Each week, a number of readings will be assigned. You must read all of the papers, as well as the book chapters, and submit reading summaries each class day. Summaries are due every Tuesday and Thursday at 1:59 pm, and should summarize all of the papers you read since the last time you turned in summaries; that is, the papers or book chapters that we will be discussing in class on any given day.
Reading summaries should be submitted in Google Forms at this link.
The summary for each paper will consist of answers to the following questions:
- What research problems were the authors investigating?
- How did the authors attempt to answer these questions? Did they succeed?
- What follow-on possibilities do you see?
If a given book chapter doesn’t fit well with the above questions, the summary for that chapter can alternately consist of answers to the following questions:
- What were the main points being made in the chapter?
- How were those points made?
- Do you agree with the points? How can you expand on or apply them?
Example reading summary
Gaze Enhanced Speech Recognition
1. The authors wanted to find out if they could improve the accuracy of speech recognition by incorporating gaze tracking. By understanding what the user has looked at, they theorize that they can restrict the speech recognizer's language model.
2. The authors ran a study to collect eye gaze data in conjunction with speech-based navigation of an interface, and then they used this information to adapt a speech recognition engine to use they eye gaze data. Their method worked well, achieving a 10% improvement in errors.
3. The authors only considered a static situation, reading a web page. It would be interesting to try to do the same task in a more realistic scenario, by having users interact with objects in the environment by looking at them.
The reading summaries are worth 10% of your final grade. Each will be evaluated based on whether you have provided the requested information, and whether it looks to me like you actually read the paper. I will discard the lowest-graded summary.