Program for the CSAIL Student Workshop, 2009
| 8:00 AM | Breakfast outside Stata |
| 8:30 AM | Buses leave Stata
|
| 9:15 AM | Registration and Snacks |
| 9:45 AM | Opening Remarks |
| 10:00 AM | Keynote Speech by Prof. Randall Davis |
| 10:45 AM | Session 1 - Learning and Robotics |
| 12:00 NN | Lunch
|
| 1:30 PM | Session 2 - Human Computer Interaction |
| 3:00 PM | Invited Talk by Mike Oltmans (ITA representative)
|
| 3:30 PM | Tea break |
| 3:45 PM | Session 3 - Theory
|
| 4:45 PM | Invited Talk by Eric Prud'hommeaux (W3C)
|
| 5:15 PM | Session 4 - Systems |
| 6:15 PM | Dinner |
| 7:45 PM | Keynote Speech by Prof. Rob Miller |
| 8:30 PM | Award Ceremony and Closing Remarks |
| 8:45 PM | Return to Cambridge |
| 9:30 PM | Buses arrive at Stata
|
Session 1 - Learning and Robotics
Chair: Harr Chen
Talk: "Toward Super maneuverable Flying Robots" by Rick Cory
Talk: "Learning to Reformulate Long Queries" by Neha Gupta
Demo: "Multiple Relative Pose Graphs for Robust Cooperative Mapping" by Been Kim
Session 2 - Human Computer Interaction
Chair: Max Van Kleek
Talk: "Sketch vs. Drag-and-Drop Input on a Tabletop Display" by Yale Song
Talk: "The Study of Content Transformation for Web Page Readability Enhancement" by Chen-Hsiang Yu
Demo: "SQL: That And a Pair of Brackets Will Buy You an XML Result of Your Choice" by Eirik Bakke
Demo: "AINS: Asynchronous Interaction System for Preschoolers" by Chen-Hsiang Yu
Session 3 - Theory
Chair: Aleksander Madry
Talk: "Universal Semantic Communication: an informal overview" by Brendan Juba
Talk: "Lower Bounds for Sparse Recovery" by Eric Price
Session 4 - Systems
Chair: Harshad Kasture
Talk: "Region Helper Locks for Fork-Join Parallel Languages" by Jim Sukha
Talk: "Secure Comparison Shopping using Trusted Execution Modules" by Victor Costan
Invited Talk by Mike Oltmans
"Needle: A structured data acquisition, integration and querying platform"
Needle is ITA's platform for creating and integrating databases from a wide variety of structured and semi-structured sources. In addition to standard imports of CSV and XML data, Needle provides a data tagging interface that users can employ to train a web scraper to extract data from semi-structured web pages. The web scraper uses the Collins perceptron algorithm to learn which data elements to extract from each page, and a grammar transformation algorithm to extract the relationships between those data elements. Data from each source is mapped to a user defined schema and stored in a custom graph-structured database. The platform performs sophisticated parsing, de-duplication, and data cleaning when importing the individual datasets. Finally, Needle's data browsing tools and custom query language allow users to easily view their data and construct powerful queries. From the end user perspective, all of this can be done without writing any code, which opens up the platform to a broad range of users.
This talk will include a live demo and presentation by Mike Oltmans (PhD from AI Lab/CSAIL 1998-2007) who has been working on the Needle team since joining ITA two years ago.
Invited Talk by Eric Prud'hommeaux (W3C)
"Introduction/Invitation to W3C’s Semantic Web Work"
On the 5th floor of the Gates tower is a cluster of web designers called W3C. Atypically, they don't design sites, but the web itself. Many are currently obsessed with designing a new, machine-useful aspect of the web called the Semantic Web. The Semantic Web is particularly well-suited to representing complex webs of information. In this slot, we will present the application of this simple, unifying data format to fine-grained representation of biological properties and interaction, through to macroscopic events like hospital visits or drug reactions. What innovation does this represent? Nothing huge, but instead the thousands of little innovations required to revolutionize the world. Scientists are sharing information and ideas in collaborative environments like Alzforum, clinics are coding rules for diagnostic and monitoring assistance, and all are sharing coding systems allowing pharmaceuticals to tailor medications to individual patients' needs. The skills learned in creating a fabric of life sciences information are teaching and motivating other industries to use the Semantic Web. If you've ever used cut and paste, or have a stack of post-its next to your keyboard, come see how your lab can participate in a global information web.
