Zero2Text: Zero-Training Cross-Domain Inversion Attacks on Textual Embeddings
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910009583992832 |
|---|---|
| author | Kim, Doohyun Kang, Donghwa Lee, Kyungjae Baek, Hyeongboo Kang, Brent Byunghoon |
| author_facet | Kim, Doohyun Kang, Donghwa Lee, Kyungjae Baek, Hyeongboo Kang, Brent Byunghoon |
| contents | The proliferation of retrieval-augmented generation (RAG) has established vector databases as critical infrastructure, yet they introduce severe privacy risks via embedding inversion attacks. Existing paradigms face a fundamental trade-off: optimization-based methods require computationally prohibitive queries, while alignment-based approaches hinge on the unrealistic assumption of accessible in-domain training data. These constraints render them ineffective in strict black-box and cross-domain settings. To dismantle these barriers, we introduce Zero2Text, a novel training-free framework based on recursive online alignment. Unlike methods relying on static datasets, Zero2Text synergizes LLM priors with a dynamic ridge regression mechanism to iteratively align generation to the target embedding on-the-fly. We further demonstrate that standard defenses, such as differential privacy, fail to effectively mitigate this adaptive threat. Extensive experiments across diverse benchmarks validate Zero2Text; notably, on MS MARCO against the OpenAI victim model, it achieves 1.8x higher ROUGE-L and 6.4x higher BLEU-2 scores compared to baselines, recovering sentences from unknown domains without a single leaked data pair. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_01757 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Zero2Text: Zero-Training Cross-Domain Inversion Attacks on Textual Embeddings Kim, Doohyun Kang, Donghwa Lee, Kyungjae Baek, Hyeongboo Kang, Brent Byunghoon Computation and Language Machine Learning The proliferation of retrieval-augmented generation (RAG) has established vector databases as critical infrastructure, yet they introduce severe privacy risks via embedding inversion attacks. Existing paradigms face a fundamental trade-off: optimization-based methods require computationally prohibitive queries, while alignment-based approaches hinge on the unrealistic assumption of accessible in-domain training data. These constraints render them ineffective in strict black-box and cross-domain settings. To dismantle these barriers, we introduce Zero2Text, a novel training-free framework based on recursive online alignment. Unlike methods relying on static datasets, Zero2Text synergizes LLM priors with a dynamic ridge regression mechanism to iteratively align generation to the target embedding on-the-fly. We further demonstrate that standard defenses, such as differential privacy, fail to effectively mitigate this adaptive threat. Extensive experiments across diverse benchmarks validate Zero2Text; notably, on MS MARCO against the OpenAI victim model, it achieves 1.8x higher ROUGE-L and 6.4x higher BLEU-2 scores compared to baselines, recovering sentences from unknown domains without a single leaked data pair. |
| title | Zero2Text: Zero-Training Cross-Domain Inversion Attacks on Textual Embeddings |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2602.01757 |