Thursday, January 26, 2006

Human-Robot Interaction Conference 2006

The advanced program is available. Check out what the state-of-the-art research of this field is.

Wednesday, January 25, 2006

CMU talk: Game theory, biology, and the binding game

Time: 3:30pm, Tuesday Jan. 24
Location: Wean 5409
Speaker: Tommi Jaakkola, MIT CSAIL

Biological processes span across vastly different scales and necessarily have to be understood at multiple levels of abstraction. Towards clarifying the role that computation plays in such understanding, we have recently developed a class of game theoretic models for capturing coordinate operation of DNA binding regulators. Our work builds in part on the argument that the roles of various molecular interactions cannot be understood in isolation but that it is necessary to also capture the context provided by other mutually constraining processes. Our game theoretic model allocates proteins to neighborhoods of sites, and to sites themselves, in a resource constrained manner, while explicitly capturing coordinate and competitive relations among proteins with affinity to the site or region. We provide examples of known biological subsystems that are naturally translated into our framework, and illustrate predictions that can be derived from the model. The focus of the talk will be on mathematical foundations of the modeling approach and requires little or no biological background. This is joint work with Luis Perez-Breva, Luis Ortiz, and Chen-Hsiang Yeang.

Speaker Bio:
Tommi S. Jaakkola received the M.Sc. degree in theoretical physics from Helsinki University of Technology, Finland, and Ph.D. from MIT in computational neuroscience. Following a postdoctoral position in computational molecular biology (DOE/Sloan fellow, UCSC) he joined the MIT EECS faculty 1998. His research interests include many aspects of machine learning, statistical inference and estimation in the context of graphical models, and analysis and development of algorithms for various modern estimation problems such as those involving multiple predominantly incomplete data sources. His applied research focuses on problems in computational biology such as transcriptional regulation.

Hi ! Everybody

I'm the new lab member, Stanley.
I'm now trying to be familiar with using Blog.
My families and I wiil go back to I-Lan tomorrow.
Happy (Chinese) new year to all of you.

My email/MSN:

Tuesday, January 24, 2006

Design Once for Both FPGA & Structured ASIC


Only Altera allows you to develop your high-density logic design using Stratix II FPGAs and then migrate to a HardCopy II structured ASIC without any need for redesign or additional timing closure efforts. Learn how Stratix II FPGAs and HardCopy II structured ASICs together provide a unique synergy from design to production.

My Talk (2005/01/24)

Detection and Tracking of Moving Objects from a Moving Platform in Presence of Strong Parallax

Abstract :
We present a novel approach to detect and track independently moving regions in a 3D scene observed by a moving camera in the presence of strong parallax. Detected moving pixels are classified into independently moving regions or parallax regions by analyzing two geometric constraints: the commonly used epipolar constraint, and the structure consistency constraint. The second constraint is implemented within a “Plane+Parallax” framework and represented by a bilinear relationship which relates the image points to their relative depths. This newly derived relationship
is related to trilinear tensor, but can be enforced into more than three frames. It does not assume a constant reference plane in the scene and therefore eliminates
the need for manual selection of reference plane. Then, a robust parallax filtering scheme is proposed to accumulate the geometric constraint errors within a sliding window and estimate a likelihood map for pixel classification. The likelihood
map is integrated into our tracking framework based on the spatio-temporal Joint Probability Data Association Filter (JPDAF). This tracking approach infers the trajectory and bounding box of the moving objects by searching the optimal path with maximum joint probability within a fixed size of buffer. We demonstrate the performance of the proposed approach on real video sequences where parallax effects are significant.


Saturday, January 21, 2006

CMU talk: Representations and Algorithms for Monitoring Dynamic Systems

Time: 3:30pm Tue., Jan. 17
Location: Wean Hall 5409.
Speaker: Avi Pfeffer, Associate Professor at Computer Science, Harvard University

Check the link for details about CMU AI Seminar.
Check the link for details about the speaker.

Continually monitoring the state of a dynamic system is an important problem for artificial intelligence. Dynamic Bayesian networks (DBNs) provide for compact representation of probabilistic dynamic models. However the monitoring task is extremely difficult even for well-factored DBNs. Therefore approximate monitoring algorithms are needed. One family of approximate monitoring algorithms is based on the idea of factoring the joint distribution over the state of the system into a product of distributions over factors consisting of subsets of variables. Factoring relies on the notion of weak interaction between subsystems. We identify a new notion of weak interaction called separability, and show that it leads to the property that, in order to compute the factor distributions at one point in time, only the factored distributions at the previous time point are needed. We also define an approximate form of separability. We show that separability and approximate separability lead to very good approximations for the monitoring task. Unfortunately, sometimes the factoring approach is computationally infeasible. An alternative approach to approximate monitoring is particle filtering (PF), in which the joint distribution over the state of the system is approximated by a set of samples, or particles. In high dimensional spaces, the variance of PF is high and too many particles are required to provide good performance. We improve the performance of PF by introducing factoring, maintaining particles over factors instead of the global state space. This has the effect of reducing the variance of PF and so reducing its error. Maintaining factored particles also allows us to improve PF by looking ahead to future evidence before deciding which particles to propagate, thus leading to much better accuracy.

Speaker bio:
Avi Pfeffer is Associate Professor at Computer Science at Harvard University. His research is directed towards achieving rational behavior in intelligent systems, based on the principles of probability theory, decision theory, Bayesian learning and game theory. He received his PhD in 2000 from Stanford University, where his dissertation on probabilistic reasoning received the Arthur Samuel Thesis Award. Dr Pfeffer has published technical papers on probabilistic reasoning, strategic reasoning, agent modeling, temporal reasoning, and database systems. He was awarded the NSF Career Award in 2001 for work on strategic reasoning, and the Alfred P. Sloan Foundation Research Fellowship in 2002. Dr Pfeffer serves on the editorial board of the Journal of Artificial Intelligence Research, and on the program committees of a number of leading conferences in artificial intelligence.

Wednesday, January 18, 2006

The Information-Form Data Association Filter

The Information-Form Data Association Filter

Brad Schumitch, Sebastian Thrun, Gary Bradski, and Kunle Olukotun

This paper presents a filter for online data association problems in high-dimensional spaces. The key innovation is a representation of the data association posterior in information form, in which the "proximity'' of objects and tracks are expressed by a numerical links. Updating these links requires linear time, compared to exponential time required for computing posterior probabilities. The paper derives the algorithm formally, and provides comparative results for using data obtained by real-world camera array and by a large-scale sensor network simulation.

The full paper is available in PDF

Tuesday, January 17, 2006

News: A genuine milestone for artificial intelligence

January 16, 2006
Richard Macey, the link

ROBOTS have a reason to party: this year is the 50th anniversary of artificial intelligence.

In 1956 John McCarthy, a scientist at the Massachusetts Institute of Technology, convened a meeting of computer specialists at Dartmouth College, New Hampshire.

"It was the dawn of the computing era," said Claude Sammut, professor of computer science and engineering at the University of NSW and leader of its Artificial Intelligence Research Group.

Professor McCarthy's meeting "brought together the small number of people who were writing AI programs. He had to invent some new name for what they were doing, so he called it artificial intelligence". The name stuck.

"When most people think of artificial intelligence they think of robots like C3PO in Stars Wars," Professor Sammut said. But intelligent robots did not always adopt humanoid shapes.

He noted a Brisbane container terminal had been fully automated, with cranes programmed to locate and collect containers. "The cranes are basically robots. They can operate without a driver and know which containers have to be taken off which ships.

"And there's a company in Sydney that makes programs that help pathologists interpret blood tests. The computer generates a detailed report. The pathologist is still there, checking, but it speeds up the pathologist's job."

Sunday, January 15, 2006

Emergency Response/Robotics Joint Topical Meeting Feb. 11-16 in Salt Lake City

The American Nuclear Society, led by the Idaho Section, is sponsoring a forum in Salt Lake City Feb. 11-16 on emergency response preparations and robotics. This topical will discuss solutions to challenges that often cut across boundaries, applications, circumstances, markets and technologies. The Environmental Sciences Division 9th Topical Meeting and the Robotics and Remote Systems Division 11th Topical Meeting have joined together around the theme "Sharing Solutions for Emergencies and Hazardous Environments." This is the first time two such ANS topical meetings have been held jointly. You can learn more about this forum and register at

Papers, panels, workshops, exhibits and demonstrations will offer research and practical field topics that will provide appeal and will promote provocative and beneficial interactions for most robotics professionals. Featured speakers will include Admiral Joseph Krol, associate administrator of emergency operations at the National Nuclear Security Administration, Ken Brockman of the International Atomic Energy Agency, Dr. Harold McFarlane of Idaho National Laboratory (INL) (who is president-elect of the American Nuclear Society) and Dr. Harold Blackman of INL. For more information, visit the above Web site or contact co-chairs Eric Loewen (208-526-9404, or Ron Lujan (208-526-4045,

Monday, January 09, 2006


Daniel Gatica-Perez, Iain McCowan, Dong Zhang, and Samy Bengio
ICASSP 2005, pp I489-492

Finding relevant segments in meeting recordings is important for summarization, browsing, and retrieval purposes. In this paper, we define relevance as the interest-level that meeting participants manifest as a group during the course of their interaction (as perceived by an external observer), and investigate the automatic detection of segments of high-interest from audio-visual cues. This is motivated by the assumption that there is a relationship between segments of interest to participants, and those of interest to the end user, e.g. of a meeting browser. We first address the problem of human annotation of group interest-level. On a 50-meeting corpus, recorded in a room equipped with multiple cameras and microphones, we found that the annotations generated by multiple people exhibit a good degree of consistency, providing a stable ground-truth for automatic methods. For the automatic detection of high-interest segments, we investigate a methodology based on Hidden Markov Models (HMMs) and a number of audio and visual features. Single- and multi-stream approaches were studied. Using precision and recall as performance measures, the results suggest that the automatic detection of group interest-level is promising, and that while audio in general constitutes the predominant modality in meetings, the use of a multi-modal approach is beneficial.


Thursday, January 05, 2006

Image Parsing: Unifying Segmentation, Detection, and Recognition

Zhuowen Tu, Xiangrong Chen, Alan L. Yuille, Song-Chun Zhu
University of California, Los Angeles


We propose a general framework for parsing images into regions and objects. In this framework, the detection and recognition of objects proceed simultaneously with image segmentation in a competitive and cooperative manner. We illustrate our approach on natural images of complex city scenes where the objects of primary interest are faces and text. This method makes use of bottom-up proposals combined with top-down generative models using the Data Driven Markov Chain Monte Carlo (DDMCMC) algorithm which is guaranteed to converge to the optimal estimate asymptotically. More precisely, we define generative models for faces, text, and generic regions– e.g. shading, texture, and clutter. These models are activated by bottom-up proposals. The proposals for faces and text are learnt using a probabilistic version of AdaBoost. The DDMCMC combines reversible jump and diffusion dynamics to enable the generative models to explain the input images in a competitive and cooperative manner. Our experiments illustrate the advantages and importance of combining bottom-up and top-down models and of performing segmentation and object detection/recognition simultaneously.

Link Here

Wednesday, January 04, 2006

CNN: Mars rovers keep exploring Red Planet

Twin robots mark second anniversary
Monday, January 2, 2006; Posted: 3:23 p.m. EST (20:23 GMT)

LOS ANGELES, California (AP) -- The warranty expired long ago on NASA's twin robots motoring around Mars.
In two years, they have traveled a total of seven miles. Not impressed? Try keeping your car running in a climate where the average temperature is well below zero and where dust devils can reach 100 mph.
These two golf cart-sized vehicles were only expected to last three months.
The full article.

Tuesday, January 03, 2006

[]: Latest News from (January 2, 2006)

COOPERS Project to Address Road-to-Vehicle Comms Techniques
The European Commission recently funded the COOPERS Integrated Project as one of a suite of projects addressing cooperative vehicle-highway systems.

INTERSAFE Reports Impressive Test Results
The PReVENT INTERSAFE project for intersection collision avoidance recently completed testing of its advanced intersection positioning and dynamic object detection system.

PReVENT UseRCams Plans Testing of 3D Camera on Vehicles
The UseRCams team held a workshop recently in Lindau, Germany.

INSAFES Defines Functions for Demonstrator Vehicles
The PReVENT INSAFES team held their second plenary meeting in early December.

ProFusion2 Publishes New Requirements
The PReVENT subproject ProFusion2 has recently published its requirements for sensor data fusion deployment in active/preventive safety applications. This report defines and presents the system requirements, use cases and test scenarios for the sensor data fusion framework to be developed in ProFusion2.

Monday, January 02, 2006

( 1/4 )my talk in lab meeting

Hi all,

This techenical report was written by Sven Ginka, who was Bob's visiting student @ ACFR, Univ. of Sydney, March-July, 2005. Also this report is mainly about Bob's research. Besides, Sven also provided solutions for some problems in SLAMMOT.

This report is available in our ftp server.


Sunday, January 01, 2006

EPFL SmartRob Contest

The SmartRob Contest is a national competition open to students of the Swiss Federal Institute of Technology in Lausanne (EPFL) and other technical high-schools. The competition is jointly organized every year by the Autonomous Systems Laboratory (ASL) and the Laboratory of Intelligent Systems (LIS).

The link.

The 50 Best Robots Ever

Wired Magazine, January 2006

They're exploring the deep sea and distant planets. They're saving lives in the operating room and on the battlefield. They're transforming factory floors and filmmaking. They're - oh c'mon, they're just plain cool! From Qrio to the Terminator, here are our absolute favorites (at least for now).

The link