Related papers: Shifting the Gradient: Understanding How Defensive…
Large language models (LLMs) require precise behavior control for safe and effective deployment across diverse applications. Activation steering offers a promising approach for LLMs' behavioral control. We focus on the question of how…
Gradient-based saliency methods such as Vanilla Gradient (VG) and Integrated Gradients (IG) are widely used to explain image classifiers, yet the resulting maps are often noisy and unstable, limiting their usefulness in high-stakes…
As Large Language Models (LLMs) gain traction across critical domains, ensuring secure and trustworthy training processes has become a major concern. Backdoor attacks, where malicious actors inject hidden triggers into training data, are…
Inductive logic programming is a form of machine learning based on mathematical logic that generates logic programs from given examples and background knowledge. In this project, we extend the Popper ILP system to make use of multi-task…
Machine Unlearning has recently garnered significant attention, aiming to selectively remove knowledge associated with specific data while preserving the model's performance on the remaining data. A fundamental challenge in this process is…
Deep neural networks (DNNs) have had many successes, but they suffer from two major issues: (1) a vulnerability to adversarial examples and (2) a tendency to elude human interpretation. Interestingly, recent empirical and theoretical…
Self-improvement through post-training methods such as iterative preference learning has been acclaimed for enhancing the problem-solving capabilities (e.g., mathematical reasoning) of Large Language Models (LLMs) without human…
Adversarial training, especially projected gradient descent (PGD), has proven to be a successful approach for improving robustness against adversarial attacks. After adversarial training, gradients of models with respect to their inputs…
Large language models are increasingly used for text classification tasks such as sentiment analysis, yet their reliance on natural language prompts exposes them to prompt injection attacks. In particular, class-directive injections exploit…
Large Language Models (LLMs) have revolutionized the field of Natural Language Processing (NLP) by automating traditional labor-intensive tasks and consequently accelerated the development of computer-aided applications. As researchers…
Large language models (LLMs) often inherit biases from vast amounts of training corpora. Traditional debiasing methods, while effective to some extent, do not completely eliminate memorized biases and toxicity in LLMs. In this paper, we…
Large language models (LLMs) have achieved great success across diverse tasks, and fine-tuning is sometimes needed to further enhance generation quality. Most existing methods rely on human supervision or parameter retraining, both of which…
Gradient inversion attack (or input recovery from gradient) is an emerging threat to the security and privacy preservation of Federated learning, whereby malicious eavesdroppers or participants in the protocol can recover (partially) the…
In distributed learning settings, models are iteratively updated with shared gradients computed from potentially sensitive user data. While previous work has studied various privacy risks of sharing gradients, our paper aims to provide a…
Large-scale pre-trained models are increasingly adapted to downstream tasks through a new paradigm called prompt learning. In contrast to fine-tuning, prompt learning does not update the pre-trained model's parameters. Instead, it only…
While neural networks demonstrate stronger capabilities in pattern recognition nowadays, they are also becoming larger and deeper. As a result, the effort needed to train a network also increases dramatically. In many cases, it is more…
Steering vectors are a lightweight method to control language model behavior by adding a learned bias to the activations at inference time. Although steering demonstrates promising performance, recent work shows that it can be unreliable or…
Speech separation has been studied in time domain because of lower latency and higher performance compared to time-frequency domain. The masking-based method has been mostly used in time domain, and the other common method (mapping-based)…
Large Vision-Language Models (LVLMs) have demonstrated their powerful multimodal capabilities. However, they also face serious safety problems, as adversaries can induce robustness issues in LVLMs through the use of well-designed…
Face Presentation Attack Detection (PAD) demands incremental learning (IL) to combat evolving spoofing tactics and domains. Privacy regulations, however, forbid retaining past data, necessitating rehearsal-free IL (RF-IL). Vision-Language…