Artificial Intelligence
Building intelligent machines that complement human reasoning and extend what people can do.
Projet
Optimization in Deep Learning
Understand the dynamics of the optimizers in deep learning and its convergence rate and implicit regularization. One important aspect to open the black-box of deep neural networks is to understand the dynamics of the optimization…
Projet
Neural Machine Translation
We are working on neural machine translation, using deep neural networks for machine translation. We achieved human parity (opens in new tab) in translating news from Chinese to English. Xueqing Wu, Yingce Xia, Jinhua Zhu, Lijun Wu,…
Projet
Ab Initio Aperiodic Molecular Dynamics
Molecular dynamics is a task for understanding and predicting physicochemical property of real-world substances from the fundamental rule of physics. It provides solution from the first principle for various impactful problems, including developing new materials…
Projet
Microbiome-based Diagnostics and Therapeutics
Human microbiota are ecological communities of microbes colonized on and inside human bodies. The diverse microbes shape our body functions and are associated with a vast majority of human diseases. Many efforts have been made…
Projet
Low Carbon Transformation Pathway
In order to reduce the risks and impacts of climate change, global efforts to keep the average temperature rise within 1.5 degrees Celsius are of great significance to all mankind. In this context, many countries…
Projet
Fast Neural PDE Solver
Understanding many sustainability issues relates fundamentally to PDEs, from macroscope to microscope, such as navier stokes equation and Schrödinger equations. Solving these PDEs enables us to understand and forecast the world and is a critical…
Projet
Learning to Decode the Immune System to Diagnose Disease
Adaptive published the original proof-of-concept for this project (Emerson, et al. Nat Genet. 2017), demonstrating that shared TCRs can be used as a basis for diagnosing viral infection. Since then, we have further demonstrated the…
Projet
Learn the Dynamics
The laws of nature are described as dynamical systems. Learning dynamics, as a new AI technique, faces the following challenges: (1) the existing pure deep learning models are not that reliable in the scientific field,…
Projet
Language-speech Pretraining
We are working on pre-trained language model, including new pre-training method, pre-trained model compression, pre-training for other tasks such as speech and music. Mingliang Zeng, Xu Tan, Rui Wang, Zeqian Ju, Tao Qin, Tie-Yan Liu,…
Projet
Graphormer
Graphormer is a deep learning package that allows researchers and developers to train custom models for molecule modeling tasks. It aims to accelerate the research and application in AI for molecule science, such as material…