Select a calendar:
Filter November Events by Event Type:
SUNMONTUEWEDTHUFRISAT
University Calendar
Events for November
-
Phd Social
Tue, Nov 05, 2013 @ 12:00 PM - 01:00 PM
Thomas Lord Department of Computer Science
University Calendar
We will host another one of our PhD student lunches!
We will be making some announcements for the PhD students, and fellow CS PhD students Rose Yu and Leandro Soriano Marcolino will share their experiences from the Heidelberg laureate forum.
You are welcome to join us. If you are interested in coming, please RSVP before this Friday (Nov 1):
http://bit.ly/1dF8XHQ
Location: Ronald Tutor Hall of Engineering (RTH) - 526
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
M.S. Preview Day 2013
Fri, Nov 08, 2013 @ 09:00 AM - 03:00 PM
Thomas Lord Department of Computer Science
University Calendar
Preview Day is the Viterbi School's Graduate open house event for students interested in pursuing their Master's degree at one of the top ten ranked graduate engineering institutions in the nation. The event is free to all prospective students.
We request that attendees have earned or are candidates to earn at least a Bachelor's degree in engineering, math, or hard science (such as physics, chemistry or biology).
Learn more and register on our special Preview Day webpageMore Information: 69afb5e7_a60d6d24_Preview_Day_Agenda_Web.pdf
Location: Town & Gown (TGF) -
Audiences: Everyone Is Invited
Contact: Ryan Rozan
-
PhD Defense - Derya Ozkan
Tue, Nov 19, 2013 @ 01:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
University Calendar
Title: Towards Intelligent Virtual Listeners: Computational Models of Social Nonverbal Behaviors
PhD Candidate: Derya Ozkan
Committee:
Louis-Philippe Morency (Chair)
Gerard Medioni
Jonathan Gratch
Stacy Marsella
Shrikanth Narayanan (outside member)
Human nonverbal communication is a highly interactive process, in which the participants dynamically send and respond to nonverbal signals. These signals play a significant role in determining the nature of a social exchange. Although human can naturally recognize, interpret and produce these nonverbal signals in social context, computers are not equipped with such abilities. Therefore, creating computational models for holding fluid interactions with human participants has become an important topic for many research fields including human-computer interaction, robotics, artificial intelligence, and cognitive sciences. Central to the problem of modeling social behaviors is the challenge of understanding the dynamics involved with listener backchannel feedbacks (i.e. the nods and paraverbals such as ``uh-hu'' and ``mm-hmm'' that listeners produce as someone is speaking).
In this thesis, I present a framework for modeling visual backchannels of a listener during a dyadic conversation. I address the four major challenges involved in modeling nonverbal human behaviors, more specifically listener backchannels: (1) high dimensional data, (2) multimodal processing, (3) mutual influence between the participants, and (4) variability in human's behaviors. We address the first challenge by proposing a sparse feature selection method that gives researchers a new tool to analyze human nonverbal communication. To address to second challenge of effective and efficient fusion of multimodal information, we introduce a new model called Latent Mixture of Discriminative Experts (LMDE) that can automatically learn the hidden dynamic between modalities. For the third challenge, we present a context-based prediction framework that models the mutual influence between the participants of a human conversation to improve the final prediction model. Finally, we propose a new approach for modeling wisdom of crowds called wisdom-LMDE, which is able to learn the variations and commonalities among different crowd members.
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
PhD Defense - Lin Quan
Fri, Nov 22, 2013 @ 02:00 PM - 04:00 PM
Thomas Lord Department of Computer Science
University Calendar
Title: LEARNING ABOUT THE INTERNET THROUGH EFFICIENT SAMPLING AND AGGREGATION
PhD Candidate: Lin Quan
Committee:
- John Heidemann (Chair)
- Ming-Deh Huang
- Ethan Katz-Bassett
- Antonio Ortega (EE, Outside)
Time: Friday Nov 22 @ 2pm-4pm
Location: SAL 322
Abstract:
The Internet is important for nearly all aspects of our society,
affecting ordinary people, businesses, and social activities. Because of its importance and wide-spread applications, we want to have good knowledge about Internet's operation, reliability and performance, through various kinds of measurements. However, despite the wide usage, we only have limited knowledge of its overall performance and reliability. The first reason of this limited knowledge is that there is no central governance of the Internet, making both active and passive measurements hard. The second reason is the huge scale of the Internet. This makes brute-force analysis hard because of practical computing resouce limits such as CPU, memory and probe rate.
This thesis states that sampling and aggregation are necessary to
overcome resource constraints in time and space to learn about better knowledge of the Internet. Many other Internet measurement studies also utilize sampling and aggregation techniques to discover properties of the Internet. We distinguish our work by exploring novel mechanisms and new knowledge in several specific areas. First, we aggregate short-time-scale observations and use an efficient multi-time-scale query scheme to discover the properties and reasons of long-lived Internet flows. Second, we sample and probe /24 blocks in the IPv4 address space, and use greedy clustering algorithms to
efficiently characterize Internet outages. Third, we show an
efficient and effective aggregation technique by visualization and clustering. This technique makes both manual inspection and automated characterization easier. Last, we develop an adaptive probing system to study global scale Internet reliability. It samples and adapts probe rate within each /24 block for accurate beliefs. By aggregation and correlation to other domains, we are also able to study broader policy effects on Internet use, such as political causes, economic conditions, and access technologies.
This thesis provides several examples of Internet knowledge discovery with new mechanisms of sampling and aggregation techniques. We believe our approaches of new sampling and aggregation mechanisms can be used by and will inspire new ways for future Internet measurement systems to overcome resource constraints, such as large amount and dispersed data.
Location: Henry Salvatori Computer Science Center (SAL) - 322
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
PhD Defense - Mrinal Kalakrishnan
Tue, Nov 26, 2013 @ 12:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
University Calendar
PhD Candidate: Mrinal Kalakrishnan
Committee members:
Stefan Schaal (chair)
Gaurav Sukhatme
Francisco Valero-Cuevas (outside member)
Time: Nov 26th 12:00pm
Location: RTH 422
Title: Learning objective functions for autonomous motion generation
Abstract:
Planning and optimization methods have been widely applied to the problem of trajectory generation for autonomous robotics. The performance of such methods, however, is critically dependent on the choice of objective function being optimized, which is non-trivial to design. On the other hand, efforts on learning autonomous behavior from user-provided demonstrations have largely been focused on reproducing behavior similar in appearance to the demonstrations, which often fails to generalize well to new situations. An alternative approach, known as Inverse Reinforcement Learning (IRL), is to learn an objective function that the demonstrations are assumed to be optimal under. With the help of a planner or trajectory optimizer, such an approach allows the system to synthesize novel behavior in situations that were not experienced in the demonstrations.
We present novel algorithms for IRL that have successfully been applied in two real-world, competitive robotics settings: (1) In the domain of rough terrain quadruped locomotion, we present an algorithm that learns an objective function for foothold selection based on "terrain templates". The learner automatically generates and selects the appropriate features which form the objective function, which reduces the need for feature engineering while attaining a high level of generalization. (2) For the domain of autonomous manipulation, we present a probabilistic model of optimal trajectories, which results in new algorithms for inverse reinforcement learning and trajectory optimization in high-dimensional settings. We apply this method to two problems in robotic manipulation: redundancy resolution in inverse kinematics, and trajectory optimization for grasping and manipulation.
Both methods have proven themselves as part of larger integrated systems in competitive settings against other teams, where testing was conducted by an independent test team in situations that were not seen during training.
Location: Ronald Tutor Hall of Engineering (RTH) - 422
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
