B652: Machine Learning
Spring 2007
Miniconference Schedule and Guidelines

B652 will conclude with a "minconference", in which students will present their projects in short conference-format talks and answer questions. This will give the opportunity to share and discuss insights and to gain experience in presenting research in a conference setting.

Two-person presentations should be 25 minutes each, plus 5 minutes for questions. Single-person presentations will be 15 minutes each, plus 5 minutes for questions. The instructor will track time to let presenters know when they need to conclude. Please be sure to plan and practice your talk with a timer so that you'll be able to cover everything and stop promptly (or even slightly early), to preserve the time for questions!

For making the presentations as informative as possible to the class, you'll want to consider points such as:

• What is the research problem? (Even though it's important to state the "big picture" problem, motivating the work with a concrete example can be very helpful.)
• What is the model or approach?
• What are your claims? What is noteworthy about your approach, compared to other alternatives?
• What lessons have you learned so far?
• What do you hope to learn by the end of the semester?

I'm looking forward to your presentations!

• Mon, Apr 16:
  • Spam Filtering with Distributed Case Bases, Jeff Cox and Toshi Uchino
  • Minimal Cost Decision Trees, Scott Dial
  • Interactive Explanation-Based Learning, Josh Bonner
• Wed, Apr 18:
  • A Movie Recommendation System Using Mixed Collaborative and Content-Based Filtering Mechanisms, Tao Huang and Zhuofeng Li
  • A General Adaptation Framework for the WebAdapt System, Jay Powell
  • Paraphrase Recognition through Event Alignment, Jason Kessler
• Mon, Apr 23:
  • Feature Learning in Direct Memory Access Parsing, Craig Michaud
  • Generalizing Case Bases Using Neural Networks, Joey Morwick