@inproceedings{wang-etal-2026-resonating,
    title = "Resonating with {R}o{PE}: Spectral Quantization for High-Fidelity Key Cache Compression",
    author = "Wang, Xuefei  and
      Tang, Haoyu  and
      Liang, Tianyuan  and
      Wang, Zhibin  and
      Hu, Yupeng  and
      Guan, Weili",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.1732/",
    pages = "37328--37348",
    ISBN = "979-8-89176-390-6",
    abstract = "The linear growth of KV cache bottlenecks long-context LLMs, yet RoPE-induced oscillations complicate Key cache quantization. To address this issue, we propose SpectrumQuant, a frequency-domain framework that utilizes the Discrete Cosine Transform (DCT) to convert these oscillations into sparse spectral representations. Specifically, our pipeline integrates dominant frequency extraction, hybrid bit-width allocation, and high-frequency pre-emphasis to maximize fidelity while minimizing memory footprint. To eliminate computational overhead, we develop fused Triton kernels featuring deferred inverse transformation and on-chip sparse accumulation. Extensive experiments on several benchmarks confirm SpectrumQuant achieves efficient compression with performance and latency comparable to FP16 baselines."
}