mlsanityData to 14 Oct 2026

Spectral Retrieval: Multi-Scale Sinc Convolution over Token Embeddings for Localized Retrieval in LLM Multi-Agent Systems

arXiv 2605.24764 ↗First posted 23 May 2026cs.IR cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

[Abridged] - Spectral Retrieval is a plug-in re-ranking stage that interpolates between per-token MaxSim and mean-pool retrieval through a multi-scale sinc convolution over token embeddings. In standard dense retrieval each document is one mean-pooled vector; when relevance localises into a short subspan, the signal averages into noise. Spectral Retrieval reuses per-token embeddings from a late-interaction index and convolves them with a normalised sinc kernel at multiple scales. At L=1 the kernel acts as the identity, recovering per-token MaxSim; as L grows it approaches a uniform filter, recovering mean pooling. The maximum cosine over positions and scales yields a score provably no less informative than either endpoint. On a controlled synthetic benchmark with 1,000 documents and planted single-position spikes, mean-pool retrieval sits at chance (Recall@10 ~ 0.02) regardless of spike strength, while Spectral Retrieval reaches Recall@10 = 1.0 once the planted cosine exceeds the corpus-level token noise floor. On LIMIT-small with a frozen all-mpnet-base-v2 encoder, Spectral Retrieval lifts Recall@10 from 0.33 to 0.90, MRR from 0.22 to 0.79, and strict Success@10 from 0.12 to 0.84, without retraining. The method fits naturally into multi-agent LLM systems, where each agent benefits from a tighter, role-specific retrieval window over a shared corpus.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 12 of its 16 references that are in the map. First posted on arXiv 23 May 2026.

Builds on

12 of 16 references.

References · 12

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  3. 2019Sentence-BERT: Sentence Embeddings using Siamese BERT-Networkscited by 9,170
  4. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  5. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  6. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,784
  7. 2016MS MARCO: A Human Generated MAchine Reading COmprehension Datasetcited by 1,605
  8. 2021BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Modelscited by 1,252
  9. 2020ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERTcited by 1,178
  10. 2021ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interactioncited by 476
  11. 2021SPLADE: Sparse Lexical and Expansion Model for First Stage Rankingcited by 286
  12. 2025On the Theoretical Limitations of Embedding-Based Retrievalcited by 104

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026From Embedding Geometry to Spectral Search: Energy Dispersion Networks For Vector Retrieval ↗
  2. 2026Visual RAG Toolkit: Scaling Multi-Vector Visual Retrieval with Training-Free Pooling and Multi-Stage Search
  3. 2026ReinPool: Reinforcement Learning Pooling Multi-Vector Embeddings for Retrieval System
  4. 2026Hydra: Unifying Document Retrieval and Generation in a Single Vision-Language Model
  5. 2026Delving into Spectral Clustering with Vision-Language Representations
  6. 2026ATFlash: Per-RoPE-Wavelength Attention Windows for Compute/Memory-Efficient LLM Inference
  7. 2026Where Post-Training Quantization Breaks Text Embedders: A Measured Map Across Four Embedder Families
  8. 2026SOLAR: Self-supervised Joint Learning for Symmetric Multimodal Retrieval