Denis McInerney (Northeastern University)

I am a Computer Science PhD student at Northeastern University studying Machine Learning and Natural Language Processing, advised by Professors Jan-Willem van de Meent and Byron Wallace. I am interested in creating interpretable unsupervised or distantly supervised neural models that can be trained with naturally occurring data, especially in domains like healthcare where annotations are difficult to obtain. I have worked on automatic summarization, continuous learning, and representation learning using contrastive and generative objectives. Currently, my focus is on using multimodal models to identify localized relationships between text and images, specifically with radiology images and reports. I want to develop systems that can learn from unlabeled and unstructured data in electronic health records to help doctors perform tasks more efficiently and with less errors.