Foundation Models that are capable of processing and generating multi-modal data have transformed AI's role in medicine. However, a key limitation of their reliability is hallucination, where inaccurate or fabricated information can impact clinical decisions and patient safety. We define medical hallucination as any instance in which a model generates misleading medical content. This paper examines the unique characteristics, causes, and implications of medical hallucinations, with a particular focus on how these errors manifest themselves in real-world clinical scenarios. Our contributions include (1) a taxonomy for understanding and addressing medical hallucinations, (2) benchmarking models using medical hallucination dataset and physician-annotated LLM responses to real medical cases, providing direct insight into the clinical impact of hallucinations, and (3) a multi-national clinician survey on their experiences with medical hallucinations. Our results reveal that inference techniques such as Chain-of-Thought (CoT) and Search Augmented Generation can effectively reduce hallucination rates. However, despite these improvements, non-trivial levels of hallucination persist. These findings underscore the ethical and practical imperative for robust detection and mitigation strategies, establishing a foundation for regulatory policies that prioritize patient safety and maintain clinical integrity as AI becomes more integrated into healthcare. The feedback from clinicians highlights the urgent need for not only technical advances but also for clearer ethical and regulatory guidelines to ensure patient safety. A repository organizing the paper resources, summaries, and additional information is available at https://github.com/mitmedialab/medical hallucination.

Many practical vision-language applications require models that understand negation, e.g., when using natural language to retrieve images which contain certain objects but not others. Despite advancements in vision-language models (VLMs) through large-scale training, their ability to comprehend negation remains underexplored. This study addresses the question: how well do current VLMs understand negation? We introduce NegBench, a new benchmark designed to evaluate negation understanding across 18 task variations and 79k examples spanning image, video, and medical datasets. The benchmark consists of two core tasks designed to evaluate negation understanding in diverse multimodal settings: Retrieval with Negation and Multiple Choice Questions with Negated Captions. Our evaluation reveals that modern VLMs struggle significantly with negation, often performing at chance level. To address these shortcomings, we explore a data-centric approach wherein we finetune CLIP models on large-scale synthetic datasets containing millions of negated captions. We show that this approach can result in a 10% increase in recall on negated queries and a 40% boost in accuracy on multiple-choice questions with negated captions.

For medical imaging AI models to be clinically impactful, they must generalize. However, this goal is hindered by (i) diverse types of distribution shifts, such as temporal, demographic, and label shifts, and (ii) limited diversity in datasets that are siloed within single medical institutions. While these limitations have spurred interest in federated learning, current evaluation benchmarks fail to evaluate different shifts simultaneously. However, in real healthcare settings, multiple types of shifts co-exist, yet their impact on medical imaging performance remains unstudied. In response, we introduce FedMedICL, a unified framework and benchmark to holistically evaluate federated medical imaging challenges, simultaneously capturing label, demographic, and temporal distribution shifts. We comprehensively evaluate several popular methods on six diverse medical imaging datasets (totaling 550 GPU hours). Furthermore, we use FedMedICL to simulate COVID-19 propagation across hospitals and evaluate whether methods can adapt to pandemic changes in disease prevalence. We find that a simple batch balancing technique surpasses advanced methods in average performance across FedMedICL experiments. This finding questions the applicability of results from previous, narrow benchmarks in real-world medical settings.

Recent progress in empirical and certified robustness promises to deliver reliable and deployable Deep Neural Networks (DNNs). Despite that success, most existing evaluations of DNN robustness have been done on images sampled from the same distribution on which the model was trained. However, in the real world, DNNs may be deployed in dynamic environments that exhibit significant distribution shifts. In this work, we take a first step towards thoroughly investigating the interplay between empirical and certified adversarial robustness on one hand and domain generalization on another. To do so, we train robust models on multiple domains and evaluate their accuracy and robustness on an unseen domain. We observe that: (1) both empirical and certified robustness generalize to unseen domains, and (2) the level of generalizability does not correlate well with input visual similarity, measured by the FID between source and target domains. We also extend our study to cover a real-world medical application, in which adversarial augmentation significantly boosts the generalization of robustness with minimal effect on clean data accuracy.

Modern machine learning pipelines are limited due to data availability, storage quotas, privacy regulations, and expensive annotation processes. These constraints make it difficult or impossible to train and update large-scale models on such dynamic annotated sets. Continual learning directly approaches this problem, with the ultimate goal of devising methods where a deep neural network effectively learns relevant patterns for new (unseen) classes, without significantly altering its performance on previously learned ones. In this paper, we address the problem of continual learning for video data. We introduce PIVOT, a novel method that leverages extensive knowledge in pre-trained models from the image domain, thereby reducing the number of trainable parameters and the associated forgetting. Unlike previous methods, ours is the first approach that effectively uses prompting mechanisms for continual learning without any in-domain pre-training. Our experiments show that PIVOT improves state-of-the-art methods by a significant 27% on the 20-task ActivityNet setup.

Current evaluations of Continual Learning (CL) methods typically assume that there is no constraint on training time and computation. This is an unrealistic assumption for any real-world setting, which motivates us to propose: a practical real-time evaluation of continual learning, in which the stream does not wait for the model to complete training before revealing the next data for predictions. To do this, we evaluate current CL methods with respect to their computational costs. We conduct extensive experiments on CLOC, a large-scale dataset containing 39 million time-stamped images with geolocation labels. We show that a simple baseline outperforms state-of-the-art CL methods under this evaluation, questioning the applicability of existing methods in realistic settings. In addition, we explore various CL components commonly used in the literature, including memory sampling strategies and regularization approaches. We find that all considered methods fail to be competitive against our simple baseline. This surprisingly suggests that the majority of existing CL literature is tailored to a specific class of streams that is not practical. We hope that the evaluation we provide will be the first step towards a paradigm shift to consider the computational cost in the development of online continual learning methods.