We encourage students (a) to choose topics related to their own research area and use deep learning to solve it. For example, cancer detection in medical healthcare, financial prediction, etc. (b) to investigate and develop some new ideas on the theory of deep leaning and/or the applications (computer vision, speech, NLP, etc.) (c) to contribute to the open-source community of the deep learning system (Caffe/Torch/TensorFlow, etc). Below we provide a rough pool of potential topics which we find could possibly fit in the promise stated above.
Note that you are always welcome to propose your own topic/project outside this scope.
There are five primary topic areas.
Within each category, we list some potential projects for you to choose. we try to depict as detailed as possible with resouce links; however, due to limitted time budget, some topics in this pool are not fully described (
it does not mean they are NOT interesting!). You can dig into the details or resort to our fellow TAs for further information.
If you are newkie to the deep learning field and may not be sure where to start, there are some cool resources related to important deep learning applications: