Information-Theoretic Objective
di: Yang, Jiaming, et al.
Pubblicazione: (2026)

Optimal and Near-Optimal Adaptive Vector Quantization
di: Ben-Basat, Ran, et al.
Pubblicazione: (2024)

StateSMix: Online Lossless Compression via Mamba State Space Models and Sparse N-gram Context Mixing
di: Tacconelli, Roberto
Pubblicazione: (2026)

GEAR: An Efficient KV Cache Compression Recipe for Near-Lossless Generative Inference of LLM
di: Kang, Hao, et al.
Pubblicazione: (2024)

Micro-Diffusion Compression - Binary Tree Tweedie Denoising for Online Probability Estimation
di: Tacconelli, Roberto
Pubblicazione: (2026)

Data-Driven Matrix Recovery via Optimal Shrinkage and Spatially Resolved Singular Vector Denoising under High-Dimensional Separable Noise
di: Su, Pei-Chun
Pubblicazione: (2025)

Lossless Image Compression Using Multi-level Dictionaries: Binary Images
di: Agnihotri, Samar, et al.
Pubblicazione: (2024)

KeepKV: Achieving Periodic Lossless KV Cache Compression for Efficient LLM Inference
di: Tian, Yuxuan, et al.
Pubblicazione: (2025)

Compressed Sensor Caching and Collaborative Sparse Data Recovery with Anchor Alignment
di: Yang, Yi-Jen, et al.
Pubblicazione: (2024)

Near-Optimal Clustering in Mixture of Markov Chains
di: Lee, Junghyun, et al.
Pubblicazione: (2025)

GPTQ-intrinsic LoRA: A Near-optimal Algorithm for Low-precision Quantization with Low-rank Adaptation
di: Zhang, Shihao, et al.
Pubblicazione: (2026)

OmniZip: Learning a Unified and Lightweight Lossless Compressor for Multi-Modal Data
di: Zhao, Yan, et al.
Pubblicazione: (2026)

Optimizing Learned Image Compression on Scalar and Entropy-Constraint Quantization
di: Borzechowski, Florian, et al.
Pubblicazione: (2025)

Sequential 1-bit Mean Estimation with Near-Optimal Sample Complexity
di: Lau, Ivan, et al.
Pubblicazione: (2025)

Hurwitz Quaternion Multiplicative Quantization for KV Cache Compression
di: Swain, Kabir, et al.
Pubblicazione: (2026)

KVTuner: Sensitivity-Aware Layer-Wise Mixed-Precision KV Cache Quantization for Efficient and Nearly Lossless LLM Inference
di: Li, Xing, et al.
Pubblicazione: (2025)

Almost Asymptotically Optimal Active Clustering Through Pairwise Observations
di: Teo, Rachel S. Y., et al.
Pubblicazione: (2026)

On the Regret of Coded Caching with Adversarial Requests
di: Nayak, Anupam, et al.
Pubblicazione: (2024)

QCM-SGM+: Improved Quantized Compressed Sensing With Score-Based Generative Models
di: Meng, Xiangming, et al.
Pubblicazione: (2023)

Optimal Quantization for Matrix Multiplication
di: Ordentlich, Or, et al.
Pubblicazione: (2024)

Orthogonal Approximate Message Passing with Optimal Spectral Initializations for Rectangular Spiked Matrix Models
di: Chen, Haohua, et al.
Pubblicazione: (2025)

VeriCache: Turning Lossy KV Cache into Lossless LLM Inference
di: Yao, Jiayi, et al.
Pubblicazione: (2026)

Quantization Dominates Rank Reduction for KV-Cache Compression
di: Salfati, Samuel
Pubblicazione: (2026)

Just Wing It: Near-Optimal Estimation of Missing Mass in a Markovian Sequence
di: Pananjady, Ashwin, et al.
Pubblicazione: (2024)

Beyond RAG: Task-Aware KV Cache Compression for Comprehensive Knowledge Reasoning
di: Corallo, Giulio, et al.
Pubblicazione: (2025)

Learnable Residual-Based Latent Denoising in Semantic Communication
di: Xu, Mingkai, et al.
Pubblicazione: (2025)

Spectral Estimators for Multi-Index Models: Precise Asymptotics and Optimal Weak Recovery
di: Kovačević, Filip, et al.
Pubblicazione: (2025)