Victim Model

Victim model research focuses on understanding and mitigating vulnerabilities in machine learning models, primarily concerning adversarial attacks and model extraction. Current efforts concentrate on developing more effective attack methods (e.g., leveraging gradient-based and generative adversarial approaches) and robust defenses, including techniques like noise injection and dataset inference. This research is crucial for enhancing the security and privacy of machine learning systems across various applications, from healthcare to finance, by identifying weaknesses and developing countermeasures against malicious exploitation.

Papers

October 18, 2024

REEF: Representation Encoding Fingerprints for Large Language Models
Jie Zhang, Dongrui Liu, Chen Qian, Linfeng Zhang, Yong Liu, Yu Qiao, Jing Shao
Open Source Large Language Model Coral Reef Fingerprint Representation Victim Model Suspect Model

July 23, 2024

MiranDa: Mimicking the Learning Processes of Human Doctors to Achieve Causal Inference for Medication Recommendation
Ziheng Wang, Xinhe Li, Haruki Momma, Ryoichi Nagatomi
Causal Inference Civil Engineering Phase Internal Medicine DOCTOR Medication Recommendation Learning Process Optimal Treatment Victim Model

April 8, 2024

Semantic Stealth: Adversarial Text Attacks on NLP Using Several Methods
Roopkatha Dey, Aivy Debnath, Sayak Kumar Dutta, Kaustav Ghosh, Arijit Mitra, Arghya Roy Chowdhury, Jaydip Sen
Adversarial Attack Adversarial Example Practical Method NLP Field BERT Based Textual Adversarial Attack Digital STEALTH Metric Victim Model

March 5, 2024

Precise Extraction of Deep Learning Models via Side-Channel Attacks on Edge/Endpoint Devices
Younghan Lee, Sohee Jun, Yungi Cho, Woorim Han, Hyungon Moon, Yunheung Paek
Deep Learning Model Extreme Edge Side Channel Model Extraction Attack Attack Vector Victim Model Endpoint Detection Efficient Extraction

February 29, 2024

Unraveling Adversarial Examples against Speaker Identification -- Techniques for Attack Detection and Victim Model Classification
Sonal Joshi, Thomas Thebaud, Jesús Villalba, Najim Dehak
Adversarial Attack Adversarial Example Speaker Verification Barzilai Borwein Technique Speaker Identification Attack Detection Victim Model Attack Classification

January 16, 2024

A Generative Adversarial Attack for Multilingual Text Classifiers
Tom Roth, Inigo Jauregi Unanue, Alsharif Abuadbba, Massimo Piccardi
Adversarial Attack Adversarial Example Generative Attack Victim Model Multilingual Classification Multilingual Text Classification

October 3, 2023

Beyond Labeling Oracles: What does it mean to steal ML models?
Avital Shafran, Ilia Shumailov, Murat A. Erdogdu, Nicolas Papernot
Machine Learning Model Model Extraction Attack Adversarial Objective Confidence Oracle Model Extraction Machine Learning a a Service Victim Model

July 3, 2023

Pareto-Secure Machine Learning (PSML): Fingerprinting and Securing Inference Serving Systems
Debopam Sanyal, Jui-Tse Hung, Manav Agrawal, Prahlad Jasti, Shahab Nikkhoo, Somesh Jha, Tianhao Wang, Sibin Mohan, Alexey Tumanov
Model Extraction Attack Model Deployment Secure Inference Fingerprinting Lemma Secure Machine Learning Victim Model

March 13, 2023

Can Adversarial Examples Be Parsed to Reveal Victim Model Information?
Yuguang Yao, Jiancheng Liu, Yifan Gong, Xiaoming Liu, Yanzhi Wang, Xue Lin, Sijia Liu
Adversarial Attack Adversarial Example Transferable Attack Adversarial Instance Victim Model

September 16, 2022

Dataset Inference for Self-Supervised Models
Adam Dziedzic, Haonan Duan, Muhammad Ahmad Kaleem, Nikita Dhawan, Jonas Guan, Yannis Cattan, Franziska Boenisch, Nicolas Papernot
Self Supervised Model Encoder Model Model Stealing Victim Model

May 1, 2022

A Word is Worth A Thousand Dollars: Adversarial Attack on Tweets Fools Stock Predictions
Yong Xie, Dakuo Wang, Pin-Yu Chen, Jinjun Xiong, Sijia Liu, Sanmi Koyejo
Adversarial Attack Real Text Word Adversarial Setting Stock Price Forecasting Adversarial Generation Victim Model

March 4, 2022