Class Project Information

Teams

The final project will be done in teams of 2-3 people. Please fill in the team assignment form before Apirl 26.

Important Milestones

Project Topic

You may choose any topic you are interested in (such as your research area) as long as it pertains to the topics of this class. You are encouraged to explore advanced machine learning techniques and novel data-driven applications. While we encourage interdisciplinary projects, the main focus should be about applying and understanding learning methods rather than the data collection and hardware design process.

Usually, your project will fall into one of the following categories:

• Application: use machine learning to solve a specific problem, e.g.

  • AI applications such as computer vision, NLP, speech recognition, etc

  • Information processing for physical systems, such as energy, material and optics

• Algorithm: develop a machine learning method that has better performance, e.g.

  • Improve the robustness of an algorithm under noisy data

  • Make a machine learning method more efficient

• Theory: theoretical problems, e.g.

  • Theoretical analysis of a specific learning algorithm regarding its generalization, consistency, or stability properties

  • Use existing theoretical concepts to create efficient learning algorithms

A good project will showcase your ability to clearly define a problem (with practical or theoretical importance), demonstate the motivation and technical details of your method, and conduct meaningful evaluation and analysis.

Grading

The final project will count for 25% of your final grade. It will be graded on a scale of 100 points, which include:

• Proposal: 15 points

• Discussion with class staff: 5 points

• Final poster: 25 points

• Poster session participation: 5 points

• Final report grade: 50 points

Sample Posters, Projects from Previous Years

Poster

• Remaining useful life prediction for Lithium-ion batteries across operation conditions by CORAL, Shengyu Tao, Huazhang Ying, Tingwei Cao (application) 2022 Best Poster Award

• Dancing In the Dark, Chengwei Ren , Jinnan He, Yuzhu Zhang (application) 2022 Best Presenter Award

• AI Coach for divers, Lekang Yuan and Xiaohang Yu (application)

• Face Rectification on Feature Maps for Recognition, Peng Lu (algorithm)

Report

• Deep Hypergraph Convolutional Network for Paper Classification, Yifei Zhu and Guanzi Chen (algorithm & theory)

• An Application of Machine Learning on Predicting CO2 Reduction Electrocatalysts, Chen Liang and Yuhang Zhang (application)

• Live Gradient Compensation for Evading Stragglers in Distributed Learning Jian Xu (algorithm) published

• Normalized LMS for Adaptive Graph Signal Estimation on Genetic Data Yi Yan (algorithm) published

• Debugging Neural Networks, Riccardo Mattesini, Sebastian Beetschen and Bunchalit Eua-arporn (algorithm)

Common Pitfalls

• Insufficient contribution (e.g. Simply run the code of existing models on some benchmark datasets): Try to come up with at least one novel contribution in any of the three catogries discussed in the Topics section. Or consider a more challenging problem scenario.

• Topic is too broad or too ambitious: reduce the project scope

• Project relies heavily on external factors (e.g. data collection, hardware equipment) : be sure to have your data ready at the start of the project. Have backup plan (e.g. alternative datasets/tasks) when external requirements can not be met by the time you submit your proposal.