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

arXiv: https://arxiv.org/abs/2603.09185 · HTML: https://mlsanity.com/paper/2603.09185/

## Key facts

- Cited by 2 papers in the map (data through 14 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.

## Placed in

- Retrieval-augmented generation / [Generation and context integration](https://mlsanity.com/rag/generation-and-context-integration/)

## Builds on (most cited first, 13 of 13)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [VQA: Visual Question Answering](https://mlsanity.com/paper/1505.00468/) (2015)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [Vision-Language Models Do Not Understand Negation](https://mlsanity.com/paper/2501.09425/) (2025)
- [TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives](https://mlsanity.com/paper/2411.02545/) (2024)
- [NevIR: Negation in Neural Information Retrieval](https://mlsanity.com/paper/2305.07614/) (2023)
- [Interpret and Control Dense Retrieval with Sparse Latent Features](https://mlsanity.com/paper/2411.00786/) (2024)
- [TNG-CLIP:Training-Time Negation Data Generation for Negation Awareness of CLIP](https://mlsanity.com/paper/2505.18434/) (2025)
- [NUDGE: Lightweight Non-Parametric Fine-Tuning of Embeddings for Retrieval](https://mlsanity.com/paper/2409.02343/) (2024)
- [Logical Consistency is Vital: Neural-Symbolic Information Retrieval for Negative-Constraint Queries](https://mlsanity.com/paper/2505.22299/) (2025)
- [Visual Question Answering based on Local-Scene-Aware Referring Expression Generation](https://mlsanity.com/paper/2101.08978/) (2021)

## Built on by (most cited first, 2 of 2)

- [flexvec: SQL Vector Retrieval with Programmatic Embedding Modulation](https://mlsanity.com/paper/2603.22587/) (2026)
- [E-SENS: Exclusion-Sensitive Penalization for Negative-Constraint Retrieval](https://mlsanity.com/paper/2608.30130/) (2026)

## 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.