mlsanityData to 7 Oct 2026

DEO: Training-Free Direct Embedding Optimization for Negation-Aware Retrieval

arXiv 2603.09185 ↗First posted 10 Mar 2026cs.CLcited by 2

On the map: Retrieval-augmented generation → Generation and context integration

Abstract (arXiv)

Recent advances in Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) have enabled diverse retrieval methods. However, existing retrieval methods often fail to accurately retrieve results for negation and exclusion queries. To address this limitation, prior approaches rely on embedding adaptation or fine-tuning, which introduce additional computational cost and deployment complexity. We propose Direct Embedding Optimization (DEO), a training-free method for negation-aware text and multimodal retrieval. DEO decomposes queries into positive and negative components and optimizes the query embedding with a contrastive objective. Without additional training data or model updates, DEO outperforms baselines on NegConstraint, with gains of +0.0738 nDCG@10 and +0.1028 MAP@100, while improving Recall@5 by +6% over OpenAI CLIP in multimodal retrieval. These results demonstrate the practicality of DEO for negation- and exclusion-aware retrieval in real-world settings.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Generation and context integration, one of 15 directions in Retrieval-augmented generation. Builds on 13 of its 19 references that are in the map. First posted on arXiv 10 Mar 2026.

Builds on

13 of 19 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024NUDGE: Lightweight Non-Parametric Fine-Tuning of Embeddings for Retrieval4 Sep
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks22 May

Other references · 11

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,779
  3. 2015VQA: Visual Question Answeringcited by 2,489
  4. 2022When and why vision-language models behave like bags-of-words, and what to do about it?cited by 603
  5. 2025Vision-Language Models Do Not Understand Negationcited by 69
  6. 2024TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negativescited by 57
  7. 2023NevIR: Negation in Neural Information Retrievalcited by 28
  8. 2024Interpret and Control Dense Retrieval with Sparse Latent Featurescited by 9
  9. 2025TNG-CLIP:Training-Time Negation Data Generation for Negation Awareness of CLIPcited by 6
  10. 2025Logical Consistency is Vital: Neural-Symbolic Information Retrieval for Negative-Constraint Queriescited by 3
  11. 2021Visual Question Answering based on Local-Scene-Aware Referring Expression Generationcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026flexvec: SQL Vector Retrieval with Programmatic Embedding Modulationcited by 0
  2. 2026E-SENS: Exclusion-Sensitive Penalization for Negative-Constraint Retrievalcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Reason to Contrast: A Cascaded Multimodal Retrieval Framework
  2. 2025ELITE: Embedding-Less retrieval with Iterative Text Exploration
  3. 2026Skim and Skip: Hierarchical Adaptive Inference for Efficient Multimodal Retrieval
  4. 2025U-MARVEL: Unveiling Key Factors for Universal Multimodal Retrieval via Embedding Learning with MLLMs
  5. 2026HAKARI-Bench: A Lightweight Benchmark for Comparing Retrieval Architectures and Efficiency Settings under Unified Conditions
  6. 2024Embedding-Informed Adaptive Retrieval-Augmented Generation of Large Language Models
  7. 2025Exploring Reasoning-Infused Text Embedding with Large Language Models for Zero-Shot Dense Retrieval
  8. 2025Direct Retrieval-augmented Optimization: Synergizing Knowledge Selection and Language Models