Links

Of Interest

UCLA Department of Statistics Speakers Series

3:00 PM-4:00 PM, Tuesday, February 9, 2010
6229 Math Sciences Bldg.

Laurent Itti

Computer Science
University of Southern California

Exploiting Bottom-up and Top-down Information in a Bayesian Framework for Vision

Humans and many other animals use a combination of stimulus-driven (bottom-up) and task- or goal-driven (top-down) signals to shape behavior, including how to deploy visual attention, guide gaze over complex visual scenes, recognize objects, and build cognitive representations of scenes. In this talk, I will focus on describing a biologically-inspired framework for perception, developed by Lior Elazary in my lab, which provides a generic architecture by which bottom-up and top-down information can be combined. The core is a hierarchical Bayesian inference system. Hypotheses about objects in a visual scene are generated "bottom-up" from sensor data. These hypotheses are refined and validated "top-down" when complex objects, hypothesized at higher levels, impose new feature and location priors on the component parts of these objects at lower levels. To efficiently implement the framework, an important new contribution is to systematically utilize the concept of bottom-up saliency maps to narrow down the space of hypotheses. In addition, we let the system hallucinate top-down (manufacture its own data) at low levels given high-level hypotheses, to overcome missing data, ambiguities and noise. The implemented system is tested against images of real scenes containing simple 2D objects against various backgrounds. The system correctly recognizes the objects in 98.71% of 621 video frames, as compared to SIFT which achieves 38.00%.

Bio: Dr. Laurent Itti received his M.S. in Electrical Engineering with a specialization in Image Processing from the Ecole Nationale Superieure des Telecommunications, Paris, France, in 1994. He received his Ph.D. in Computation and Neural Systems from the California Institute of Technology, Pasadena, California, in 2000. He has since then been an Assistant (2000-2006) and Associate (2006-present) professor of Computer Science and voting faculty member of the cross-disciplinary Neuroscience Graduate Program at the University of Southern California (USC), Los Angeles, California. Dr. Itti has authored over 100 peer-reviewed publications in journals, books, and top-ranked conferences. Dr. Itti teaches Artificial Intelligence, Brain Theory and Neural Networks, Introduction to Robotics, Visual Processing, Neuroscience Core Course, Neural Basis for Visually Guided Behavior, and Computational Architectures in Biological Vision. Dr. Itti's laboratory comprises 15 students, postdocs and engineers, and is recipient of grants by the National Science Foundation, DARPA, the National Geospatial-Intelligence Agency, the Human Frontier Science Program (HFSP), the Office of Naval Research, the Army Research Office, and the National Institutes of Health. Dr. Itti has been distinguished through a number of awards, including the 2008 Okawa Foundation Research Award, being one of the 16 nationally selected speakers at the 2007 National Academy of Engineering's Frontiers of Engineering Symposium, and serving on Program Committees for several conferences by IEEE.

UCLA Statistics' Blended Instruction Cited by 2009 WASC Review

The instructional work of Dr. Mahtash Esfandiari incorporating a blend of in-class instruction and Moodle’s Quiz Tool has been cited in UCLA's 2009 WASC[1] (Western Association of Schools and Colleges) reaccreditation review.

Dr. Esfandiari's work is cited as part of UCLA's third accreditation phase in the report "Educational Effectiveness Review"[3]. The history, description, and results of Dr. Esfandiari's experiment are given in essay D of the report[4]. Some of the highlights of the essay:

  • Winter 2005, Dr. Mahtash Esfandiari (Senior Lecturer in Statistics) extensively redesigned Statistics 10 by using online quizzes, weekly labs, and homework to maximize students’ roles as active learners, and minimize their roles as passive recipients of information.
  • Moodle’s Quiz Tool function allowed Dr. Esfandiari to develop an automated test bank of nearly 1,500 multiple-choice statistics questions that engage students’ higher order thinking skills, including application, analysis, synthesis, and evaluation.
  • The first quiz is administered via Moodle the day after lecture, prior to the week’s discussion section in order to measure student comprehension of lecture material. Teaching assistants have the ability to monitor student progress online, allowing future instruction and discussion to be tailored to the strengths and weakness of students’ quiz results. They can also assess which students are progressing similarly and create compatible small groups for in-class discussions. This allows the teaching assistant to focus on the groups’ needs more efficiently and encourages students to discuss their misconceptions as a group, thus further developing their knowledge through peer collaboration.
  • After students attend the week’s discussion section, the second quiz is administered via Moodle. This quiz addresses the same concepts as the first quiz by using similar questions from the test bank, which are easily identifiable through the test bank’s search function.
  • An experimental study was designed to investigate the educational effectiveness of blending standard in-class teaching methods with Moodle’s Quiz Tool when teaching introductory statistics to a large group of students (100 or more). For the purpose of this study, educational effectiveness was defined by a student’s ability to apply statistical principles to solve or interpret real world questions versus simply mastering statistical formulas. This included the ability for higher order thinking such as application, analysis, synthesis, and evaluation.
  • At the end of the term, students from both the experimental and control groups were surveyed about their impressions of how, and why, they developed understanding throughout the course. Attendance, homework, student interactions, active learning, memorization, knowledge application, and critical thinking elements were all addressed. Students also completed a final examination to assess learning outcomes that ranged from solving mathematical equations to evaluating real world cases.
  • Overall, the findings supported Dr. Esfandiari’s premise that blended instruction would foster student reflection and self-generated learning and lead to higher order thinking. For example, students from the experimental and control groups performed equally well when asked to respond to open-ended questions that related to hypothesis testing calculations. However, experimental group students performed much better than control group students on open-ended questions that related to the Central Limit Theorem, which involved analysis and evaluation. When control group students responded to these questions they were able to use the correct statistical terminology to describe the problem, but they were unable to elaborate on what the terminology meant.

About WASC

The Western Association of Schools and Colleges (WASC) is one of the six regional accrediting associations in the United States. WASC was formed on July 1, 1962, to evaluate and accredit schools, colleges, and universities in California, Hawaii, the territories of Guam, American Samoa, Federated States of Micronesia, Republic of Palau, and the Commonwealth of the Northern Mariana Islands. WASC functions through a board of directors and three accrediting commissions: the Accrediting Commission for Senior Colleges and Universities, the Accrediting Commission for Community and Junior Colleges, and the Accrediting Commission for Schools. The board of directors consists of nine members, with each accrediting commission electing three members.

Each commission, with the involvement of all participating institutions, develops its own standards, proce- dures, and fiscal policies, under the authority and subject to the approval of the WASC board of directors. The accreditation actions of each commission are certified annually by the board of directors of WASC. Accredi- tation ceases whenever an institution fails to pay its annual fees, requests in writing that its accreditation be withdrawn, or when the Commission formally acts to terminate accreditation.

About WASC Accreditation

"UCLA aims to use the reaccreditation process to highlight and enhance ongoing efforts to improve our educational effectiveness. An effective process requires a high level of cooperation and collaboration among students, faculty, and administrators. Together, we engage in reflection and self-study, and then apply the insights gained to refining our goals and strengthening our programs. A variety of qualitative and quantitative data inform these endeavors."[2]

The Western Association of Schools and Colleges (WASC) is the regional organization that provides umbrella accreditation for UCLA as a whole, through its Accrediting Commission for Senior Colleges and Universities. This site displays the plans, processes, data, analyses and reports associated with UCLA’s reaccreditation by WASC.

To become or remain accredited, institutions must meet WASC Standards and satisfy WASC Criteria for Review. Accreditation confers a variety of benefits, especially the right to administer federal financial aid. The accreditation process is described in detail in WASC’s Handbook of Accreditation. The Policies Manual provides additional information.

[1] http://www.wascsenior.org/
[2] http://www.wasc.ucla.edu/approach.shtml
[3] http://www.wasc.ucla.edu/2009effectiveness.shtml
[4] http://www.wasc.ucla.edu/EER_EssayD.pdf


Date Posted: 2010-01-27, 05:41:00


Part-time Lecturer Position Available for 2009-2010 Academic Year

Status: Open

The UCLA Department of Statistics seeks a temporary part-time lecturer to teach statistics during the spring quarter in the 2009-10 academic year. Ideal candidates will have a Ph.D. in Statistics, demonstrated excellence in teaching, knowledge of and experience with R, experience in applied statistics, and a commitment to excellent undergraduate teaching. If interested, please send a CV and cover letter to:

Robert Gould
Vice-Chair of Undergraduate Studies
UCLA Department of Statistics
8125 Math Sciences Building
Box 951554
Los Angeles, CA 90095-1554
rgould@stat.ucla.edu

In cover letter, please address teaching experience and provide evidence of teaching excellence. Women and underrepresented minorities are encouraged to apply. Full consideration is guaranteed to those who apply by November 13, 2009, however, the search will remain open until all positions are filled.

The University of California Los Angeles is an Equal Opportunity/Affirmative Action Employer.

Contact: Rob Gould

Posted: 2009-11-02, 11:17:00 by Jose Hales-Garcia

New Online Calendars

The department has updated its online workgroup calendaring system. You can see the calendars being served at http://calendars.stat.ucla.edu. Two calendars are presently listed:


Date Posted: 2010-01-28, 11:33:00


New. Improved. Support Pages

The technical (formerly known as Support) and academic knowledge-bases of the department have been combined into a single wiki. Answers[1] is the one-stop location for finding answers to departmental questions.

Articles include: Taking the Oral Exam, Computation and Data Set Management Guidelines, For the Web: Start Here.

You can explore these and other articles at Answers[1].

[1] http://answers.stat.ucla.edu


Date Posted: 2009-12-11, 07:44:00


Welcome

Why Statistics?

The world is becoming more and more quantitative. Many professions depend on numerical measurements to make decisions in the face of uncertainty. Statisticians use quantitative abilities, statistical knowledge, and communication skills to work on many challenging problems.

Want to Learn More?

The Department of Statistics at UCLA coordinates undergraduate and graduate statistics teaching and research within the College of Letters and Sciences. We teach a large number of undergraduates and we have a substantial graduate program. Our research and teaching have a strong emphasis on computational and applied statistics.  We have an active consulting center for both on-campus and off-campus clients.

You can find the Department’s catalog on-line. We offer nearly one hundred courses; an undergraduate B.S. degree and Minor; and M.S. and Ph.D. degrees in Statistics.

Want to Help?

Donations and contributions of any amount to the Department of Statistics are greatly appreciated. Generous giving from the community, alumni, and parents helps the department to maintain high standards of computational technology and continue with its rapid growth. Thank you for your consideration. For more information please visit our webpage for online giving.

News & Events

« Exploiting Bottom-up and Top-down Information in a Bayesian Framework for Vision[Seminar: Tue, Feb 9]

« UCLA Statistics' Blended Instruction Cited by 2009 WASC Review

« Part-time Lecturer Position Available for 2009-2010 Academic Year

« New Online Calendars

« New. Improved. Support Pages

« Welcome


UCLA Department of Statistics
8125 Math Sciences Bldg.
Box 951554
Los Angeles, CA
90095-1554
310-825-8430
Fax: 310-206-5658