Select a calendar:
Filter March Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for March
-
CS Colloquium: Oded Stein (MIT) - Mathematical Foundations of Robust Geometry and Fabrication
Tue, Mar 01, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Oded Stein, MIT
Talk Title: Mathematical Foundations of Robust Geometry and Fabrication
Series: CS Colloquium
Abstract: Current geometry methods for creating and manipulating shapes on computers can sometimes be unreliable and fail unpredictably. Such failures make geometry tools hard to use, prevent non-experts from creating geometry on their computers, and limit the use of geometry methods in domains where reliability is critical. We will discuss my recent efforts in proving when existing methods work as intended, my work in making methods more robust to imperfect input, my work in the creation of new reliable tools with mathematical guarantees, and my future efforts towards a reliable geometry pipeline.
When used for computational fabrication, geometry methods can be expensive, finicky, and require a controlled environment. I will show how simple and economical manufacturing techniques can be used for computational fabrication by exploiting the geometric constraints inherent in specific materials and fabrication methods. We will take a look at how I create geometric tools to design for constrained fabrication techniques, and discuss how computational fabrication can be made both economical as well as accessible in difficult environments.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Oded Stein is a postdoc at MIT at the geometric data processing group. He obtained his MSc from ETH Zurich in 2015, and his PhD from Columbia University in 2020.
Oded is interested in geometry, computer graphics, and applied mathematics. He works on smoothness energies, partial differential equations, discretization of geometric quantities, and their applications to computer graphics and digital fabrication.
Host: Jernej Barbic
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Amy Ousterhout (UC Berkeley) - Optimizing CPU Efficiency and Tail Latency in Datacenters
Tue, Mar 01, 2022 @ 02:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Amy Ousterhout , UC Berkeley
Talk Title: Optimizing CPU Efficiency and Tail Latency in Datacenters
Series: CS Colloquium
Abstract: The slowing of Moore's Law and increased concerns about the environmental impacts of computing are exerting pressure on datacenter operators to use resources such as CPUs and memory more efficiently. However, it is difficult to improve efficiency without degrading the performance of applications.
In this talk, I will focus on CPU efficiency and how we can increase efficiency while maintaining low tail latency for applications. The key innovation is to reallocate cores between applications on the same server very quickly, every few microseconds. First I will describe Shenango, a system design that makes such frequent core reallocations possible. Then I will show how policy choices for core reallocation and load balancing impact CPU efficiency and tail latency, and present the policies that yield the best combination of both.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Amy is a postdoctoral researcher in the Department of Electrical Engineering and Computer Sciences at UC Berkeley. She received her PhD in Computer Science from MIT and her BSE in Computer Science from Princeton University. Her research is on operating systems and distributed systems, and focuses on improving the efficiency, performance, and usability of applications in datacenters. She is a recipient of a Jacobs Presidential Fellowship at MIT, an NSF Graduate Research Fellowship, and a Hertz Foundation Fellowship.
Host: Barath Raghavan
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Dmitry Berenson (University of Michigan) - Learning Where to Trust Unreliable Dynamics Models for Motion Planning and Manipulation
Tue, Mar 01, 2022 @ 04:15 PM - 05:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dmitry Berenson, University of Michigan
Talk Title: Learning Where to Trust Unreliable Dynamics Models for Motion Planning and Manipulation
Series: Computer Science Colloquium
Abstract: **New time: 4:15pm-5:20pm, SGM 124**
The world outside our labs seldom conforms to the assumptions of our models. This is especially true for dynamics models used in control and motion planning for complex high-DOF systems like deformable objects. We must develop better models, but we must also accept that, no matter how powerful our simulators or how big our datasets, our models will sometimes be wrong. This talk will present our recent work on using unreliable dynamics models for motion planning and manipulation. Given a dynamics model, our methods learn where that model can be trusted given either batch data or online experience. These approaches allow imperfect dynamics models to be useful for a wide range of tasks in novel scenarios, while requiring much less data than baseline methods. This data-efficiency is a key requirement for scalable and flexible motion planning and manipulation capabilities.
Prof. Dmitry Berenson will give his talk in person at SGM 124 and we will also host the talk over Zoom.
Register in advance for this webinar at:
https://usc.zoom.us/webinar/register/WN_prfowdXjR7iOn1mPLTnXog
After registering, attendees will receive a confirmation email containing information about joining the webinar.
This lecture satisfies requirements for CSCI 591: Research Colloquium.
Biography: Dmitry Berenson is an Associate Professor in Electrical Engineering and Computer Science and the Robotics Institute at the University of Michigan, where he has been since 2016. Before coming to University of Michigan, he was an Assistant Professor at WPI (2012-2016). He received a BS in Electrical Engineering from Cornell University in 2005 and received his Ph.D. degree from the Robotics Institute at Carnegie Mellon University in 2011, where he was supported by an Intel PhD Fellowship. He was also a post-doc at UC Berkeley (2011-2012). He has received the IEEE RAS Early Career Award and the NSF CAREER award. His current research focuses on robotic manipulation, robot learning, and motion planning.
Host: Stefanos Nikolaidis
Webcast: https://usc.zoom.us/webinar/register/WN_prfowdXjR7iOn1mPLTnXogLocation: Seeley G. Mudd Building (SGM) - 124
WebCast Link: https://usc.zoom.us/webinar/register/WN_prfowdXjR7iOn1mPLTnXog
Audiences: Everyone Is Invited
Contact: Computer Science Department
-
CS Colloquium: Saining Xie (Facebook AI Research (FAIR)) - Towards Scalable Representation Learning for Visual Recognition
Thu, Mar 03, 2022 @ 10:00 AM - 11:00 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Saining Xie, Facebook AI Research (FAIR)
Talk Title: Towards Scalable Representation Learning for Visual Recognition
Series: CS Colloquium
Abstract: A powerful biological and cognitive representation is essential for humans' remarkable visual recognition abilities. Deep learning has achieved unprecedented success in a variety of domains over the last decade. One major driving force is representation learning, which is concerned with learning efficient, accurate, and robust representations from raw data that are useful for a downstream classifier or predictor. A modern deep learning system is composed of two core and often intertwined components: 1) neural network architectures and 2) representation learning algorithms. In this talk, we will present several studies in both directions. On the neural network modeling side, we will examine modern network design principles and how they affect the scaling behavior of ConvNets and recent Vision Transformers. Additionally, we will demonstrate how we can acquire a better understanding of neural network connectivity patterns through the lens of random graphs. In terms of representation learning algorithms, we will discuss our recent efforts to move beyond the traditional supervised learning paradigm and demonstrate how self-supervised visual representation learning, which does not require human annotated labels, can outperform its supervised learning counterpart across a variety of visual recognition tasks. The talk will encompass a variety of vision application domains and modalities (e.g. 2D images and 3D scenes). The goal is to show existing connections between the techniques specialized for different input modalities and provide some insights about diverse challenges that each modality presents. Finally, we will discuss several pressing challenges and opportunities that the "big model era" raises for computer vision research.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Saining Xie is a research scientist at Facebook AI Research (FAIR). He received his Ph.D. and M.S. degrees in computer science from the University of California San Diego, advised by Zhuowen Tu. Prior to that, he received his Bachelor's degree from Shanghai Jiao Tong University. He has broad research interests in deep learning and computer vision, with a focus on developing deep representation learning techniques to push the boundaries of core visual recognition. He is a recipient of the Marr Prize Honorable Mention at ICCV 2015.
Host: Ramakant Nevatia
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Lars Lindemann (University of Pennsylvania) - Safe AI-Enabled Autonomy
Thu, Mar 03, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Lars Lindemann, University of Pennsylvania
Talk Title: Safe AI-Enabled Autonomy
Series: CS Colloquium
Abstract: AI-enabled autonomous systems show great promise to enable many future technologies such as autonomous driving, intelligent transportation, and robotics. Over the past years, there has been tremendous success in the development of autonomous systems, which was especially accelerated by the computational advances in machine learning and AI. At the same time, however, new fundamental questions were raised regarding the safety and reliability of these increasingly complex systems that often operate in uncertain and dynamic environments. In this seminar, I will provide new insights and exciting opportunities to address these challenges.
In the first part of the seminar, I will present a data-driven optimization framework to learn safe control laws for dynamical systems. For most safety-critical systems such as self-driving cars, safe expert demonstrations in the form of system trajectories that show safe system behavior are readily available or can easily be collected. At the same time, accurate models of these systems can be identified from data or obtained from first order modeling principles. To learn safe control laws, I present a constrained optimization problem with constraints on the expert demonstrations and the system model. Safety guarantees are provided in terms of the density of the data and the smoothness of the system model. Two case studies on a self-driving car and a bipedal walking robot illustrate the presented method. In the past years, it was shown that neural networks are fragile and that their use in AI-enabled systems has resulted in systems taking excessive risk. The second part of the seminar is motivated by this fact and presents a data-driven verification framework to quantify and assess the risk of AI-enabled systems. I particularly show how risk measures, classically used in finance, can be used to quantify the risk of not being robust to failure, and how we can estimate this risk from data. We will compare and verify four different neural network controllers in terms of their risk for a self-driving car. I will conclude by sharing exciting research directions in this area.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Lars Lindemann is currently a Postdoctoral Researcher in the Department of Electrical and Systems Engineering at the University of Pennsylvania. He received his B.Sc. degrees in Electrical and Information Engineering and his B.Sc. degree in Engineering Management in 2014 from the Christian-Albrechts-University (CAU), Kiel, Germany. He received his M.Sc. degree in Systems, Control and Robotics in 2016 and his Ph.D. degree in Electrical Engineering in 2020, both from KTH Royal Institute of Technology, Stockholm, Sweden. His current research interests include systems and control theory, formal methods, data-driven control, and autonomous systems. Lars received the Outstanding Student Paper Award at the 58th IEEE Conference on Decision and Control and was a Best Student Paper Award Finalist at the 2018 American Control Conference. He also received the Student Best Paper Award as a co-author at the 60th IEEE Conference on Decision and Control.
Host: Jyo Deshmukh
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Sai-Kit Yeung (Hong Kong University of Science and Technology (HKUST).) - Computer Vision and Graphics for Real-World Challenges
Fri, Mar 04, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Sai-Kit Yeung, Hong Kong University of Science and Technology (HKUST)
Talk Title: Computer Vision and Graphics for Real-World Challenges
Series: CS Colloquium
Abstract: With the recent advancements in sensing technology and pervasive computing devices, the fields of computer vision and graphics are witnessing renewed importance in addressing real-world problems. In this talk, I will be discussing my research relating to 3D reconstruction, scene understanding, content generation, and fabrication. My talk will also overview ways this core research can be used in multidisciplinary projects involving city planning, seafloor surveying, and fishery design. I will conclude my talk by discussing potential collaborative projects between computer vision, graphics, and other disciplines to address challenging issues related to human empowerment and the building of sustainable environments.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Dr. Sai-Kit Yeung is an Associate Professor at the Division of Integrative Systems and Design (ISD) and the Department of Computer Science and Engineering (CSE) at the Hong Kong University of Science and Technology (HKUST). His research interests include 3D vision and graphics, content generation, fabrication, and novel computational techniques and integrative systems for marine-related problems.
Dr. Yeung has published extensively in premiere computer vision and graphics venues including numerous full oral papers in CVPR, ICCV, and AAAI. His work has received best paper honorable mention awards at ICCP 2015 and 3DV 2016. He has served as a Senior Program Committee member in IJCAI and AAAI, and as a Course Chair for SIGGRAPH Asia 2019. In addition, he regularly serves as a Technical Papers Committee member for SIGGPAPH & SIGGRAPH Asia and is currently an Associate Editor of the ACM Transactions on Graphics (TOG).
Host: Jernej Barbic
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CAIS Seminar: Patrick Fowler - Equity in Data-Driven Policies
Mon, Mar 07, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Patrick Fowler,
Talk Title: Equity in Data-Driven Policies
Series: USC Center for Artificial Intelligence in Society (CAIS) Seminar Series
Abstract: Dr. Fowler will present work investigating group fairness in algorithmic decision-making. The team uses administrative records on homeless service delivery to demonstrate inherent tradeoffs in fairness that depend on the operationalization of equity. The findings inform data-driven policy-making in homelessness and social services broadly.
Register in advance for this webinar at:
https://usc.zoom.us/webinar/register/WN_b3XxZm-rRfibw54-40F-UA
After registering, attendees will receive a confirmation email containing information about joining the webinar.
This lecture satisfies requirements for CSCI 591: Research Colloquium.
Biography: Dr. Patrick Fowler's research centers on homelessness prevention and the negative consequences of homelessness on youth, families, and communities. His recent research focuses on homelessness among youth aging out of foster care and child maltreatment prevention among families experiencing homelessness.
Using linked administrative data, Dr. Fowler designs and tests big data applications to improve the fairness and efficiency of homelessness services delivery. He employs a complex systems approach to create developmentally appropriate and culturally tailored responses to homelessness.
Host: USC Center for Artificial Intelligence in Society (CAIS)
Webcast: https://usc.zoom.us/webinar/register/WN_b3XxZm-rRfibw54-40F-UALocation: Online - Zoom Webinar
WebCast Link: https://usc.zoom.us/webinar/register/WN_b3XxZm-rRfibw54-40F-UA
Audiences: Everyone Is Invited
Contact: Computer Science Department
-
CS Colloquium: Fei Miao (University of Connecticut) - Learning, Optimization and Control for Efficiency and Security of Cyber-Physical Systems
Tue, Mar 08, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Fei Miao, University of Connecticut
Talk Title: Learning, Optimization and Control for Efficiency and Security of Cyber-Physical Systems
Series: CS Colloquium
Abstract: Ubiquitous sensing enables large-scale multi-source data of cyber-physical systems (CPS) collected in real-time and poses both challenges and opportunities for a paradigm-shift to data-driven CPS. For instance, how to capture the complexity and analyze the dynamical state information from data, and make decisions to improve safety, efficiency and security of the networked CPS is still challenging. In this talk, we present our research that integrates optimization, machine learning, control, and game theory to address these challenges, including theoretical contributions, algorithmic design, and experimental validations. We first present data-driven distributionally robust optimization (DRO) methods for CPS efficiency, with application on smart city resource allocation. We design algorithms to construct the uncertainty sets of the model prediction based on spatial temporal data. We prove the computationally tractable forms or equivalent convex optimization forms of the DRO problems to guarantee the worst-case expected cost of real-time decisions. We show the improvement of autonomous mobility-on-demand (AMoD) service fairness and efficiency based on large-scale dataset. Second, we summarize our research contribution for CPS security. We mainly present a hybrid state stochastic game model to guarantee the worst-case cost of the system, and a proved suboptimal algorithm to calculate the mixed policies. Finally, based on our active awarded projects, we briefly discuss future research directions on robust multi-agent reinforcement learning and data-driven robust optimization based decision-making, to address CPS safety, efficiency, and security challenges.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Fei Miao is an Assistant Professor of the Department of Computer Science & Engineering, a Courtesy Faculty of the Department of Electrical & Computer Engineering, University of Connecticut since 2017. She is also affiliated to the Institute of Advanced Systems Engineering and Eversource Energy Center. Her research interests lie in the optimization, machine learning, control, and game theory, to address safety, efficiency, and security challenges of cyber-physical systems. She received the Ph.D. degree and the Best Doctoral Dissertation Award in Electrical and Systems Engineering from the University of Pennsylvania in 2016. She received the B.S. degree majoring in Automation from Shanghai Jiao Tong University. She was a postdoc researcher at the GRASP Lab and the PRECISE Lab of Upenn from 2016 to 2017. Dr. Miao is a receipt of the NSF CAREER Award, the title of the project is "Distributionally Robust Learning, Control, and Benefits Analysis of Information Sharing for Connected and Autonomous Vehicles". Dr. Miao has also received a couple of other awards from NSF, including awards from the Smart & Autonomous Systems, the Cyber-Physical Systems, and the Smart & Connected Communities programs. She received Best Paper Award and Best Paper Award Finalist at the ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS) in 2021 and 2015, respectively.
Host: Jyo Deshmukh
Location: 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Christoforos Mavrogiannis (University of Washington) - Building Robots that Humans Accept
Wed, Mar 09, 2022 @ 10:00 AM - 11:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Christoforos Mavrogiannis, University of Washington
Talk Title: Building Robots that Humans Accept
Series: CS Colloquium
Abstract: Robotics has transformed sectors like manufacturing and fulfillment which now rely on robots to meet their goals. Conventionally, these robots operate in isolation from humans due to safety and efficiency considerations. Lately, there have been efforts towards bringing robots closer to humans to assist in everyday-life tasks, enhance productivity, and augment human capabilities. Despite these efforts, robotic technology has not reached widespread acceptance outside of factories; robot autonomy is often not robust, producing new problems that outweigh its benefits for users. Inspired by theories of technology acceptance, my research strives to develop highly functional, safe, and comfortable robots that humans accept. In this talk, I argue that the path towards acceptance requires imbuing robots with a deeper understanding of how users perceive and react to them. To motivate this perspective, I will share insights on robot navigation in dynamic environments, a fundamental task with many crucial applications ranging from collaborative manufacturing to warehouse automation and healthcare. I will describe a human-inspired algorithmic framework for crowd navigation, highlighting how mathematical abstractions of multiagent behavior enable safe, efficient, and positively perceived robot motion across a series of extensive empirical studies involving real robots and human subjects. Inspired by field-deployment challenges, I will then present a data-driven framework that enables robots to recover from failure via bystander help without overloading users. I will conclude with future directions on the development of shared and full robot autonomy that explicitly reasons about human perceptions to produce safe, trustworthy, and comfortable robot behavior.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Christoforos Mavrogiannis is a postdoctoral Research Associate in the Paul G. Allen School of Computer Science & Engineering at the University of Washington, working with Prof. Siddhartha Srinivasa. His interests lie at the intersection of robotics, human-robot interaction, and artificial intelligence. His research often draws insights from algebraic topology and dynamical systems, tools from machine learning, planning and control, and inspiration from social sciences. He is a full-stack roboticist, passionate about real-world deployment of robot systems, and extensive benchmarking with users. He has been a best-paper award finalist at the ACM/IEEE International Conference on Human-Robot Interaction (HRI), and selected as a Pioneer at the HRI and RSS conferences. He has also led open-source initiatives (Openbionics, MuSHR), for which he has been a finalist for the Hackaday Prize and a winner of the Robotdalen International Innovation Award. His work has received coverage from many media outlets including Wired, IEEE Spectrum, GeekWire, RoboHub, and the Hellenic Broadcasting Corporation. Christoforos holds M.S. and Ph.D. degrees from Cornell University, and a Diploma in mechanical engineering from the National Technical University of Athens.
Host: Jesse Thomason
Location: Ronald Tutor Hall of Engineering (RTH) - 109
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: David Held (Carnegie Mellon University) - Perceptual Robot Learning
Thu, Mar 10, 2022 @ 05:00 PM - 06:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: David Held, Carnegie Mellon University
Talk Title: Perceptual Robot Learning
Series: Computer Science Colloquium
Abstract: *New time: 5:00-6:20PM*
Robots today are typically confined to interact with rigid, opaque objects with known object models. However, the objects in our daily lives are often non-rigid, can be transparent or reflective, and are diverse in shape and appearance. One reason for the limitations of current methods is that computer vision and robot planning are often considered separate fields. I argue that, to enhance the capabilities of robots, we should design state representations that consider both the perception and planning algorithms needed for the robotics task. I will show how we can develop novel perception and planning algorithms to assist with the tasks of manipulating cloth, articulated objects, and transparent and reflective objects. By thinking about the downstream task while jointly developing perception and planning algorithms, we can significantly improve our progress on difficult robots tasks.
Register in advance for this webinar at:
https://usc.zoom.us/webinar/register/WN_X9bmT5afSU2gjC03nttQHg
After registering, attendees will receive a confirmation email containing information about joining the webinar.
This lecture satisfies requirements for CSCI 591: Research Colloquium.
Biography: David Held is an assistant professor at Carnegie Mellon University in the Robotics Institute and is the director of the RPAD lab: Robots Perceiving And Doing. His research focuses on perceptual robot learning, i.e. developing new methods at the intersection of robot perception and planning for robots to learn to interact with novel, perceptually challenging, and deformable objects. David has applied these ideas to robot manipulation and autonomous driving. Prior to coming to CMU, David was a post-doctoral researcher at U.C. Berkeley, and he completed his Ph.D. in Computer Science at Stanford University. David also has a B.S. and M.S. in Mechanical Engineering at MIT. David is a recipient of the Google Faculty Research Award in 2017 and the NSF CAREER Award in 2021.
Host: Stefanos Nikolaidis
Webcast: https://usc.zoom.us/webinar/register/WN_X9bmT5afSU2gjC03nttQHgLocation: Online - Zoom Webinar
WebCast Link: https://usc.zoom.us/webinar/register/WN_X9bmT5afSU2gjC03nttQHg
Audiences: Everyone Is Invited
Contact: Computer Science Department
-
CS Colloquium: Harsha V. Madhyastha (University of Michigan) - Inter-connecting society across space and time
Fri, Mar 11, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Harsha V. Madhyastha, University of Michigan
Talk Title: Inter-connecting society across space and time
Series: CS Colloquium
Abstract: Thanks to the Internet and a range of services that have been developed to take advantage of it -- web, email, social media, instant messaging, etc. -- being in the same place at the same time is no longer a requirement for all of us to share information with each other. Instead, we are able to store our ideas, opinions, and observations on services which enable others to access this information later from anywhere in the world.
In this talk, I will discuss my group/s work over the past several years to address some of the fundamental challenges faced by the providers of such global-scale services. I will provide examples of two broad research thrusts: 1) enabling cost-effective development and deployment of geo-distributed services, and 2) optimizing the availability and performance of client-service interactions. I will also briefly discuss my ongoing research in facilitating
information exchange in domains such as web archival, federated learning, and 3D printing.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Harsha V. Madhyastha is an Associate Professor in CSE at the University of Michigan. His research broadly spans the areas of distributed systems and
networking. Two of his papers have received the IRTF's Applied Networking Research Prize, and he has also co-authored award papers at OSDI, NSDI, and IMC. He has received multiple Google Faculty Research awards, a NetApp Faculty Fellowship, a Facebook Faculty Award, and an NSF CAREER award.
Host: Barath Raghavan
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Amir Houmansadr (UMass Amherst) - Communication Secrecy in the Age of AI
Thu, Mar 17, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Amir Houmansadr, UMass Amherst
Talk Title: Communication Secrecy in the Age of AI
Series: CS Colloquium
Abstract: Internet users face constant threats to the secrecy of their communications: repressive regimes deprive them of open access to the Internet, corporations and surveillance organizations monitor their online behavior, advertising companies and social networks collect and share their private information, and cybercriminals hurt them financially by stealing their private information. In this talk, I will present the key research challenges facing communication secrecy in a world overtaken by the AI. In particular, I will introduce new ML-specific mechanisms to defeat AI-enabled surveillance. I will also discuss crucial AI trustworthiness research problems that are essential to the secrecy of Internet communications in the age of AI.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Amir Houmansadr is an associate professor of computer science at UMass Amherst. He received his Ph.D. from the University of Illinois at Urbana-Champaign in 2012, and spent two years at the University of Texas at Austin as a postdoctoral scholar. Amir is broadly interested in the security and privacy of networked systems. To that end, he designs and deploys privacy-enhancing technologies, analyzes network protocols and services (e.g., messaging apps and machine learning APIs) for privacy leakage, and performs theoretical analysis to derive bounds on privacy (e.g., using game theory and information theory). Amir has received several awards and recognitions including the 2013 IEEE S&P Best Practical Paper Award, a 2015 Google Faculty Research Award, an NSF CAREER Award in 2016, a CSAW 2019 Applied Research Competition Finalist, an IMC 2020 Best Paper Award Runner-up, and a Facebook 2021 Privacy Enhancing Technologies Award Finalist. He is an Associate Editor of the IEEE TDSC and frequently serves on the program committees of major security conferences.
Host: Barath Raghavan
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Matthew Mirman (ETH Zürich) - Trustworthy Deep Learning: methods, systems and theory
Thu, Mar 17, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Matthew Mirman , ETH Zürich
Talk Title: Trustworthy Deep Learning: methods, systems and theory
Series: CS Colloquium
Abstract: Deep learning models are quickly becoming an integral part of a plethora of high stakes applications, including autonomous driving and health care. As the discovery of vulnerabilities and flaws in these models has become frequent, so has the interest in ensuring their safety, robustness and reliability. My research addresses this need by introducing new core methods and systems that can establish desirable mathematical guarantees of deep learning models.
In the first part of my talk I will describe how we leverage abstract interpretation to scale verification to orders of magnitude larger deep neural networks than prior work, at the same time demonstrating the correctness of significantly more properties. I will then show how these techniques can be extended to ensure, for the first time, formal guarantees of probabilistic semantic specifications using generative models.
In the second part, I will show how to fuse abstract interpretation with the training phase so as to improve a model's amenability to certification, allowing us to guarantee orders of magnitude more properties than possible with prior work. Finally, I will discuss exciting theoretical advances which address fundamental questions on the very existence of certified deep learning.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Matthew Mirman is a final-year PhD student at ETH Zürich, supervised by Martin Vechev. His main research interests sit at the intersection of programming languages, machine learning, and theory with applications to creating safe and reliable artificial intelligence systems. Prior to ETH, he completed his B.Sc. and M.Sc. at Carnegie-Mellon University supervised by Frank Pfenning.
Host: Mukund Raghothaman
Location: 115
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Jieyu Zhao (UMD) - Building Accountable NLP Models: on Social Bias Detection and Mitigation
Fri, Mar 18, 2022 @ 02:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Jieyu Zhao , UMD
Talk Title: Building Accountable NLP Models: on Social Bias Detection and Mitigation
Series: CS Colloquium
Abstract: Natural Language Processing (NLP) plays an important role in many applications, including resume filtering, text analysis, and information retrieval. Despite the remarkable accuracy enabled by the advances in machine learning used in many applications, the technique may discover and generalize the societal biases implicit in the data. For example, an automatic resume filtering system may unconsciously select candidates based on their gender and race due to implicit associations between applicant names and job titles, causing the societal disparity discovered by researchers. Various laws and policies have been designed to ensure social equality and diversity. However, there is no such mechanism for a machine learning model for sensitive applications. My research analyzes the potential stereotypes in various machine learning models and develops computational approaches to enhance fairness in a wide range of NLP applications. The broader impact of my research aligns with one the concerns of machine learning community: how can we do AI for social good.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Jieyu Zhao is a postdoctoral research at UMD, working together with Prof. Hal Daumé III. She obtained her PhD from the department of Computer Science at UCLA where she was advised by Prof. Kai-Wei Chang. Her research interest lies in fairness of ML/NLP models. Her paper got the EMNLP Best Long Paper Award (2017). She was one of the recipients of 2020 Microsoft PhD Fellowship and has been selected to participate in 2021 Rising Stars in EECS workshop. Her research has been covered by news media such as Wires, The Daily Mail and South China Morning Post. She was invited by UN-WOMEN Beijing on a panel discussion about gender equality and social responsibility. More detail can be found at https://jyzhao.net/.
Host: Xiang Ren
Location: Ronald Tutor Hall of Engineering (RTH) - 105
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Yue Wang (MIT) - Learning 3D representations with minimal supervision
Mon, Mar 21, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Yue Wang , MIT
Talk Title: Learning 3D representations with minimal supervision
Series: CS Colloquium
Abstract: Deep learning has demonstrated considerable success embedding images and more general 2D representations into compact feature spaces for downstream tasks like recognition, registration, and generation. Learning on 3D data, however, is the missing piece needed for embodied agents to perceive their surrounding environments. To bridge the gap between 3D perception and robotic intelligence, my present efforts focus on learning 3D representations with minimal supervision from a geometry perspective.
In this talk, I will discuss two key aspects to reduce the amount of human supervision in current 3D deep learning algorithms. First, I will talk about how to leverage geometry of point clouds and incorporate such inductive bias into point cloud learning pipelines. These learning models can be used to tackle object recognition problems and point cloud registration problems. Second, I will present our works on leveraging natural supervision in point clouds to perform self-supervised learning. In addition, I will discuss how these 3D learning algorithms enable human-level perception for robotic applications such as self-driving cars. Finally, the talk will conclude with a discussion about future inquiries to design complete and active 3D learning systems.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Yue Wang is a final year PhD student with Prof. Justin Solomon at MIT. His research interests lie in the intersection of computer vision, computer graphics, and machine learning. His major field is learning from point clouds. His paper "Dynamic Graph CNN" has been widely adopted in 3D visual computing and other fields. He is a recipient of the Nvidia Fellowship and is named the first place recipient of the William A. Martin Master's Thesis Award for 2021. Yue received his BEng from Zhejiang University and MS from University of California, San Diego. He has spent time at Nvidia Research, Google Research and Salesforce Research.
Host: Ramakant Nevatia
Location: online only
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Erdem Bıyık (Stanford University) - Learning Preferences for Interactive Autonomy
Mon, Mar 21, 2022 @ 02:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Erdem Bıyık , Stanford University
Talk Title: Learning Preferences for Interactive Autonomy
Series: CS Colloquium
Abstract: In human-robot interaction or more generally multi-agent systems, we often have decentralized agents that need to perform a task together. In such settings, it is crucial to have the ability to anticipate the actions of other agents. Without this ability, the agents are often doomed to perform very poorly. Humans are usually good at this, and it is mostly because we can have good estimates of what other agents are trying to do. We want to give such an ability to robots through reward learning and partner modeling. In this talk, I am going to talk about active learning approaches to this problem and how we can leverage preference data to learn objectives. I am going to show how preferences can help reward learning in the settings where demonstration data may fail, and how partner-modeling enables decentralized agents to cooperate efficiently.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Erdem Bıyık is a fifth-year Ph.D. candidate in the Electrical Engineering department at Stanford. He has received his B.Sc. degree from Bilkent University, Turkey, in 2017; and M.Sc. degree from Stanford University in 2019. He is interested in enabling robots to actively learn from various forms of human feedback and designing altruistic robot policies to improve the efficiency of multi-agent systems both in cooperative and competitive settings. He also worked at Google as a research intern in 2021 where he adapted his active robot learning algorithms to recommender systems.
Host: Heather Culbertson
Location: Ronald Tutor Hall of Engineering (RTH) - 105
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Rachee Singh (Microsoft Azure for Operators) - Leveraging over-provisioned cloud networks for next-generation services
Tue, Mar 22, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Rachee Singh , Microsoft Azure for Operators
Talk Title: Leveraging over-provisioned cloud networks for next-generation services
Series: CS Colloquium
Abstract: The last decade has seen a large-scale commercialization of cloud computing and the emergence of global cloud providers. Cloud providers are expanding their datacenter deployments and backbone capacity, preparing their infrastructure to meet the challenges of rapidly evolving workloads in the cloud. In this talk, I will re-examine the design and operation choices made by cloud providers during this phase of exponential growth using a cross-layer empirical analysis of the wide-area network (WAN) of a large commercial cloud provider. Despite their crucial role in enabling high performance cloud applications and expensive infrastructure, there are several inefficiencies in both the design and operation of cloud WANs. In this talk, I will focus on improving the performance and cost efficiency of the fiber optical network underpinning cloud WANs. First, I will demonstrate how rate-adaptive physical links can harness 75% more capacity from 80% of the optical wavelengths in a cloud WAN, leading to a gain of over 100 terabits of network capacity with 25% fewer link failures. Second, I will show how to achieve a 40% reduction in the hardware costs of provisioning long-haul WAN capacity by optically bypassing network hops where conversion of light signals from optical to electrical domain is unnecessary and uneconomical. I will show that identifying and fixing these inefficiencies in today's cloud networks is crucial for enabling next-generation cloud services.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Rachee Singh is a senior researcher in the office of the CTO of Microsoft Azure for Operators. Before this, she was a researcher in the Mobility and Networking group of Microsoft Research, Redmond. Her research interests are in computer networking with a focus on wide area network performance and monitoring. Her PhD dissertation was supported by a Google PhD Fellowship in Systems and Networking and it received the CICS Outstanding dissertation award at the University of Massachusetts, Amherst. Recently, she was named a rising star in computer networking by N2Women and a rising star in EECS by UC Berkeley. In a previous life, she developed routing protocol features for Ethernet switches at Arista Networks.
Host: Ramesh Govindan
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Ariel Barel (Technion, Israeli Institute of Technology) - Applied Deep-Learning methods for Expediting Path Selection in Real-Time MAP
Tue, Mar 22, 2022 @ 01:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Ariel Barel, Technion, Israeli Institute of Technology
Talk Title: Applied Deep-Learning methods for Expediting Path Selection in Real-Time MAP
Series: Computer Science Colloquium
Abstract: Multi-Agent Path Finding (MAPF) is an NP-hard problem that plays a key role in numerous domains ranging from warehouse automation to computer games. In this research we are given a large pre-calculated set of legal paths from all possible sources to all possible destinations. The aim is to select paths from this set such that they do not collide with static obstacles nor with each other and minimize the maximal execution time of all tasks (Makespan). This selection should be calculated in near real-time, i.e., extremely fast compared to classic MAPF algorithms.
We investigate how Deep-Learning methods may speed up the search process, as trained Neural Networks have potential to make computations extremely fast. Training dataset may be generated by solving the "online" problem offline. The idea is to train the network to recognize patterns in the training examples and apply them to new, previously unseen, settings of the problem, i.e., new pairs of sources and destinations. The main challenges are definition of NN architecture and input representation.
This work addresses well-formed environments where agents may wait indefinitely at their sources but must follow a wait-free path once deployed. Moreover, our framework allows assignment of multiple agents per source and requires that all calculations complete before the first deployment, making scheduling a key component of the solution.
This lecture satisfies requirements for CSCI 591: Research Colloquium.
Join Zoom Meeting
https://usc.zoom.us/j/98857434920
Meeting ID: 988 5743 4920
One tap mobile
+16699006833,,98857434920# US (San Jose)
+12532158782,,98857434920# US (Tacoma)
Dial by your location
+1 669 900 6833 US (San Jose)
+1 253 215 8782 US (Tacoma)
+1 346 248 7799 US (Houston)
+1 312 626 6799 US (Chicago)
+1 646 876 9923 US (New York)
+1 301 715 8592 US (Washington DC)
Biography: Dr. Ariel Barel is an academic visitor at the Technion, Israeli Institute of Technology. He received the PhD degrees in Computer Science from the Technion in the field of Distributed Control of Multi-Agent Systems. His current interest also includes Machine Learning implementations to expedite traditional planning algorithms. For more info and publications visit his personal web page https://arielba.cswp.cs.technion.ac.il/
Host: Christopher Leet (cjleet@usc.edu), Sven Koenig (skoenig@usc.edu)
Webcast: https://usc.zoom.us/j/98857434920Location: Online - Zoom
WebCast Link: https://usc.zoom.us/j/98857434920
Audiences: Everyone Is Invited
Contact: Computer Science Department
-
CS Colloquium: Natasha Jacques (Google Brain / UC Berkeley) - Social Reinforcement Learning
Tue, Mar 22, 2022 @ 02:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Natasha Jacques, Google Brain / UC Berkeley
Talk Title: Social Reinforcement Learning
Series: CS Colloquium
Abstract: Social learning helps humans and animals rapidly adapt to new circumstances, coordinate with others, and drives the emergence of complex learned behaviors. What if it could do the same for AI? This talk describes how Social Reinforcement Learning in multi-agent and human-AI interactions can address fundamental issues in AI such as learning and generalization, while improving social abilities like coordination. I propose a unified method for improving coordination and communication based on causal social influence. I then demonstrate that multi-agent training can be a useful tool for improving learning and generalization. I present PAIRED, in which an adversary learns to construct training environments to maximize regret between a pair of learners, leading to the generation of a complex curriculum of environments. Agents trained with PAIRED generalize more than 20x better to unknown test environments. Ultimately, the goal of my research is to create intelligent agents that can assist humans with everyday tasks; this means leveraging social learning to interact effectively with humans. I show that learning from human social and affective cues scales more effectively than learning from manual feedback. However, it depends on accurate recognition of such cues. Therefore I discuss how to dramatically enhance the accuracy of affect detection models using personalized multi-task learning to account for inter-individual variability. Together, this work argues that Social RL is a valuable approach for developing more general, sophisticated, and cooperative AI, which is ultimately better able to serve human needs.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Natasha Jaques holds a joint position as a Senior Research Scientist at Google Brain and Postdoctoral Fellow at UC Berkeley. Her research focuses on Social Reinforcement Learning in multi-agent and human-AI interactions. Natasha completed her PhD at MIT, where her thesis received the Outstanding PhD Dissertation Award from the Association for the Advancement of Affective Computing. Her work has also received Best Demo at NeurIPS, an honourable mention for Best Paper at ICML, Best of Collection in the IEEE Transactions on Affective Computing, and Best Paper at the NeurIPS workshops on ML for Healthcare and Cooperative AI. She has interned at DeepMind, Google Brain, and was an OpenAI Scholars mentor. Her work has been featured in Science Magazine, Quartz, MIT Technology Review, Boston Magazine, and on CBC radio. Natasha earned her Masters degree from the University of British Columbia, and undergraduate degrees in Computer Science and Psychology from the University of Regina.
Host: Jesse Thomason
Location: Ronald Tutor Hall of Engineering (RTH) - 105
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Ram Alagappan (VMware Research Group) - Co-designing Distributed Systems and Storage Stacks for Improved Reliability
Wed, Mar 23, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Ram Alagappan , VMware Research Group
Talk Title: Co-designing Distributed Systems and Storage Stacks for Improved Reliability
Series: CS Colloquium
Abstract: Distributed storage systems form the core of modern cloud services. Like many systems software, these systems are built using layering: designers layer distributed protocols (e.g., Paxos, 2PC) upon local storage stacks. Such layering abstracts details about the local storage stack to the layers above, easing development. I will show that such black-box layering, unfortunately, masks vital information, resulting in poor reliability. I will then demonstrate that reliability can be significantly improved by co-designing these layers.
In the first half of the talk, I will show how local storage-layer faults in one node can lead to serious vulnerabilities such as global data loss, corruption, and unavailability in many widely used systems. I then present CTRL, a new foundation that uses the co-design approach to avoid such problems, improving reliability. I implement CTRL in two practical systems and show that CTRL greatly improves resiliency to storage faults while incurring little performance overhead.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Ram Alagappan is a postdoctoral researcher at the VMware Research Group. He earned his Ph.D., working with Professors Andrea Arpaci-Dusseau and Remzi Arpaci-Dusseau at the University of Wisconsin - Madison. His work has been published at top systems venues and has won three best paper awards (FAST 17, 18, and 20). His dissertation also won an honorable mention for the UW CS Best Dissertation. His open-source frameworks have had a practical impact: these tools have exposed more than 80 severe vulnerabilities across 20 widely used systems. Ideas from his work have been adopted by a financial database to make it more robust.
Host: Ramesh Govindan
Location: online only
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Tesca Fitzgerald (Carnegie Mellon University) - Learning to address novel situations through human-robot collaboration
Thu, Mar 24, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Tesca Fitzgerald, Carnegie Mellon University
Talk Title: Learning to address novel situations through human-robot collaboration
Series: CS Colloquium
Abstract: As our expectations for robots' adaptive capacities grow, it will be increasingly important for them to reason about the novel objects, tasks, and interactions inherent to everyday life. Rather than attempt to pre-train a robot for all potential task variations it may encounter, we can develop more capable and robust robots by assuming they will inevitably encounter situations that they are initially unprepared to address. My work enables a robot to address these novel situations by learning from a human teacher's domain knowledge of the task, such as the contextual use of an object or tool. Meeting this challenge requires robots to be flexible not only to novelty, but to different forms of novelty and their varying effects on the robot's task completion. In this talk, I will focus on (1) the implications of novelty, and its various causes, on the robot's learning goals, (2) methods for structuring its interaction with the human teacher in order to meet those learning goals, and (3) modeling and learning from interaction-derived training data to address novelty.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Dr. Tesca Fitzgerald is a Postdoctoral Fellow in the Robotics Institute at Carnegie Mellon University. Her research is centered around interactive robot learning, with the aim of developing robots that are adaptive, robust, and collaborative when faced with novel situations. Before joining Carnegie Mellon, Dr. Fitzgerald received her PhD in Computer Science at Georgia Tech and completed her B.Sc at Portland State University. She is an NSF Graduate Research Fellow (2014), Microsoft Graduate Women Scholar (2014), and IBM Ph.D. Fellow (2017).
www.tescafitzgerald.com
Host: Heather Culbertson
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Seo Jin Park (MIT CSAIL) - Towards Interactive Big Data Processing Through Flash Burst Parallel Systems
Thu, Mar 24, 2022 @ 02:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Seo Jin Park , MIT CSAIL
Talk Title: Towards Interactive Big Data Processing Through Flash Burst Parallel Systems
Series: CS Colloquium
Abstract: Today, many organizations store big data on the cloud and lease relatively small clusters of instances to run analytics queries, train machine learning models, and more. However, the exponential data growth, combined with the slowdown of Moore's law, makes it challenging (if not impossible) to run such big data processing tasks in real-time. Most applications run big data workloads on timescales of several minutes or hours and resort to complex, application-specific optimizations to reduce the amount of data processing required for interactive queries. This design pattern hurts developer productivity and restricts the scope of applications that can use big data. However, as we have many servers in a cloud datacenter, a natural question is "can we borrow thousands of servers briefly to accelerate big data processing enough to be interactive?"
In this talk, I'll share my vision to enable massively parallel data processing even for very short-duration (1-10 ms), which I call "flash bursts." This will empower interactive, real-time applications (e.g., cyber security attack defense, self-driving cars or drones, etc) to utilize much larger data than before. For this moonshot, I take a two-pronged approach. First, I restructure important big data applications (analytics and DNN training) so that they can run efficiently in a flash burst fashion. On this prong, the talk will focus on how I efficiently scaled distributed sorting to 100+ servers even for a 1-millisecond time budget. Second, I rebuild various layers in distributed systems to reduce overheads of flash burst scaling. On this prong, I will focus on how I removed the overheads of consistent replication.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Seo Jin Park is a postdoctoral researcher at MIT CSAIL. He received a Ph.D. in Computer Science from Stanford University in 2019, advised by John Ousterhout. He is broadly interested in distributed systems, focusing on low-latency systems: scaling low-latency data processing, optimizing consensus protocols (both standard and byzantine), suppressing tail-latencies, and building efficient performance debugging tools. His Ph.D. study was supported by Samsung Scholarship.
Host: Barath Raghavan
Location: Michelson Center for Convergent Bioscience (MCB) - 101
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Feras Saad (MIT) - Scalable Structure Learning and Inference via Probabilistic Programming
Thu, Mar 24, 2022 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Feras Saad, MIT
Talk Title: Scalable Structure Learning and Inference via Probabilistic Programming
Series: CS Colloquium
Abstract: Probabilistic programming supports probabilistic modeling, learning, and inference by representing sophisticated probabilistic models as computer programs in new programming languages. This talk presents efficient probabilistic programming-based techniques that address two fundamental challenges in scaling and automating structure learning and inference over complex data.
First, I will describe scalable structure learning methods that make it possible to automatically synthesize probabilistic programs in an online setting by performing Bayesian inference over hierarchies of flexibly structured symbolic program representations, for discovering models of time series data, tabular data, and relational data. Second, I will present fast compilers and symbolic analyses that compute exact answers to a broad range of inference queries about these learned programs, which lets us extract interpretable patterns and make accurate predictions in real time.
I will demonstrate how these techniques deliver state-of-the-art performance in terms of runtime, accuracy, robustness, and programmability by drawing on several examples from real-world applications, which include adapting to extreme novelty in economic time series, online forecasting of flu rates given sparse multivariate observations, discovering stochastic motion models of zebrafish hunting, and verifying the fairness of machine learning classifiers.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Feras Saad is a PhD candidate in Computer Science at MIT working at the intersection of programming languages, probabilistic machine learning, and computational statistics. His research is accompanied with a collection of popular open-source probabilistic programming systems used by collaborators at Intel, Takeda, Liberty Mutual, IBM, and the Bill & Melinda Gates Foundation for practical applications of structure learning and probabilistic inference. Feras' MEng thesis on probabilistic programming and data science has been recognized with the 1st Place Computer Science Thesis Award at MIT.
Host: Mukund Raghothaman
Location: Ronald Tutor Hall of Engineering (RTH) - 105
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Chuang Gan (MIT-IBM Watson AI Lab) - Neuro-Symbolic AI for Machine Intelligence
Fri, Mar 25, 2022 @ 10:00 AM - 11:00 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Chuang Gan, MIT-IBM Watson AI Lab
Talk Title: Neuro-Symbolic AI for Machine Intelligence
Series: CS Colloquium
Abstract: Machine intelligence is characterized by the ability to understand and reason about the world around us. While deep learning has excelled at pattern recognition tasks such as image classification and object recognition, it falls short of deriving the true understanding necessary for complex reasoning and physical interaction. In this talk, I will introduce a framework, neuro-symbolic AI, to reduce the gap between machine and human intelligence in terms of data efficiency, flexibility, and generalization. Our approach combines the ability of neural networks to extract patterns from data, symbolic programs to represent and reason from prior knowledge, and physics engines for inference and planning. Together, they form the basis of enabling machines to effectively reason about underlying objects and their associated dynamics as well as master new skills efficiently and flexibly.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Chuang Gan is a principal research staff member at MIT-IBM Watson AI Lab. He is also a visiting research scientist at MIT, working closely with Prof. Antonio Torralba and Prof. Josh Tenenbaum. Before that, he completed his Ph.D. with the highest honor at Tsinghua University, supervised by Prof. Andrew Chi-Chih Yao. His research interests sit at the intersection of computer vision, machine learning, and cognitive science. His research works have been recognized by Microsoft Fellowship, Baidu Fellowship, and media coverage from BBC, WIRED, Forbes, and MIT Tech Review. He has served as an area chair of CVPR, ICCV, ECCV, ICML, ICLR, NeurIPS, ACL, and an associate editor of IEEE Transactions on Image Processing.
Host: Ram Nevatia
Location: online only
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Aishwarya Ganesan (VMware Research) - Consistency and Performance in Distributed Storage Systems
Mon, Mar 28, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Aishwarya Ganesan , VMware Research
Talk Title: Consistency and Performance in Distributed Storage Systems
Series: CS Colloquium
Abstract: Talk abstract: Computer systems underpin every modern application that we interact with today. When designing systems, one must often tradeoff strong guarantees for performance or vice-versa. The same tradeoff exists in distributed storage systems as well; designers must often choose consistency or performance. In this talk, I will show how we can build distributed storage systems that provide strong consistency yet also perform well. My key insight to achieving this goal is to defer enforcing consistency until state is externally visible. Based on this insight, I design two novel distributed storage systems.
First, I present Skyros, a new replication protocol that exploits storage-interface properties to defer expensive coordination. Skyros realizes that many update interfaces are nil-externalizing: they do not expose system state immediately. By taking advantage of nil-externality, Skyros offers significantly lower latencies than traditional replication protocols while still providing strong consistency.
Second, I present consistency-aware durability (CAD), a new durability primitive that enables stronger consistency. CAD shifts the point of durability from writes to reads. By delaying writes, CAD enables high performance; however, by ensuring durability before serving reads, CAD enables the construction of stronger consistency models.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Aishwarya Ganesan is a postdoctoral researcher at VMware Research. She completed her PhD from the University of Wisconsin - Madison in Computer Sciences, advised by Andrea Arpaci-Dusseau and Remzi Arpaci-Dusseau. She is broadly interested in distributed systems and storage systems. Her work has been recognized with best-paper awards at FAST 20 and FAST 18 and a best paper award nomination at FAST 17. She was selected for the Rising Stars in EECS workshop and a recipient of Facebook 2019 PhD Fellowship. She also received the graduate student instructor award for teaching graduate-level distributed systems at UW Madison.
Host: Ramesh Govindan
Location: online only
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Souti Chattopadhyay (Oregon State University) - When cognition works against us! Transforming Software to reduce the cost of cognitive processes.
Mon, Mar 28, 2022 @ 02:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Souti Chattopadhyay, Oregon State University
Talk Title: When cognition works against us! Transforming Software to reduce the cost of cognitive processes.
Series: CS Colloquium
Abstract: 86 billion neurons make up our brains! Naturally, these 100 trillion neural connections give rise to a complex process of making decisions, interpreting information, and taking intended actions. This is especially true when programming, whether to build software systems or analyze data. Cognitive processes like selective interpretation and biases affect these programming decisions and actions frequently and significantly. In a recent study, we found that biases are associated with 45.7% of actions that developers take (like editing a line or navigating to a part of code). Eventually, developers reversed or undid 70% of the actions associated with biases which made up 25% of their entire worktime [1]. Similarly, data scientists report spending a lot of time in a "tortuous, multi-step adventure" for getting the data set up for analysis based on familiarity and preferences [2]. Programmers pay the necessary price of being human when working with tools without support for the negative impacts of cognitive processes. In this talk, I will present findings on how some cognitive processes affect programming. To reduce the friction between software and cognition, we will discuss how to design tools to be vigilant and provide desired support using automated and empirical approaches.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Souti Chattopadhyay (Rini) is a Ph.D. candidate at Oregon State University in the Department of EECS. She works at the intersection of Human-Computer Interaction, Software Engineering, and Cognitive Science, focusing on assisting software engineers and data scientists.
Her research is on human-centered tools and interfaces that align with the human cognitive processes when solving problems. Her work is focused on understanding how humans make decisions when interacting with interfaces, specifically programming interfaces. She studies developers, data scientists, and end-user programmers to identify the process behind their technical decisions and social interactions.
During her internship at Microsoft Research, she worked on a project related to the next generation of developers, specifically how they express their identity on social media platforms like YouTube. Some of her works were awarded best papers and honorable mentions by ACM and IEEE, including understanding cognitive biases in programmers and exploring a plethora of challenges data scientists face. Her work on cognitive biases was also recognized as research highlights by CACM and that on data scientists was featured on Nature articles.
Host: Chao Wang
Location: Ronald Tutor Hall of Engineering (RTH) - 105
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Lingjie Liu (Max Planck Institute for Informatics) - Neural Representation and Rendering of 3D Real-world Scenes
Tue, Mar 29, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Lingjie Liu , Max Planck Institute for Informatics
Talk Title: Neural Representation and Rendering of 3D Real-world Scenes
Series: CS Colloquium
Abstract: High-quality reconstruction and photo-realistic rendering of real-world scenes are two important tasks that have a wide range of applications in AR/VR, movie production, games, and robotics. These tasks are challenging because real-world scenes contain complex phenomena, such as occlusions, motions and interactions. Approaching these tasks using classical computer graphics techniques is a highly difficult and time-consuming process, which requires complicated capture procedures, manual intervention, and a sophisticated global illumination rendering process. In this talk, I will introduce our recent work that integrates deep learning techniques into the classical graphics pipeline for modelling humans and static scenes in an automatic way. Specifically, I will talk about creating photo-realistic animatable human characters from only RGB videos, high-quality reconstruction and fast novel view synthesis of general static scenes from RGB image inputs, and scene generation with a 3D generative model. Finally, I will discuss challenges and opportunities in this area for future work.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Lingjie Liu is Lise Meitner Postdoctoral Research Fellow working with Prof. Christian Theobalt in the Visual Computing and AI Department at the Max Planck Institute for Informatics. She received her Ph.D. degree at the University of Hong Kong in 2019. Before that, she got her B.Sc. degree in Computer Science at Huazhong University of Science and Technology in 2014. Her research interests include neural scene representations, neural rendering, human performance modeling and capture, and 3D reconstruction. Webpage: https://lingjie0206.github.io/.
Host: Jernej Barbic
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Roopsha Samanta (Purdue University) - Semantics-Guided Inductive Program Synthesis
Tue, Mar 29, 2022 @ 01:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Roopsha Samanta, Purdue University
Talk Title: Semantics-Guided Inductive Program Synthesis
Series: CS Colloquium
Abstract: The dream of program synthesis seeks to automatically develop programs that conform to a user's intent. Classically, program synthesis has been framed as a problem of generation of correct-by-construction programs from complete, formal specifications of their expected behavior. An increasingly promising and more tractable paradigm of program synthesis, however, is inductive program synthesis. Broadly construed, inductive program synthesis can be framed as a problem of program discovery from partial specifications such as input-output examples, program traces, and natural language descriptions. While the last decade has witnessed several breakthroughs in improving the scalability and applicability of inductive program synthesis, the true potential of this synthesis paradigm remains to be unleashed.
In this talk, I will describe my group's ongoing endeavors to advance the frontiers of inductive program synthesis. Further, I will emphasize the need to tackle a fundamental, yet often neglected, challenge of inductive synthesis-”reliability. Because inductive synthesizers generalize from partial observations, they often suffer from overfitting, ambiguity, and brittleness-”the synthesized program may indeed conform to its partial specification, but it may not exhibit the intended behavior on all inputs. I will present my group's novel semantics-guided approach-”based on surprising notions of program semantics-”to improve the reliability of inductive program synthesis.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Roopsha Samanta is an Assistant Professor in the Department of Computer Science at Purdue University. Before joining Purdue in 2016, she completed her PhD at UT Austin in 2013, advised by E. Allen Emerson and Vijay K. Garg, and was a postdoctoral researcher at IST Austria from 2014-2016 with Thomas A. Henzinger. She is a recipient of 2019 NSF CAREER award and 2021 Amazon Research Award. Roopsha's research seeks to help programmers write programs that conform to their intent. She develop tools and techniques for algorithmic program verification, synthesis, and repair for a spectrum of application domains, correctness specifications, and programmer expertise. Her current research agenda is centered around two themes-”semantics-guided inductive program synthesis and repair and modular, bounded verification of unbounded distributed systems
Host: Mukund Raghothaman
Location: Ronald Tutor Hall of Engineering (RTH) - 105
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Marco Pavone (Stanford University) - Towards safe, data-driven autonomy
Tue, Mar 29, 2022 @ 02:30 PM - 03:50 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Marco Pavone, Stanford University
Talk Title: Towards safe, data-driven autonomy
Series: Computer Science Colloquium
Abstract: *New time: 2:30 - 3:50 PM PT*
AI-powered autonomous vehicles that can learn, reason, and interact with people are no longer science fiction. Self-driving cars, unmanned aerial vehicles, and autonomous spacecraft, among others, are continually increasing in capability and seeing incremental deployment in more and more domains. However, fundamental research questions still need to be addressed in order to achieve full and widespread vehicle autonomy. In this talk, I will discuss our work on addressing key open problems in the field of vehicle autonomy, particularly in pursuit of safe, data-driven autonomy stacks. Specifically, I will discuss (1) robust human prediction models for both simulation and real-time decision making, (2) AI safety frameworks for autonomous systems, and (3) novel, highly integrated autonomy architectures that are amenable to end-to-end training while retaining a modular, interpretable structure. The discussion will be grounded in autonomous driving and aerospace robotics applications.
**Dr. Marco Pavone will give his talk in person at SGM 124 and we will also host the talk over Zoom.**
Register in advance for this webinar at:
https://usc.zoom.us/webinar/register/WN_tR3q2DuTRwulhaNsexKXAw
After registering, attendees will receive a confirmation email containing information about joining the webinar.
This lecture satisfies requirements for CSCI 591: Research Colloquium.
Biography: Dr. Marco Pavone is an Associate Professor of Aeronautics and Astronautics at Stanford University, where he is the Director of the Autonomous Systems Laboratory and Co-Director of the Center for Automotive Research at Stanford. He is currently on a partial leave of absence at NVIDIA serving as Director of Autonomous Vehicle Research. Before joining Stanford, he was a Research Technologist within the Robotics Section at the NASA Jet Propulsion Laboratory. He received a Ph.D. degree in Aeronautics and Astronautics from the Massachusetts Institute of Technology in 2010. His main research interests are in the development of methodologies for the analysis, design, and control of autonomous systems, with an emphasis on self-driving cars, autonomous aerospace vehicles, and future mobility systems. He is a recipient of a number of awards, including a Presidential Early Career Award for Scientists and Engineers from President Barack Obama, an Office of Naval Research Young Investigator Award, a National Science Foundation Early Career (CAREER) Award, a NASA Early Career Faculty Award, and an Early-Career Spotlight Award from the Robotics Science and Systems Foundation. He was identified by the American Society for Engineering Education (ASEE) as one of America's 20 most highly promising investigators under the age of 40.
Host: Stefanos Nikolaidis
Webcast: https://usc.zoom.us/webinar/register/WN_tR3q2DuTRwulhaNsexKXAwLocation: Seeley G. Mudd Building (SGM) - 124
WebCast Link: https://usc.zoom.us/webinar/register/WN_tR3q2DuTRwulhaNsexKXAw
Audiences: Everyone Is Invited
Contact: Computer Science Department
-
CS Colloquium: Nengkun Yu (University of Technology Sydney) - Efficient verification and testing of quantum programs
Wed, Mar 30, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Nengkun Yu , University of Technology Sydney
Talk Title: Efficient verification and testing of quantum programs
Series: CS Colloquium
Abstract: Quantum can solve complex problems that classical computers will never be able to. In recent years, significant efforts have been devoted to building quantum computers to solve real-world problems. To ensure the correctness of quantum programs, we develop verification techniques and testing algorithms for quantum programs. In the first part of this talk, I will overview efficient reasoning about quantum programs by developing verification techniques and tools that leverage the power of Birkhoff & von Neumann quantum logic. In the second part, I will review my work on quantum state tomography, i.e., learning the classical description of quantum states, which closes a long-standing gap between the upper and lower bounds for dynamic testing the properties of a quantum program's output.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Nengkun Yu is an associate professor in the Centre for Quantum Software and Information, the University of Technology Sydney. He received his B.S. and PhD degrees from the Department of Computer Science and Technology, Tsinghua University, Beijing, China, in July of 2008 and 2013. He won a distinguished paper award at OOPSLA 2020 and a distinguished paper award at PLDI 2021. His research interest focuses on quantum computing.
Host: Todd Brun / Jyo Deshmukh
Location: online only
Audiences: By invitation only.
Contact: Assistant to CS chair
-
CS Colloquium: Weihang Wang (State University of New York at Buffalo) - Understanding WebAssembly via Program Transformation
Thu, Mar 31, 2022 @ 11:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Weihang Wang , State University of New York at Buffalo
Talk Title: Understanding WebAssembly via Program Transformation
Series: CS Colloquium
Abstract: WebAssembly is the newest language for the web, aiming to enable high-performance applications and provide languages such as C/C++ a compilation target so that they can be run on the web. WebAssembly defines a portable binary instruction set, as well as a corresponding textual assembly format. However, WebAssembly's syntax is difficult to interpret for human readers because of the stack machine-based implementation. As a result, distributed third-party WebAssembly modules need to be implicitly trusted by developers as verifying the functionality requires significant effort.
In this talk, I will describe my work towards building analysis tools for developers to understand WebAssembly programs. The first section of the talk will focus on identifying limitations of current analysis tools: I will introduce a code obfuscation technique for obfuscating JavaScript malware by translating parts of the computation into WebAssembly. By pinpointing limitations of current malware detectors, my work motivates future efforts on detecting multi-language malware on the web that uses WebAssembly. The second section of the talk will focus on a set of abstraction rules for WebAssembly instructions, which can be used to lift WebAssembly to a high-level representation that abstracts the underlying semantics of the code. I have applied the abstraction rules in detecting WebAssembly-based cryptomining malware. My detection relies on program semantics unique to cryptomining, which is resilient to variants.
This lecture satisfies requirements for CSCI 591: Research Colloquium
Biography: Weihang Wang is an Assistant Professor at the State University of New York at Buffalo. She received her Ph.D. degree in Computer Science from Purdue University in 2018. Weihang's interests are in Software Engineering, with a focus on building tools for improving the reliability and security of software systems. She was awarded an NSF CAREER Award in 2021, a Facebook Testing and Verification Research Award in 2019, a Mozilla Research Award in 2019, and a Maurice H. Halstead Memorial Research Award in 2018.
Host: Chao Wang
Location: Olin Hall of Engineering (OHE) - 132
Audiences: By invitation only.
Contact: Assistant to CS chair
