Imagine a radiologist examining a chest X-ray from a new patient. She notices the patient has swelling in the tissue but does not have an enlarged heart. Looking to speed up diagnosis, she might use a vision-language machine-learning model to search for reports from similar patients.
But if the model mistakenly identifies reports with both conditions, the most likely diagnosis could be quite different: If a patient has tissue swelling and an enlarged heart, the condition is very likely to be cardiac related, but with no enlarged heart there could be several underlying causes.
In a new study, MIT researchers have found that vision-language models are extremely likely to make such a mistake in real-world situations because they don't understand negation - words like "no" and "doesn't" that specify what is false or absent.
"Those negation words can have a very significant impact, and if we are just using these models blindly, we may run into catastrophic consequences," says Kumail Alhamoud, an MIT graduate student and lead author of this study .
The researchers tested the ability of vision-language models to identify negation in image captions. The models often performed as well as a random guess. Building on those findings, the team created a dataset of images with corresponding captions that include negation words describing missing objects.
They show that retraining a vision-language model with this dataset leads to performance improvements when a model is asked to retrieve images that do not contain certain objects. It also boosts accuracy on multiple choice question answering with negated captions.
But the researchers caution that more work is needed to address the root causes of this problem. They hope their research alerts potential users to a previously unnoticed shortcoming that could have serious implications in high-stakes settings where these models are currently being used, from determining which patients receive certain treatments to identifying product defects in manufacturing plants.
"This is a technical paper, but there are bigger issues to consider. If something as fundamental as negation is broken, we shouldn't be using large vision/language models in many of the ways we are using them now - without intensive evaluation," says senior author Marzyeh Ghassemi, an associate professor in the Department of Electrical Engineering and Computer Science (EECS) and a member of the Institute of Medical Engineering Sciences and the Laboratory for Information and Decision Systems.
Ghassemi and Alhamoud are joined on the paper by Shaden Alshammari, an MIT graduate student; Yonglong Tian of OpenAI; Guohao Li, a former postdoc at Oxford University; Philip H.S. Torr, a professor at Oxford; and Yoon Kim, an assistant professor of EECS and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT. The research will be presented at Conference on Computer Vision and Pattern Recognition.
Neglecting negation
Vision-language models (VLM) are trained using huge collections of images and corresponding captions, which they learn to encode as sets of numbers, called vector representations. The models use these vectors to distinguish between different images.
