As large Vision-Language Models (VLMs) grow in prominence, ensuring their safe deployment has become critical. Recent studies have explored VLM robustness against jailbreak attacks--techniques that exploit model vulnerabilities to elicit harmful outputs. However, the limited availability of diverse multi-modal data has led current approaches to rely heavily on adversarial or manually crafted images derived from harmful text datasets, which may lack effectiveness and diversity across different contexts. In this paper, we propose a novel jailbreak method named IDEATOR, which autonomously generates malicious image-text pairs for black-box jailbreak attacks. IDEATOR is based on the insight that VLMs themselves could serve as powerful red team models for generating multimodal jailbreak prompts. Specifically, IDEATOR uses a VLM to create targeted jailbreak texts and pairs them with jailbreak images generated by a state-of-the-art diffusion model. Our extensive experiments demonstrate IDEATOR's high effectiveness and transferability. Notably, it achieves a 94% success rate in jailbreaking MiniGPT-4 with an average of only 5.34 queries, and high success rates of 82%, 88%, and 75% when transferred to LLaVA, InstructBLIP, and Meta's Chameleon, respectively. IDEATOR uncovers specific vulnerabilities in VLMs under black-box conditions, underscoring the need for improved safety mechanisms.

Patient-centered research is increasingly important in narrowing the gap between research and patient care, yet incorporating patient perspectives into health research has been inconsistent. We propose an automated framework leveraging innovative natural language processing (NLP) and artificial intelligence (AI) with patient portal messages to generate research ideas that prioritize important patient issues. We further quantified the quality of AI-generated research topics. To define patient clinical concerns, we analyzed 614,464 patient messages from 25,549 individuals with breast or skin cancer obtained from a large academic hospital (2013 to 2024), constructing a 2-staged unsupervised NLP topic model. Then, we generated research topics to resolve the defined issues using a widely used AI (ChatGPT-4o, OpenAI Inc, April 2024 version) with prompt-engineering strategies. We guided AI to perform multi-level tasks: 1) knowledge interpretation and summarization (e.g., interpreting and summarizing the NLP-defined topics), 2) knowledge generation (e.g., generating research ideas corresponding to patients issues), 3) self-reflection and correction (e.g., ensuring and revising the research ideas after searching for scientific articles), and 4) self-reassurance (e.g., confirming and finalizing the research ideas). Six highly experienced breast oncologists and dermatologists assessed the significance and novelty of AI-generated research topics using a 5-point Likert scale (1-exceptional, 5-poor). One-third of the AI-suggested research topics were highly significant and novel when both scores were lower than the average. Two-thirds of the AI-suggested topics were novel in both cancers. Our findings demonstrate that AI-generated research topics reflecting patient perspectives via a large volume of patient messages can meaningfully guide future directions in patient-centered health research.

There is an urgent need to identify both short and long-term risks from newly emerging types of Artificial Intelligence (AI), as well as available risk management measures. In response, and to support global efforts in regulating AI and writing safety standards, we compile an extensive catalog of risk sources and risk management measures for general-purpose AI (GPAI) systems, complete with descriptions and supporting examples where relevant. This work involves identifying technical, operational, and societal risks across model development, training, and deployment stages, as well as surveying established and experimental methods for managing these risks. To the best of our knowledge, this paper is the first of its kind to provide extensive documentation of both GPAI risk sources and risk management measures that are descriptive, self-contained and neutral with respect to any existing regulatory framework. This work intends to help AI providers, standards experts, researchers, policymakers, and regulators in identifying and mitigating systemic risks from GPAI systems. For this reason, the catalog is released under a public domain license for ease of direct use by stakeholders in AI governance and standards.

In today's digitally driven world, dialogue systems play a pivotal role in enhancing user interactions, from customer service to virtual assistants. In these dialogues, it is important to identify user's goals automatically to resolve their needs promptly. This has necessitated the integration of models that perform Intent Detection. However, users' intents are diverse and dynamic, making it challenging to maintain a fixed set of predefined intents. As a result, a more practical approach is to develop a model capable of identifying new intents as they emerge. We address the challenge of Intent Discovery, an area that has drawn significant attention in recent research efforts. Existing methods need to train on a substantial amount of data for correctly identifying new intents, demanding significant human effort. To overcome this, we introduce IntentGPT, a novel training-free method that effectively prompts Large Language Models (LLMs) such as GPT-4 to discover new intents with minimal labeled data. IntentGPT comprises an \textit{In-Context Prompt Generator}, which generates informative prompts for In-Context Learning, an \textit{Intent Predictor} for classifying and discovering user intents from utterances, and a \textit{Semantic Few-Shot Sampler} that selects relevant few-shot examples and a set of known intents to be injected into the prompt. Our experiments show that IntentGPT outperforms previous methods that require extensive domain-specific data and fine-tuning, in popular benchmarks, including CLINC and BANKING, among others.

Large language models memorize parts of their training data. Memorizing short snippets and facts is required to answer questions about the world and to be fluent in any language. But models have also been shown to reproduce long verbatim sequences of memorized text when prompted by a motivated adversary. In this work, we investigate an intermediate regime of memorization that we call non-adversarial reproduction, where we quantify the overlap between model responses and pretraining data when responding to natural and benign prompts. For a variety of innocuous prompt categories (e.g., writing a letter or a tutorial), we show that up to 15% of the text output by popular conversational language models overlaps with snippets from the Internet. In worst cases, we find generations where 100% of the content can be found exactly online. For the same tasks, we find that human-written text has far less overlap with Internet data. We further study whether prompting strategies can close this reproduction gap between models and humans. While appropriate prompting can reduce non-adversarial reproduction on average, we find that mitigating worst-case reproduction of training data requires stronger defenses -- even for benign interactions.

The Structured Dialogue System, referred to as SuDoSys, is an innovative Large Language Model (LLM)-based chatbot designed to provide psychological counseling. SuDoSys leverages the World Health Organization (WHO)'s Problem Management Plus (PM+) guidelines to deliver stage-aware multi-turn dialogues. Existing methods for employing an LLM in multi-turn psychological counseling typically involve direct fine-tuning using generated dialogues, often neglecting the dynamic stage shifts of counseling sessions. Unlike previous approaches, SuDoSys considers the different stages of counseling and stores essential information throughout the counseling process, ensuring coherent and directed conversations. The system employs an LLM, a stage-aware instruction generator, a response unpacker, a topic database, and a stage controller to maintain dialogue flow. In addition, we propose a novel technique that simulates counseling clients to interact with the evaluated system and evaluate its performance automatically. When assessed using both objective and subjective evaluations, SuDoSys demonstrates its effectiveness in generating logically coherent responses. The system's code and program scripts for evaluation are open-sourced.

The increasing electricity demands of personal computers, communication networks, and data centers contribute to higher atmospheric greenhouse gas emissions, which in turn lead to global warming and climate change. Therefore the energy consumption of code must be minimized. Code can be generated by large language models. We look at the influence of prompt modification on the energy consumption of the code generated. We use three different Python code problems of varying difficulty levels. Prompt modification is done by adding the sentence ``Give me an energy-optimized solution for this problem'' or by using two Python coding best practices. The large language models used are CodeLlama-70b, CodeLlama-70b-Instruct, CodeLlama-70b-Python, DeepSeek-Coder-33b-base, and DeepSeek-Coder-33b-instruct. We find a decrease in energy consumption for a specific combination of prompt optimization, LLM, and Python code problem. However, no single optimization prompt consistently decreases energy consumption for the same LLM across the different Python code problems.

Theoretical linguistics seeks to explain what human language is, and why. Linguists and cognitive scientists have proposed different theoretical models of what language is, as well as cognitive factors that shape it, and allow humans to 'produce', 'understand', and 'acquire' natural languages. However, humans may no longer be the only ones learning to 'generate', 'parse', and 'learn' natural language: artificial intelligence (AI) models such as large language models are proving to have impressive linguistic capabilities. Many are thus questioning what role, if any, such models should play in helping theoretical linguistics reach its ultimate research goals? In this paper, we propose to answer this question, by reiterating the tenets of generative linguistics, a leading school of thought in the field, and by considering how AI models as theories of language relate to each of these important concepts. Specifically, we consider three foundational principles, finding roots in the early works of Noam Chomsky: (1) levels of theoretical adequacy; (2) procedures for linguistic theory development; (3) language learnability and Universal Grammar. In our discussions of each principle, we give special attention to two types of AI models: neural language models and neural grammar induction models. We will argue that such models, in particular neural grammar induction models, do have a role to play, but that this role is largely modulated by the stance one takes regarding each of these three guiding principles.

Large Multi-Modal Models (LMMs) have demonstrated impressive capabilities as general-purpose chatbots that can engage in conversations about a provided input, such as an image. However, their responses are influenced by societal biases present in their training datasets, leading to undesirable differences in how the model responds when presented with images depicting people of different demographics. In this work, we propose a novel debiasing framework for LMMs that directly removes biased representations during text generation to decrease outputs related to protected attributes, or even representing them internally. Our proposed method is training-free; given a single image and a list of target attributes, we can ablate the corresponding representations with just one step of gradient descent on the image itself. Our experiments show that not only can we can minimize the propensity of LMMs to generate text related to protected attributes, but we can improve sentiment and even simply use synthetic data to inform the ablation while retaining language modeling capabilities on real data such as COCO or FACET. Furthermore, we find the resulting generations from a debiased LMM exhibit similar accuracy as a baseline biased model, showing that debiasing effects can be achieved without sacrificing model performance.

The success of deep neural networks has driven numerous research studies and applications from Euclidean to non-Euclidean data. However, there are increasing concerns about privacy leakage, as these networks rely on processing private data. Recently, a new type of privacy attack, the model inversion attacks (MIAs), aims to extract sensitive features of private data for training by abusing access to a well-trained model. The effectiveness of MIAs has been demonstrated in various domains, including images, texts, and graphs. These attacks highlight the vulnerability of neural networks and raise awareness about the risk of privacy leakage within the research community. Despite the significance, there is a lack of systematic studies that provide a comprehensive overview and deeper insights into MIAs across different domains. This survey aims to summarize up-to-date MIA methods in both attacks and defenses, highlighting their contributions and limitations, underlying modeling principles, optimization challenges, and future directions. We hope this survey bridges the gap in the literature and facilitates future research in this critical area. Besides, we are maintaining a repository to keep track of relevant research at https://github.com/AndrewZhou924/Awesome-model-inversion-attack.