Reaction of Siebel Scholars

Description of Siebel Scholoars' Research

Siebel Systems Scholarships Awarded to Five SCS Graduate Students

Siebel Systems, Inc. announced a $2.5 million gift to establish five fellowships at the Carnegie Mellon School of Computer Science (SCS). The awardees or Siebel Scholars as they will be know will receive $25,000 each for educational expenditures.

The five chosen graduate students who demonstrate academic excellence and leadership qualities are:

• Frank Dellaert, CS Ph.D.
  Advisors: Chuck Thorpe and Sebastian Thrun
• Rayid Ghani, CALD MS
  Advisor: Tom Mitchell
• Craig Olinsky, LTI MS
  Advisor: Alan Black
• Jovan Popovic, CS Ph.D.
  Advisors: Steve Seitz and Mike Erdmann
• Jeffrey Smith, Robotics Ph.D.
  Advisor: Irving Oppenheim

Reaction from the new Siebel Scholars was a mixture of surprise, gratitude, and relief. Frank was grateful the Siebel Systems scholarship would now cover the cost of graduate school. "After having been generously supported by SCS over all these years, it feels good to be able to contribute part of it myself by means of this grant," said Dellaert. Normally, the School of Computer Science awards most students accepted into a doctoral program a Graduate Fellowship for the academic year, covering full tuition and providing a living allowance that is usually renewable for the duration of the program. Similarly, Jeff Smith's response was one of relief. "My advisor has no funded research," said Smith, "so I've been living on the generosity of better-funded professors for the past two years."

Rayid Ghani was totally surprised upon hearing he had received the $25,000 Siebel Systems scholarship. "It was completely unexpected and I feel extremely honored to have been chosen to receive this award," exclaimed Ghani. Craig Olinsky responded to the news with a renewed dedication to his research. "I suppose that it would, in a sense, encourage me to work harder knowing that there has now been set this external influence with such lofty expections of my overall research output," explained Olinsky

Olinsky gives an accurate read on Siebel System's and other's expectations. Siebel donated the gift to recognize the excellent reputation Carnegie Mellon has in the field of computer science. "Siebel Systems wanted to invest money in supporting the leaders of tomorrow's technology companies," said Justin Dooley, Siebel System's Vice President of Finance. "By providing graduate fellowships at schools like Carnegie Mellon's School of Computer Science we are able to do just that."

Living up to the School's reputation, the Siebel Scholars' research focuses on area of computer science that will undoubtably impact technology, from text and web mining, speech synthesizer technology, computer vision, and computer animation.

Rayid Ghani, a master's student in the Computer Automated Learning and Discovery (CALD) program, is investigating text and web mining. Given that Lycos technology was born at Carnegie Mellon, Ghant is following a strong tradition. Specifically, he works on applying Machine Learning techniques to automatic text classification and information extraction on the web as well as using unlabeled data to improve text classifiers. After graduation, he is considering a career at a corporate research lab where he could apply his machine learning skills to real-world problems.

Craig Olinsky, a master's student in the Language Technology program, is working on training an intontation model for a Chinese speech synthesizer (text-to-speech system) for his master's thesis. After graduation, he hopes to to pursue a Ph.D and then to find a corporate research position involing east asian language processing (working in Speech synthesis, recognition, text analyais, or such...)

Frank Dellaert, a doctoral student in Computer Science, conducts research that concerns the reconstruction of three dimensional objects and environments from a set of pictures or video images. "I believe computer vision will become an increasingly visible application for current-day multimedia compute rs, and there is a lot of interest in the industry to gain an edge in this domain," says Dellaert.His research uses the language of probablities and tools for probabilistic inference to solve for the 3D reconstruction and the correspondences simultaneously. This greatly extends the range of situations for which one could obtain 3D reconstructions, without having to resort to manual intervention, as is commonly done now. After graduation, Dellaert plans to look at faculty positions at research universities and industry research labs that have a strong computer vision group or are in the process of building one.

Jovan Popovic, also a doctoral student in Computer Science, is working to develop and improve computer animation techniques to make animation an accessible form of expression for everyone. For his thesis, Popovic is developing a method that provides a direct, intuitive control of physical simulations. Instead of having to specify the physical parameters, the animator can specify how the motion should look, such as sketching out the desired look, and the system automatically computes a realistic motion that has the desired look.

Like Popovic, Jeff Smith, a Ph.D student in Robotics, is also focusing on physically-based computer animation. Specifically, he is studying human and animal motion synthesis, such as generating walking, running and other kinds of motion for arbitrary, animator-specified, creatures (which may or may not be actual animals). After graduating; Smith is planning on applying for post-docs, both here in the US and in other countries, in the areas of computer animation and electronic/new-media art.