Embedding Compression

Ask about this topic

Overview

Embedding compression aims to make vector representations cheaper to store, transmit, and use without losing the information needed for downstream tasks. Methods include dimensionality reduction such as PCA, scalar and product quantization, binary hashing, pruning or pooling multi-vector representations, and learned codebooks or autoencoders. In retrieval, systems are often evaluated on benchmarks such as BEIR and MTEB using ranking quality alongside index size and latency; recommendation and on-device applications also measure parameter and compute costs. Very compact representations can work well, but the best method and acceptable tradeoff depend on the task, embedding geometry, and whether compression preserves reconstruction, similarity, or ranking quality.

Latest papers

CardsList

49 papers10 papers in the last four weeks, up 43% on the four weeks before. 0.1% of all new papers.

  1. Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent Vector Quantization

    Oct 22, 2026Jiacheng Liu, Yaxin Luo, Xiaohan Zhao +5Embedding CompressionLLM Compression

    Mohamed bin Zayed University of Artificial Intelligence · Westlake University · Zhejiang University +2

  2. A KAN-based Non-linear Factorization for Embedding Layer Compression

    Oct 21, 2026Oliver Giudice, Anna Russo RussoEmbedding CompressionLow-Rank Compression

    Applied Research Team, IT Dept. Banca d’Italia *−

  3. SpecMRL: Redesign Matryoshka Representation Learning within Parametric Spectrum

    Oct 19, 2026Zhongxin Chen, Songhao Wu, Heng Cui +3Representation LearningMatryoshka Representation Learning

    Gaoling School of Artificial Intelligence, Renmin University of China · School of Computer Science, Wuhan University · Lenovo Group

  4. Embeddings Are Overpriced? A Frozen 16 MB Vocabulary for Small and Ternary Language Models

    Oct 15, 2026Sai Kiran BathulaEmbedding CompressionSmall Language Models

    QLNI, New South Wales, Australia

  5. Geometric Structure Predicts Per-Example Neighborhood Loss Under PCA Truncation

    Oct 13, 2026Sriram Loganathan, Gokul Anand, Ragavan ArivazhaganText Embedding EvaluationEmbedding Compression

    Department of Computer Science, San Jose State University, San Jose, CA, USA

  6. Simple Linear Corrections for Compression in Late Interaction

    Oct 9, 2026Mihir Agarwal, Sabyasachi Basu, Viraj AgrawalLate Interaction RetrievalMulti-Vector Retrieval

    Microsoft Research India

  7. Lossy Compressive Text Autoencoders

    Oct 7, 2026Vinko Sabolčec, Angelos Katharopoulos, David GrangierAutoencodersRepresentation Learning

    Apple

  8. OrBIT: Structure-Guided Embedding Compression

    Oct 7, 2026Yunied Puig, Amit Kumar JaiswalEmbedding CompressionLLM Compression

    Jay Chaudhry Software Innovation Centre Indian Institute of Technology (BHU) Varanasi, India

  9. MRVQ: One Resident Index for Dimension- and Rate-Elastic Vector Search

    Oct 2, 2026Sean Culatana, Shang-En Huang, Kang LiEmbedding CompressionDense Retrieval

    Atlassian Mountain View, CA · National Taiwan University Taipei, Taiwan · Atlassian Bellevue, WA

  10. Planetary Feature Fields are Scalable Earth Representations

    Sep 29, 2026Arjun Rao, Sebastian Loeschcke, Anthony Fuller +3Implicit Neural RepresentationsGeospatial Representation Learning

    University of British Columbia · University of Copenhagen · Carleton University +2

  11. MAGIC: Marginal-Guided Compression with Optimal Transport for Efficient Visual Document Retrieval

    Sep 17, 2026Xu Yuan, Hua Liu, Wenqi Fan +1Visual Document RetrievalEmbedding Compression

    The Hong Kong Polytechnic University, HK SAR · Southern University of Science and Technology, China

  12. Exact semantic readout from compressed vector representations

    Sep 16, 2026Daniel QuigleyLinear ProbingEmbedding Compression

    Center for Possible Minds Indiana University Bloomington Bloomington, IN 47408

  13. Hybrid Quantum-Classical NLP Classification with Compact Semantic Representations: An Experimental Analysis of Representation Compression

    Sep 9, 2026Ali Hassan, Zijia Zhao, Maha A. MetaweiHybrid Quantum-Classical MLText Classification

    German University in Cairo, Cairo, Egypt · University of Melbourne, Melbourne, Australia · High Performance Computing Lab, Electronics Research Institute, Cairo, Egypt

  14. FastE: Readout-Triggered Token Compression for LLM Embedding Inference

    Sep 8, 2026Jinsong Shu, Jinyong Wen, Baokun Wang +4Token PruningLLM Inference Acceleration

    Zhejiang University · Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security · Ant Group +1

  15. Matryoshka Hash Representations for Model-Aware Compact Semantic Retrieval

    Sep 7, 2026Peichun Hua, Yunming XiaoMatryoshka Representation LearningEmbedding Compression

    The Chinese University of Hong Kong, Shenzhen

  16. Spruce: Scalable Private Outsourced Retrieval Using Compact Embeddings

    Sep 3, 2026Peichun Hua, Yunming XiaoSecure Multi-Party ComputationEmbedding Compression

    The Chinese University of Hong Kong, Shenzhen · State Key Laboratory of Internet Architecture, Tsinghua University

  17. RSLM: Training-Free Vector Quantization for Approximate Nearest Neighbor Search

    Aug 31, 2026Rastislav Lenhardt, Teodora Dobos, Thomas Vecchiato +2Rotation-Based QuantizationVector Quantization

    Google Zurich, Switzerland · Technical University of Munich Munich, Germany · University of Copenhagen Copenhagen, Denmark +1

  18. Scalable partial information decomposition for symptom networks via supervised embeddings

    Aug 14, 2026Cillian Hourican, Eric Dignum, Rick Quax +1Embedding CompressionMental Health

    Computational Science Lab, Institute for Informatics, University of Amsterdam

  19. Tokens are All You Need: Dual-purpose Semantic IDs for Achieving LLM-Level I/O Efficiency in recommendation systems

    Jul 26, 2026Baolei Li, Yiping Yuan, Yilin Zheng +6Semantic ID RecommendationEmbedding Compression

    YouTube · Google Deepmind

  20. Progressive Cramming: Reliable Token Compression and What It Reveals

    Jul 23, 2026Dmitrii Tarasov, Timofei Lashukov, Elizaveta Goncharova +1LLM CompressionEmbedding Compression

    FusionBrain Lab, AXXX, Moscow, Russia · HSE University, Moscow, Russia · Artificial Intelligence Institute, Innopolis University, Innopolis, Russia

  21. TurboVec: A Case Study in Cost-Efficient Private Retrieval for Enterprise RAG via Codebook-Oblivious Quantization

    Jul 18, 2026Navnit Shukla, Kamal Pandey, Omsankar TiwariRetrieval-Augmented GenerationVector Quantization

    Snowflake Inc. Tustin, California, USA · Rivian Automotive Inc. Irvine, United States of America · Google Deepmind Sunnyvale, California, USA

  22. BitNet Text Embeddings

    Jun 24, 2026Zhen Li, Xin Huang, Liang Wang +8LLM QuantizationEmbedding Compression

    Peking University · Microsoft Research

  23. Geometric and Information Compression of Representations in Deep Learning

    Jun 19, 2026Linara Adilova, Henning Petzka, Asja Fischer +1Neural Representation GeometryEmbedding Compression

    Research Center Trustworthy Data Science and Security, Dortmund · Faculty of Computer Science, Ruhr University Bochum · Signal Processing and Speech Communication Laboratory, Graz University of Technology +2

  24. Fast-TurboQuant: A Multiplier-Free Online Vector Quantization Approach

    Jun 19, 2026Pedro M. R. Pereira, Felipe A. P. de Figueiredo, Rausley A. A. de SouzaLLM QuantizationRotation-Based Quantization

    National Institute of Telecommunications (Inatel), Santa Rita do Sapucaí, MG 37536-001, Brazil

  25. Lost in a Single Vector: Improving Long-Document Retrieval with Chunk Evidence Aggregation

    Jun 17, 2026Shanshan Lyu, Yiwei Wang, Yujun Cai +2Long-Context RetrievalDocument Chunking

    Chongqing University · State Key Laboratory of AI Safety · Institute of Computing Technology, Chinese Academy of Sciences +2

  26. Sustainable Face Recognition on Low-Power Devices with VQ-VAE Embeddings

    Jun 13, 2026Christos Chronis, Georgios Th. Papadopoulos, Iraklis VarlamisEnergy-Efficient MLEdge Inference

    Dep. of Informatics and Telematics Harokopio University of Athens Athens, Greece

  27. ColBERTSaR: Sparsified ColBERT Index via Product Quantization

    Jun 4, 2026Eugene Yang, Andrew Yates, Dawn Lawrie +3Vector QuantizationMulti-Vector Retrieval

    Johns Hopkins University, Baltimore, MD, USA

  28. When Is 0.1% Enough? Analyzing the Combined Effects of Dimensionality Reduction and Quantization on Text Embedding Compression

    May 31, 2026Riku Kisako, Hayato Tsukagoshi, Ryohei SasanoEmbedding CompressionDimensionality Reduction

    Graduate School of Informatics, Nagoya University

  29. No More K-means: Single-Stage Sparse Coding for Efficient Multi-Vector Retrieval

    May 28, 2026Lixuan Guo, Yifei Wang, Tiansheng Wen +3Learned Sparse RetrievalMulti-Vector Retrieval

    Stony Brook University · Xidian University · Amazon AGI SF Lab +4

  30. Clark Hash: Stateless Sparse Johnson-Lindenstrauss Quantization for Neural Embeddings

    May 27, 2026Stanislav Kirdey, Clark Labs IncEmbedding CompressionNeural Network Quantization

    Clark Labs Inc.