Lune

EMNLP Findings2022

Invernet: An Inversion Attack Framework to Infer Fine-Tuning Datasets through Word Embeddings

Ishrak Hayet, Zijun Yao, Bo Luo

Abstract

Word embedding aims to learn the dense representation of words and has become a regular input preparation in many NLP tasks. Due to the data and computation intensive nature of learning embeddings from scratch, a more affordable way is to borrow the pretrained embedding available in public and fine-tune the embedding through a domain specific downstream dataset. A privacy concern can arise if a malicious owner of the pretrained embedding gets access to the fine-tuned embedding and tries to infer the critical information from the downstream datasets. In this study, we propose a novel embedding inversion framework called Invernet that materializes the privacy concern by inferring the context distribution in the downstream dataset, which can lead to key information breach. With extensive experimental studies on two real-world news datasets: Antonio Gulli's News and New York Times, we validate the feasibility of proposed privacy attack and demonstrate the effectiveness of Invernet on inferring downstream datasets based on multiple word embedding methods.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

2 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Exploiting Unintended Feature Leakage in Collaborative Learning cites Membership Inference Attacks Against Machine Learning Models
  • Text Embedding Inversion Security for Multilingual Language Models cites Membership Inference Attacks Against Machine Learning Models
  • Text Embedding Inversion Security for Multilingual Language Models cites Information Leakage in Embedding Models
  • Against All Odds: Overcoming Typology, Script, and Language Confusion in Multilingual Embedding Inversion Attacks cites Text Embedding Inversion Security for Multilingual Language Models
  • Information Leakage in Embedding Models cites Exploiting Unintended Feature Leakage in Collaborative Learning
  • Information Leakage in Embedding Models cites Membership Inference Attacks Against Machine Learning Models
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers2

  1. Text Embedding Inversion Security for Multilingual Language ModelsYiyi Chen, Heather C. Lent, Johannes BjervaACL 2024 · 10 citations
  2. Against All Odds: Overcoming Typology, Script, and Language Confusion in Multilingual Embedding Inversion AttacksYiyi Chen, Russa Biswas, Heather C. Lent, Johannes BjervaAAAI 2025 · 9 citations

Builds on4

  1. Membership Inference Attacks Against Machine Learning ModelsReza Shokri, Marco Stronati, Congzheng Song, Vitaly ShmatikovS&P 2017 · 5,137 citations
  2. Exploiting Unintended Feature Leakage in Collaborative LearningLuca Melis, Congzheng Song, Emiliano De Cristofaro, Vitaly ShmatikovS&P 2019 · 1,736 citations
  3. Information Leakage in Embedding ModelsCongzheng Song, Ananth RaghunathanCCS 2020 · 200 citations
  4. Intermediate-Task Transfer Learning with Pretrained Language Models: When and Why Does It Work?Yada Pruksachatkun, Jason Phang, Haokun Liu, Phu Mon Htut et al.ACL 2020 · 168 citations

Related papers

  1. Transferable Embedding Inversion Attack: Uncovering Privacy Risks in Text Embeddings without Model QueriesYu-Hsiang Huang, Yu-Che Tsai, Hsiang Hsiao, Hong-Yi Lin et al.ACL 2024 · 5 citations
  2. Forget to Flourish: Leveraging Machine-Unlearning on Pretrained Language Models for Privacy LeakageMd. Rafi Ur Rashid, Jing Liu, Toshiaki Koike-Akino, Ye Wang et al.AAAI 2025 · 17 citations
  3. Black-Box Embedding Inversion Attack on Vector DatabasesLichao Sun, Yuncheng Wu, Haichao Sha, Xinjian Luo et al.KDD 2026
  4. Humpty Dumpty: Controlling Word Meanings via Corpus PoisoningRoei Schuster, Tal Schuster, Yoav Meri, Vitaly ShmatikovS&P 2020 · 42 citations
  5. ALGEN: Few-shot Inversion Attacks on Textual Embeddings via Cross-Model Alignment and GenerationYiyi Chen, Qiongkai Xu, Johannes BjervaACL 2025
Dusk over the sea between two cliffs drawn in fine vertical lines