60 citations · 105 across the 17 of their papers we have counts for
17 papers
REER-PT: Reverse-Engineered Reasoning for Perplexity-Guided Pre-training Data Augmentation
Haoran Que, Jiajun Shi, Ting Huang +7
As language-model compute continues to scale, high-quality training data is becoming an increasingly important bottleneck. Conventional next-token prediction supervises what follow…
Correct, Concise and Complete: Multi-stage Training For Adaptive Reasoning
Nathanaël Carraz Rakotonirina, Ren Pang, Neha Anna John +2
The reasoning capabilities of large language models (LLMs) have improved substantially through increased test-time computation, typically in the form of intermediate tokens known a…
Watch the Watcher! Backdoor Attacks on Security-Enhancing Diffusion Models
Changjiang Li, Ren Pang, Bochuan Cao +4
Thanks to their remarkable denoising capabilities, diffusion models are increasingly being employed as defensive tools to reinforce the security of other models, notably in purifyi…
On the Difficulty of Defending Contrastive Learning against Backdoor Attacks
Changjiang Li, Ren Pang, Bochuan Cao +4
Recent studies have shown that contrastive learning, like supervised learning, is highly vulnerable to backdoor attacks wherein malicious functions are injected into target models,…
Model Extraction Attacks Revisited
Jiacheng Liang, Ren Pang, Changjiang Li +1
Model extraction (ME) attacks represent one major threat to Machine-Learning-as-a-Service (MLaaS) platforms by ``stealing'' the functionality of confidential machine-learning model…
Defending Pre-trained Language Models as Few-shot Learners against Backdoor Attacks
Zhaohan Xi, Tianyu Du, Changjiang Li +5
Pre-trained language models (PLMs) have demonstrated remarkable performance as few-shot learners. However, their security risks under such settings are largely unexplored. In this…
On the Security Risks of Knowledge Graph Reasoning
Zhaohan Xi, Tianyu Du, Changjiang Li +6
Knowledge graph reasoning (KGR) -- answering complex logical queries over large knowledge graphs -- represents an important artificial intelligence task, entailing a range of appli…
Neural Architectural Backdoors
Ren Pang, Changjiang Li, Zhaohan Xi +2
This paper asks the intriguing question: is it possible to exploit neural architecture search (NAS) as a new attack vector to launch previously improbable attacks? Specifically, we…
An Embarrassingly Simple Backdoor Attack on Self-supervised Learning
Changjiang Li, Ren Pang, Zhaohan Xi +4
As a new paradigm in machine learning, self-supervised learning (SSL) is capable of learning high-quality representations of complex data without relying on labels. In addition to…
Reasoning over Multi-view Knowledge Graphs
Zhaohan Xi, Ren Pang, Changjiang Li +4
Recently, knowledge representation learning (KRL) is emerging as the state-of-the-art approach to process queries over knowledge graphs (KGs), wherein KG entities and the query are…
Towards Robust Reasoning over Knowledge Graphs
Zhaohan Xi, Ren Pang, Changjiang Li +4
Answering complex logical queries over large-scale knowledge graphs (KGs) represents an important artificial intelligence task, entailing a range of applications. Recently, knowled…
On the Security Risks of AutoML
Ren Pang, Zhaohan Xi, Shouling Ji +2
Neural Architecture Search (NAS) represents an emerging machine learning (ML) paradigm that automatically searches for models tailored to given tasks, which greatly simplifies the…
i-Algebra: Towards Interactive Interpretability of Deep Neural Networks
Xinyang Zhang, Ren Pang, Shouling Ji +2
Providing explanations for deep neural networks (DNNs) is essential for their use in domains wherein the interpretability of decisions is a critical prerequisite. Despite the pleth…
TrojanZoo: Towards Unified, Holistic, and Practical Evaluation of Neural Backdoors
Ren Pang, Zheng Zhang, Xiangshan Gao +5
Neural backdoors represent one primary threat to the security of deep learning systems. The intensive research has produced a plethora of backdoor attacks/defenses, resulting in a…
AdvMind: Inferring Adversary Intent of Black-Box Attacks
Ren Pang, Xinyang Zhang, Shouling Ji +2
Deep neural networks (DNNs) are inherently susceptible to adversarial attacks even under black-box settings, in which the adversary only has query access to the target models. In p…
Graph Backdoor
Zhaohan Xi, Ren Pang, Shouling Ji +1
One intriguing property of deep neural networks (DNNs) is their inherent vulnerability to backdoor attacks -- a trojan model responds to trigger-embedded inputs in a highly predict…
A Tale of Evil Twins: Adversarial Inputs versus Poisoned Models
Ren Pang, Hua Shen, Xinyang Zhang +5
Despite their tremendous success in a range of domains, deep learning systems are inherently susceptible to two types of manipulations: adversarial inputs -- maliciously crafted sa…