This project tackles machine comprehension on colloquial writing such as dialog or email. We design models to interpret implicit and explicit contexts such as interpersonal feelings or personal identification.
This project develops core NLP components on the cloud for large scale computing. Our components use advanced techniques in deep learning, and show a great strength in handling big data.
This project explores language structures in all levels. We take linguistic and distributional approaches to parse text into deep dependency graphs conveying practical relations for real applications.
This project challenges to infer diagnostic information from medical reports. We develop an automated pipeline that reads medical reports, makes appropriate diagnosis, and sends the info to the patients.