Description
Sources
Most of these papers will also be posted on arXiv.org and it’s fine to select a paper from arXiv, but since any crap can be posted on arXiv, please select only papers that have been peer reviewed and published, preferably in one of the proceedings above.
The only restriction on the paper is that it should address neural network models. Ideally, you will find a paper that is on a topic of interest to you and that might lead to a final project topic or a tie-in to your main research interest. The paper can be on an application of neural nets or a more theoretical contribution. I would like you to pick some work we have not yet talked about in class, though it’s fine if you cover articles I have on the syllabus for the last month of class.
I expect that half of your effort on this assignment will be on finding a ‘just right’ paper. Select wisely, and have fun looking through articles.
Commentary
The remainder of your commentary could include such reactions as:
- questions you have about the paper, either clarification questions or points of disagreement with the authors (“I don’t see how such and such will work as the author claims…”).
- comments on how the paper relates to other work you’ve read;
- a critique of the work. What are the flaws in the ideas presented? What are the limitations? Did the speaker place their work in the appropriate theoretical context? Did the authors overstate their results?
- If the paper gave you inspiration for new research ideas, describe these ideas. The ideas might be extensions of the work, new directions to move in, or applications of the work to other subfields.

