Cathy Yuanchen Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Cathy Yuanchen Li has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Ben Y. Zhao
University of Chicago, Chicago, IL, USA
2 shared publications - Haitao Zheng
University of Chicago, Chicago, IL, USA
2 shared publications - Arman Bhalla
University of Chicago, Chicago, USA
1 shared publication - Emily Wenger
Department of Electrical and Computer Engineering, Duke University, Durham, North Carolina, USA
1 shared publication - Evrard Garcelon
Meta AI, Paris, France
1 shared publication - François Charton
FAIR, Meta
1 shared publication
Latest Publications
- Open Access
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Towards Hijacking Interactions in Virtual Reality
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessTowards Hijacking Interactions in Virtual Reality
- Open Access
- 2 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1211-1225
Understanding Implosion in Text-to-Image Generative Models
Recent works show that text-to-image generative models are surprisingly vulnerable to a variety of poisoning attacks. Empirical results find that these models can be corrupted by altering associations between individual text prompts and associated visual features. Furthermore, a number of concurrent poisoning attacks can induce "model implosion," where the model becomes unable to produce meaningful images for unpoisoned prompts. These intriguing findings highlight the absence of an intuitive framework to understand poisoning attacks on these models. In this work, we establish the first analytical framework on robustness of image generative models to poisoning attacks, by modeling and analyzing the behavior of the cross-attention mechanism in latent diffusion models. We model cross-attention training as an abstract problem of "supervised graph alignment" and formally quantify the impact of training data by the hardness of alignment, measured by an Alignment Difficulty (AD) metric. The higher the AD, the harder the alignment. We prove that AD increases with the number of individual prompts (or concepts) poisoned. As AD grows, the alignment task becomes increasingly difficult, yielding highly distorted outcomes that frequently map meaningful text prompts to undefined or meaningless visual representations. As a result, the generative model implodes and outputs random, incoherent images at large. We validate our analytical framework through extensive experiments, and we confirm and explain the unexpected (and unexplained) effect of model implosion while producing new, unforeseen insights. Our work provides a useful tool for studying poisoning attacks against diffusion models and their defenses.
- 2 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1211-1225
Open AccessUnderstanding Implosion in Text-to-Image Generative Models
Recent works show that text-to-image generative models are surprisingly vulnerable to a variety of poisoning attacks. Empirical results find that these models can be corrupted by altering associations between individual text prompts and associated visual features. Furthermore, a number of concurrent poisoning attacks can induce "model implosion," where the model becomes unable to produce meaningful images for unpoisoned prompts. These intriguing findings highlight the absence of an intuitive framework to understand poisoning attacks on these models. In this work, we establish the first analytical framework on robustness of image generative models to poisoning attacks, by modeling and analyzing the behavior of the cross-attention mechanism in latent diffusion models. We model cross-attention training as an abstract problem of "supervised graph alignment" and formally quantify the impact of training data by the hardness of alignment, measured by an Alignment Difficulty (AD) metric. The higher the AD, the harder the alignment. We prove that AD increases with the number of individual prompts (or concepts) poisoned. As AD grows, the alignment task becomes increasingly difficult, yielding highly distorted outcomes that frequently map meaningful text prompts to undefined or meaningless visual representations. As a result, the generative model implodes and outputs random, incoherent images at large. We validate our analytical framework through extensive experiments, and we confirm and explain the unexpected (and unexplained) effect of model implosion while producing new, unforeseen insights. Our work provides a useful tool for studying poisoning attacks against diffusion models and their defenses.
- Open Access
- 15 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 2606-2620
SalsaPicante: A Machine Learning Attack on LWE with Binary Secrets
Learning with Errors (LWE) is a hard math problem underpinning many proposed post-quantum cryptographic (PQC) systems. The only PQC Key Exchange Mechanism (KEM) standardized by NIST [13] is based on module LWE [2], and current publicly available PQ Homomorphic Encryption (HE) libraries are based on ring LWE. The security of LWE-based PQ cryptosystems is critical, but certain implementation choices could weaken them. One such choice is sparse binary secrets, desirable for PQ HE schemes for efficiency reasons. Prior work SALSA[51] demonstrated a machine learning-based attack on LWE with sparse binary secrets in small dimensions (n ≤ = 128) and low Hamming weights (h ≤ = 4). However, this attack assumes access to millions of eavesdropped LWE samples and fails at higher Hamming weights or dimensions. We present PICANTE, an enhanced machine learning attack on LWE with sparse binary secrets, which recovers secrets in much larger dimensions (up to n=350) and with larger Hamming weights (roughly n/10, and up to h=60 for n=350). We achieve this dramatic improvement via a novel preprocessing step, which allows us to generate training data from a linear number of eavesdropped LWE samples (4n) and changes the distribution of the data to improve transformer training. We also improve the secret recovery methods of SALSA and introduce a novel cross-attention recovery mechanism allowing us to read off the secret directly from the trained models. While PICANTE does not threaten NIST's proposed LWE standards, it demonstrates significant improvement over SALSA and could scale further, highlighting the need for future investigation into machine learning attacks on LWE with sparse binary secrets.
- 15 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 2606-2620
Open AccessSalsaPicante: A Machine Learning Attack on LWE with Binary Secrets
Learning with Errors (LWE) is a hard math problem underpinning many proposed post-quantum cryptographic (PQC) systems. The only PQC Key Exchange Mechanism (KEM) standardized by NIST [13] is based on module LWE [2], and current publicly available PQ Homomorphic Encryption (HE) libraries are based on ring LWE. The security of LWE-based PQ cryptosystems is critical, but certain implementation choices could weaken them. One such choice is sparse binary secrets, desirable for PQ HE schemes for efficiency reasons. Prior work SALSA[51] demonstrated a machine learning-based attack on LWE with sparse binary secrets in small dimensions (n ≤ = 128) and low Hamming weights (h ≤ = 4). However, this attack assumes access to millions of eavesdropped LWE samples and fails at higher Hamming weights or dimensions. We present PICANTE, an enhanced machine learning attack on LWE with sparse binary secrets, which recovers secrets in much larger dimensions (up to n=350) and with larger Hamming weights (roughly n/10, and up to h=60 for n=350). We achieve this dramatic improvement via a novel preprocessing step, which allows us to generate training data from a linear number of eavesdropped LWE samples (4n) and changes the distribution of the data to improve transformer training. We also improve the secret recovery methods of SALSA and introduce a novel cross-attention recovery mechanism allowing us to read off the secret directly from the trained models. While PICANTE does not threaten NIST's proposed LWE standards, it demonstrates significant improvement over SALSA and could scale further, highlighting the need for future investigation into machine learning attacks on LWE with sparse binary secrets.