{"ID":2890234,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.20030","arxiv_id":"2507.20030","title":"FAEDKV: Infinite-Window Fourier Transform for Unbiased KV Cache Compression","abstract":"The efficacy of Large Language Models (LLMs) in long-context tasks is often hampered by the substantial memory footprint and computational demands of the Key-Value (KV) cache. Current compression strategies, including token eviction and learned projections, frequently lead to biased representations -- either by overemphasizing recent/high-attention tokens or by repeatedly degrading information from earlier context -- and may require costly model retraining. We present FAEDKV (Frequency-Adaptive Infinite-Window for KV cache), a novel, training-free KV cache compression framework that ensures unbiased information retention. FAEDKV operates by transforming the KV cache into the frequency domain using a proposed Infinite-Window Fourier Transform (IWDFT). This approach allows for the equalized contribution of all tokens to the compressed representation, effectively preserving both early and recent contextual information. A preliminary frequency ablation study identifies critical spectral components for layer-wise, targeted compression. Experiments on LongBench benchmark demonstrate FAEDKV's superiority over existing methods by up to 22\\%. In addition, our method shows superior, position-agnostic retrieval accuracy on the Needle-In-A-Haystack task compared to compression based approaches.","short_abstract":"The efficacy of Large Language Models (LLMs) in long-context tasks is often hampered by the substantial memory footprint and computational demands of the Key-Value (KV) cache. Current compression strategies, including token eviction and learned projections, frequently lead to biased representations -- either by overemp...","url_abs":"https://arxiv.org/abs/2507.20030","url_pdf":"https://arxiv.org/pdf/2507.20030v1","authors":"[\"Runchao Li\",\"Yao Fu\",\"Mu Sheng\",\"Xianxuan Long\",\"Haotian Yu\",\"Pan Li\"]","published":"2025-07-26T18:20:25Z","proceeding":"cs.CL","tasks":"[\"cs.CL\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}