Select a calendar:
Filter May Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for May
-
CS Colloquium: Ari Shapiro (ICT) - Models of Motion, Movement and Interaction for Digital Characters
Fri, May 01, 2015 @ 01:00 PM - 02:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Ari Shapiro, USC Institute for Creative Technologies
Talk Title: Models of Motion, Movement and Interaction for Digital Characters
Series: CS Colloquium
Abstract: Research in animation has progressed where capture technologies have allowed recording and playback of human motion. For example, a human face can be recorded speaking an utterance, then accurately modeled in 3D. However, making the 3D face produce an utterance that has not previously been recorded requires an understanding of how the face reacts to the speech that is generated, how the head and neck must move to accommodate that sound as well as that expression, and how the other parts of the face and eyes act during the speech. Similarly, motion capture techniques allow the capture and replication of human walking or running as performed by the original actor, but arbitrary movement through uneven terrain with obstacles cannot be synthesized accurately, since the complexity of the human balance and structure is not accurately modeled using only kinematic points in space over time.
Thus, while motion replication into a 3D environment is fairly well understood across a number of areas, the fundamental question of how to synthesize movement through a controllable model of humans remains elusive. The human body is extremely complex, and models of movement for high energy activities such as running differ greatly from other complex phenomena such as talking or gesturing. Thus, while it is possible to replicate a recorded motion, generating a controllable model of movement for a virtual human remains an open research problem for many different areas, ranging from facial expression to speech to gross movement. In addition, the motivations for human movement and motion are often driven by cognitive functions, so a better understanding of human movement requires a similar understanding of the cognitive aspects that motivate it.
In this talk, I will describe my research in generating various controllable models of motion and movement for animated 3D characters. My objective is to better understand how people physically move, interact and respond to people and objects in their environment By better understanding how people move about and the motivations for doing so, we can create models of human movement and behavior that can be controlled within a virtual or digital space, thus enabling convincing virtual characters that can be used for various types of training and simulation. The embodiment of movement and behavior of a person into a controllable, digital model allows for the creation of complicated scenarios that can be effective substitutes and training environments for real-world experiences.
The lecture will be available to stream HERE. (Right Click, New Tab for optimal results.)
Biography: Ari Shapiro currently works as a Research Scientist at the USC Institute for Creative Technologies, where his focus is on synthesizing realistic animation for virtual characters as lead of the Character Animation and Simulation research group. Shapiro has published many academic articles in the field of computer graphics and animation for virtual characters, and is a seven-time SIGGRAPH speaker.
For several years, he worked on character animation tools and algorithms in the research and development departments of visual effects and video games companies such as Industrial Light and Magic, LucasArts and Rhythm and Hues Studios. He has worked on many feature-length films, and holds film credits in The Incredible Hulk and Alvin and the Chipmunks 2. In addition, he holds video games credits in the Star Wars: The Force Unleashed series.
He completed his Ph.D. in computer science at UCLA in 2007 in the field of computer graphics with a dissertation on character animation using motion capture, physics and machine learning. He also holds an M.S. in computer science from UCLA, and a B.A. in computer science from the University of California, Santa Cruz.
Host: CS Department
Webcast: https://bluejeans.com/577232541Location: Grace Ford Salvatori Hall Of Letters, Arts & Sciences (GFS) - 108
WebCast Link: https://bluejeans.com/577232541
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Wei Cheng (UCLA) - Integrating Multiple Networks for Big Data Analysis
Tue, May 05, 2015 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Wei Cheng, UCLA
Talk Title: Integrating Multiple Networks for Big Data Analysis
Series: CS Colloquium
Abstract: In many big data applications, data with complex structures can usually be modeled as network data. Usually, for one data mining problem, we have multiple networks. For one thing, data about the same object can be obtained from various. For another, the different objects may have complex structures and can be interrelated in a complex way. Integration of different network data is valuable for reaching a more accurate decision and discovering novel patterns. The task is challenging because of the inherent characteristics of the networks: 1) variety (e.g., complex structures, heterogeneous types and data sources); and 2) poor quality; 3) massive volume. In this talk, I will present our research efforts to use big data technologies to integrate multiple networks for both supervised and unsupervised data mining problems. First, I will begin by presenting the work of integrative analyzing multi-domain heterogeneous data for graph clustering. Next, I will present the work on robust sparse regression algorithm that integrates multi-source heterogeneous networks.
Biography: Wei Cheng is a Ph.D. candidate in Computer Science at University of North Carolina at Chapel Hill. He has been visiting Department of Computer Science of UCLA since 2013. He received a Master's and Bachelor's degree from Tsinghua University and Nanjing University, in 2010 and 2006, respectively. His research interests include big data, data mining, bioinformatics, computational biology, and machine learning. He is especially interested in scalable data analysis problems for data science with an emphasis on biological applications Previously, he also conducted research at Microsoft Research and IBM Research as an intern.
Host: Yan Liu
Location: Henry Salvatori Computer Science Center (SAL) - 213
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Kate Saenko (University of Massachusetts Lowell) - From Video to Sentences: A Deep Learning Approach
Wed, May 06, 2015 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Kate Saenko , University of Massachusetts Lowell
Talk Title: From Video to Sentences: A Deep Learning Approach
Series: CS Colloquium
Abstract: I will describe several recent advances in automatic generation of natural language descriptions for video. Video description has important applications in human-robot interaction, video indexing, and describing movies for the blind. Real-world videos often have complex dynamics, but current methods are insensitive to temporal structure and do not allow both input (sequence of frames) and output (sequence of words) of variable length. I will describe a novel sequence-to-sequence neural network that learns to generate captions for brief videos. The model is trained on video-sentence pairs and is naturally able to learn the temporal structure of the sequence of frames as well as the sequence model of the generated sentences, i.e. a language model. To further handle the ambiguity over multiple objects and locations, the model incorporates convolutional networks with Multiple Instance Learning (MIL) to consider objects in different positions and at different scales simultaneously. The multi-scale multi-instance convolutional network is integrated with a sequence-to-sequence recurrent neural network to generate sentence descriptions based on the visual representation. This architecture is the first end-to-end trainable deep neural network that is capable of multi-scale region processing for video description. I will show results of captioning YouTube videos and Hollywood movies.
Biography: Kate Saenko is an Assistant Professor of Computer Science at the University of Massachusetts Lowell. She received her PhD from MIT, followed by postdoctoral work at UC Berkeley and Harvard. Her research spans the areas of computer vision, machine learning, speech recognition, and human-robot interfaces. Dr Saenko's current research interests include domain adaptation for object recognition and joint modeling of language and vision. She is involved in a large multi-institution NSF-sponsored project, conducting research in statistical scene understanding and physics-based visual reasoning. She is also a recipient of an NSF EAGER award to analyse domain invariance of deep learning models. Previously, she was involved in DARPA's Mind's Eye project, developing methods for recognizing and describing human activities in video.
Host: Fei Sha
Location: Grace Ford Salvatori Hall Of Letters, Arts & Sciences (GFS) - 222
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
Teamcore Seminar: Dr. William Haskell (National University of Singapore) - Approximate Dynamic Programming
Wed, May 13, 2015 @ 10:30 AM - 11:30 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dr. William Haskell, National University of Singapore
Talk Title: Approximate Dynamic Programming
Series: Teamcore Seminar
Abstract: We develop a new technique for analyzing the convergence of stochastic algorithms. This technique is based on the notion of stochastic dominance and allows us to get sample complexity results. We apply this technique to study the convergence of several approximate dynamic programming algorithms for MDPs on continuous state spaces, as well as to propose some new algorithms.
Biography: Dr. William Haskell is an assistant professor in the department of the industrial & systems engineering at National University of Singapore. He obtained a PhD from the department of industrial engineering and operation research from UC Berkeley. He was a visiting scholar at USC ISE department from August 2011 to May 2013 and then a Postdoctoral Research Associate from June 2011 to May 2014 at the USC EE and CS department. His research has focused on risk-aware decision making, sequential and large-scale optimization and data-driven decision making.
Host: Teamcore Research Group
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Jelena Marašević (Columbia U.) - Full-Duplex Wireless: Resource Allocation and Rate Gains for Realistic Hardware Models
Tue, May 26, 2015 @ 10:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Jelena MaraÅ¡eviÄ, Columbia University
Talk Title: Full-Duplex Wireless: Resource Allocation and Rate Gains for Realistic Hardware Models
Series: CS Colloquium
Abstract: Full-duplex communication -“ simultaneous transmission and reception on the same frequency channel -- has the potential to substantially increase the throughput in wireless networks. The achievable rate gains and the effect of full-duplex capabilities on physical and medium access control (MAC) layers, however, are still not fully understood. In this talk, I will present our recent results on power allocation in single-channel and multi-channel settings, where the objective is to maximize the sum of the rates on uplink and downlink full-duplex channels. Specifically, I will discuss power allocation in the single-channel use cases, and present a sufficient condition under which the sum of uplink and downlink rates on a full-duplex channel is concave in the transmission power levels. This condition is essential for the design of a power allocation algorithm in the multi-channel setting. For the multi-channel setting, I will present a new realistic model of a small form-factor (e.g., a smartphone) full-duplex receiver, demonstrating its accuracy via measurement results. For the problem of jointly allocating power levels to different channels, where the objective is maximizing the sum of the rates over uplink and downlink OFDM channels, I will present two algorithms. The first is a polynomial-time algorithm that is nearly optimal under very mild restrictions. The second algorithm reduces the running time substantially, and is nearly-optimal under the high SINR approximation. Overall, our results provide a precise quantification of the achievable rate gains as a function of signal-to-noise ratios and (self-)interference-to-noise-ratios.
Based on joint work with J. Zhou, H. Krishnaswamy, Y. Zhong, and G. Zussman that will appear in Proc. ACM SIGMETRICS '15.
Biography: Jelena MaraÅ¡eviÄ is a Ph.D. student at Columbia University. Her research focuses on algorithms for fair resource allocation problems, with applications in wireless networks. She received her B.Sc. degree from University of Belgrade, School of Electrical Engineering, in 2011, and her M.S. degree in electrical engineering from Columbia University in 2012. Jelena is a recipient of the M.S. Award of Excellence and the Jacob Millman Prize for Excellence in Teaching Assistance from Columbia University. She is also a winner of the Qualcomm Innovation Fellowship 2015 award.
Host: CS Department
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Jelena Marasevic (Columbia U.) - Full-Duplex Wireless: Resource Allocation and Rate Gains for Realistic Hardware Models
Tue, May 26, 2015 @ 10:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Jelena Marasevic, Columbia University
Talk Title: Full-Duplex Wireless: Resource Allocation and Rate Gains for Realistic Hardware Models
Series: CS Colloquium
Abstract: Full-duplex communication -“ simultaneous transmission and reception on the same frequency channel -- has the potential to substantially increase the throughput in wireless networks. The achievable rate gains and the effect of full-duplex capabilities on physical and medium access control (MAC) layers, however, are still not fully understood. In this talk, I will present our recent results on power allocation in single-channel and multi-channel settings, where the objective is to maximize the sum of the rates on uplink and downlink full-duplex channels. Specifically, I will discuss power allocation in the single-channel use cases, and present a sufficient condition under which the sum of uplink and downlink rates on a full-duplex channel is concave in the transmission power levels. This condition is essential for the design of a power allocation algorithm in the multi-channel setting. For the multi-channel setting, I will present a new realistic model of a small form-factor (e.g., a smartphone) full-duplex receiver, demonstrating its accuracy via measurement results. For the problem of jointly allocating power levels to different channels, where the objective is maximizing the sum of the rates over uplink and downlink OFDM channels, I will present two algorithms. The first is a polynomial-time algorithm that is nearly optimal under very mild restrictions. The second algorithm reduces the running time substantially, and is nearly-optimal under the high SINR approximation. Overall, our results provide a precise quantification of the achievable rate gains as a function of signal-to-noise ratios and (self-)interference-to-noise-ratios.
Based on joint work with J. Zhou, H. Krishnaswamy, Y. Zhong, and G. Zussman that will appear in Proc. ACM SIGMETRICS '15.
Biography: Jelena Marasevic is a Ph.D. student at Columbia University. Her research focuses on algorithms for fair resource allocation problems, with applications in wireless networks. She received her B.Sc. degree from University of Belgrade, School of Electrical Engineering, in 2011, and her M.S. degree in electrical engineering from Columbia University in 2012. Jelena is a recipient of the M.S. Award of Excellence and the Jacob Millman Prize for Excellence in Teaching Assistance from Columbia University. She is also a winner of the Qualcomm Innovation Fellowship 2015 award.
Host: CS Department
Location: 248
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Hyun Soo Park (University of Pennsylvania) - Computational Social Cognition
Thu, May 28, 2015 @ 12:00 PM - 01:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Hyun Soo Park, University of Pennsylvania
Talk Title: Computational Social Cognition
Series: CS Colloquium
Abstract: Humans interact with one another by sending visible social signals such as facial expressions, body gestures, and gaze directions.
Computational understanding of these social signals is becoming more important for artificial agents such as service robots because they are increasingly integrated in our social space.
In this talk, I will present a computational framework for social cognition - the ability to perceive, model, and predict social signals.
The main challenges of developing computational social cognition are that 1) social signals are too subtle to be detected by current computer vision solutions and 2) they cannot be understood by analyzing an individual signal in isolation as they are reliant upon each other. I will argue that first person cameras, e.g., head-mounted cameras, are an ideal sensor placement to capture such subtlety and will show that the relationship between the signals can be modeled by leveraging a 3D reconstruction of human body motion. In the first part of my talk, I will focus on joint attention that encodes the relationship between gaze directions and present its predictive model to recognize social interactions. This predictive model is applied various tasks, e.g., event video editing, social anomaly recognition, and region of interest detection. In the second part, I will introduce a large scale motion capture system (510 cameras) to recover subtle social signals. This system reconstructs dense 3D trajectories of body gestures at unprecedented level of high spatial resolution (~20,000 trajectories per body). Then, I will demonstrate applications of computational social cognition in behavioral analysis, sport analytics, and robotics.
Biography: Hyun Soo Park is a Postdoctoral Fellow in Computer and Information Science at the University of Pennsylvania working with Prof. Jianbo Shi. He earned Ph.D. degree from Carnegie Mellon University in 2014 under the supervision of Prof. Yaser Sheikh. His research aims to develop a computational representation of social behaviors. He has over 15 publications in top tier conferences and journals that include computer vision (IJCV, ICCV, CVPR, ECCV), graphics (SIGGRAPH), machine learning (NIPS), and robotics (IJRR, ICRA, IROS). He organized Workshop on Human Behavior Understanding (2014) in conjunction with ECCV 2014 and will give a tutorial on Group Behavioral Analysis and its Applications in conjunction with CVPR 2015 based on his Ph.D. thesis. His work has been covered by various major media including Discovery Channel, MSNBC, WIRED, NSF, and Slashdot. Prior to his Ph.D., he received his M.S. degree from Carnegie Mellon University and his B.S. degree from POSTECH.
Host: Hao Li
Location: Henry Salvatori Computer Science Center (SAL) - 322
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Baoquan Chen (Shandong University) - 3D Urban Sensing and Visualization
Fri, May 29, 2015 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Baoquan Chen, Shandong University
Talk Title: 3D Urban Sensing and Visualization
Series: CS Colloquium
Abstract: 3D modeling of urban environments starts to play an increasingly important role in the emerging technologies from self-driving car to augmented reality. Beyond helping a human or a vehicle navigate, 3D urban models provide a base for spatially registering otherwise chaotic urban data, both sensor sensed and user generated, for better 'mapping' of urban big data. In this talk, I will introduce our decade long effort on acquiring and modeling large urban environments as well as analyzing and visualizing urban activities. I will also discuss future developments in this direction.
Biography: Baoquan Chen is a Professor and Dean (CS & Software) of Shandong University. Prior to the current post, he was the founding director of the Visual Computer Research Center, Shenzhen Institutes of Advanced Technology (SIAT), Chinese Academy of Sciences, and a faculty member at CS&E at the University of Minnesota at Twin Cities. His research interests generally lie in computer graphics, visualization, and human-computer interaction. Chen received PhD in CS from SUNY@Stony Brook, and MS in EE from Tsinghua. He received NSF CAREER award in 2003 and IEEE Visualization Best Paper Award in 2005. Chen served as conference chair of IEEE Visualization 2005, and more recently SIGGRAPH Asia 2014. More at: http://www.cs.sdu.edu.cn/~baoquan/
Host: Hao Li
Location: Grace Ford Salvatori Hall Of Letters, Arts & Sciences (GFS) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
