Intelligent medical diagnosis system: With the continuous development of machine learning, algorithms are becoming more widely used in the medical field. We build a multimodal-big-data based intelligent medical diagnosis system with the computation supported by Tianhe-2 supercomputer. This system can intelligently identify the patient's medical pictures and make appropriate diagnosis.
Model-Lite Planning: Previous planning systems require complete domain models as input, which is often difficult in real-world applicaitons. A more realistic assumption is that, provided limited domain knowledge (e.g., background knowledge, represented by incomplete domain models for example) and a set of historical plan cases, we can efficiently generate robust plan solutions to new planning problems. In this project, we build a model-lite planning system that can compute plans for new planning problems with incomplete domain models.
Action Model Learning: Creating domain models for planning is time-consuming and tedious, even for domain experts. In this project we aim to automatically learn domain models (or action models) from historical plan traces.
Question Answering: We propose to build deep networks to extract answers to questions. We build a demonstration based on insurance domains.
From Images to Poems:

We aim to generate poems based on images taken by users.

PlanTool: PlanTool is a tool for efficient application and research in planning area, which integrating various popular planners. For now there are a GUI for planning application and a planning package with a lot of planners.
Intelligent Manufacturing: We built an intelligent manufacturing platform to do domain knowledge engineering and planning.