Project presentations will be from 1pm-4pm in DBH 3011. Note the room and time change!
The alphabetical schedule of presentations is:
1:00 America Holloway & Sidharth Shekar "Augmenting Classifiers with Domain Knowledge"pdf
1:10 Arjun Satish and Ish Rishabh "Human detection in RGB images"pdf
1:20 Carlos Agell "Neural Networks for recognizing pen characters" pdf
1:30 Minas Gjoka and Fabio Soldo "Exploring collaborative filters" pdf
1:40 Goutham Patnaikuni "Human action recognition with LDA"pdf
1:50 Hamed Pirsiavash "Human activity recognition using space-time correlation"pdf
2:00 Jing Zhang and Dan Pan "Web service classification"pdf
2:10 Julian Yarkony "Evidence for the User Interface Theory of Perception"pdf
2:30 Kensuke Ohta "Evaluation of collaborative prediction algorithms"pdf
2:40 Matt Kayala "Classifying DNA microarrays"pdf
2:50 Denis Park and Pornpat Nikamanon "Image segmentation as a classification task"pdf
3:00 Sergi Perez "Training neural networks with genetic algoritms" pdf
3:10 Suman Tatiraju and Avi Mehta "Image segmentation using kmeans vs ncuts" pdf
3:20 Tien Bau "Gabor filters for handwritten digit recognition" pdf
3:30 Tommy Cheng "Video summarization with clustering"pdf
3:40 Yasser Ganjisaffar "Automatic web page classification"pdf
3:50 Yutian Chen "Complex cell models using products of experts"pdf
To re-iterate, the presentation should be 5 minutes max. Assuming 2 minutes for questions and 2 minutes for laptop switching, this brings us to 9-10 minutes per project.
The class project is meant to be a opportunity to explore a specific machine learning concept in depth, or an application-driven project focused on a particular real-world dataset. A typical project might be applying a technique from a research paper on one of the datasets linked to below.
The choice of projects can be very open-ended; ideally you can incorporate their own research. But your class project must be about new things you
have done this semester; you can't use results you have developed in previous semesters. People can work individually or in teams of two.
Project grading Recall the project is worth half the class grade.
A project proposal is due on March 3rd. This is worth 10% of the project grade. The proposal should include a title, a pointer to the dataset, roughly 2 paragraphs describing the project idea, and a list of 1-3 relevant papers.
You must prepare a writeup in the format of a NIPS paper (8 pages maximum in NIPS format, including references; this page limit is strict), due March 21rst at 1pm in the "Final Projects" folder in the dropbox. The writeup is worth 60% of the project grade. A single writeup is required per project.
You will also be responsible for a 5 minute project presentation, also on March 21rst. This will count toward 30% of the project grade.
Again, you are strongly encouraged to find a project aligned with their own research interests. But if you are looking for inspiration, there are numerous similar courses out there, with a plethora of interesting projects!
Research paper references
Good places to look for papers:
These include the major machine learning conferences such as NIPS and ICML. Vision papers that also have a large learning component can be found in ICCV and CVPR. Relevant journals include
Journal of ML Research | Data Mining and Knowledge Discovery | Journal of AI Research | IEEE Pattern Analysis and Machine Intelligence.
Here are a collection of links of various datasets
UC Irvine ML Archive
Caltech Object Recognition Database
Pascal Visual Object Class Challenge
Various (MNIST digits, faces, text)
Face Recognition datasets
Face Detection datasets