Kai-Wei Chang

A View on Vulnerabilites: The Security Challenges of XAI (Academic Track)

2018 · arXiv (Cornell University)

Deep neural networks (DNNs) are vulnerable to adversarial examples, perturbations to correctly classified examples which can cause the model to misclassify. In the image domain, these perturbations are often virtually indistinguishable to human perception, causing …

Intent-aware Query Obfuscation for Privacy Protection in Personalized Web Search

2018

Modern web search engines exploit users' search history to personalize search results, with a goal of improving their service utility on a per-user basis. But it is this very dimension that leads to the risk …

Examining Gender Bias in Languages with Grammatical Gender

2019 · arXiv (Cornell University)

Recent studies have shown that word embeddings exhibit gender bias inherited from the training corpora. However, most studies to date have focused on quantifying and mitigating such bias only in English. These analyses cannot be …

The Woman Worked as a Babysitter: On Biases in Language Generation

2019 · arXiv (Cornell University)

We present a systematic study of biases in natural language generation (NLG) by analyzing text generated from prompts that contain mentions of different demographic groups. In this work, we introduce the notion of the regard …

Automatic Perturbation Analysis for Scalable Certified Robustness and Beyond

2020 · arXiv (Cornell University)

Linear relaxation based perturbation analysis (LiRPA) for neural networks, which computes provable linear bounds of output neurons given a certain amount of input perturbation, has become a core component in robustness verification and certified defense. …

"Nice Try, Kiddo": Investigating Ad Hominems in Dialogue Responses

2020 · arXiv (Cornell University)

Ad hominem attacks are those that target some feature of a person's character instead of the position the person is maintaining. These attacks are harmful because they propagate implicit biases and diminish a person's credibility. …

What do Bias Measures Measure

2021 · arXiv (Cornell University)

Natural Language Processing (NLP) models propagate social biases about protected attributes such as gender, race, and nationality. To create interventions and mitigate these biases and associated harms, it is vital to be able to detect …

Dynamic Prompt Learning via Policy Gradient for Semi-structured Mathematical Reasoning

2022 · arXiv (Cornell University)

Mathematical reasoning, a core ability of human intelligence, presents unique challenges for machines in abstract thinking and logical reasoning. Recent large pre-trained language models such as GPT-3 have achieved remarkable progress on mathematical reasoning tasks …

Efficient Shapley Values Estimation by Amortization for Text Classification

2023 · arXiv (Cornell University)

Despite the popularity of Shapley Values in explaining neural text classification models, computing them is prohibitive for large pretrained models due to a large number of model evaluations. In practice, Shapley Values are often estimated …

Symbolic Chain-of-Thought Distillation: Small Models Can Also "Think" Step-by-Step

2023 · arXiv (Cornell University)

Chain-of-thought prompting (e.g., "Let's think step-by-step") primes large language models to verbalize rationalization for their predictions. While chain-of-thought can lead to dramatic performance gains, benefits appear to emerge only for sufficiently large models (beyond 50B …

LLMs in Biomedicine: A study on clinical Named Entity Recognition

2024 · arXiv (Cornell University)

Large Language Models (LLMs) demonstrate remarkable versatility in various NLP tasks but encounter distinct challenges in biomedical due to the complexities of language and data scarcity. This paper investigates LLMs application in the biomedical domain …

Unlearning as multi-task optimization: A normalized gradient difference approach with an adaptive learning rate

2024 · arXiv (Cornell University)

Machine unlearning has been used to remove unwanted knowledge acquired by large language models (LLMs). In this paper, we examine machine unlearning from an optimization perspective, framing it as a regularized multi-task optimization problem, where …

Vulnerability of Large Language Models to Output Prefix Jailbreaks: Impact of Positions on Safety

2024

Previous research on jailbreak attacks has mainly focused on optimizing the adversarial snippet content injected into input prompts to expose LLM security vulnerabilities. A significant portion of this research focuses on developing more complex, less …

DRS: Deep Question Reformulation With Structured Output

2025

Question answering represents a core capability of large language models (LLMs).However, when individuals encounter unfamiliar knowledge in texts, they often formulate questions that the text itself cannot answer due to insufficient understanding of the underlying …

Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods

2018 · arXiv (Cornell University)

We introduce a new benchmark, WinoBias, for coreference resolution focused on gender bias. Our corpus contains Winograd-schema style sentences with entities corresponding to people referred by their occupation (e.g. the nurse, the doctor, the carpenter). …

Learning Gender-Neutral Word Embeddings

2018

Word embedding models have become a fundamental component in a wide range of Natural Language Processing (NLP) applications. However, embeddings trained on human-generated corpora have been demonstrated to inherit strong gender stereotypes that reflect social …

Gender Bias in Contextualized Word Embeddings

2019

Jieyu Zhao, Tianlu Wang, Mark Yatskar, Ryan Cotterell, Vicente Ordonez, Kai-Wei Chang. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and …

Papers by this author