CS Colloquium (BMAC)

The Department of Computer Science of Colorado State University, in cooperation with ISTeC (Information Science and Technology Center), offers the CS Colloquium series as a service to all who are interested in computer science. Most seminars are scheduled for Monday 11:00 AM -- 11:50 AM in CSB 130 or Morgan Library Event Hall. For help finding the locations of our seminar meetings, consult the on-line CSU campus

For questions about this page or to schedule talks, please contact Louis-Noel Pouchet (pouchet AT colostate dot edu). Here is a list of past seminar schedules.

CS692 Info: [Link]


Upcoming Events

NOTE: All CS colloquiums scheduled after March 20, 2020 have been cancelled, due to the closure of CSU buildings for the rest of the semester. We will reschedule these talks in Fall 2020.

CS Colloquium Schedule, Spring 2020


cs Computer Science Department Colloquium
Specialized Compilation for High-Performance Computing

Speaker: Louis-Noel Pouchet, Associate Professor, Department of Computer Science, Colorado State University

When: 11:00AM ~ 11:50AM, Monday February 3, 2020
Where: CSB 130 map

Abstract: The computing landscape has become increasingly diverse, driven by power density and energy considerations for embedded, mainstream, and peta ⁄ exascale computing at the high end. Non-homogeneous CPU cores, increasingly complex System-on-Chips as well as domain-specific hardware accelerators have become available platforms for which high-performance execution is demanded. But determining whether the algorithm is at fault or its implementation when facing sub-par performance is a considerably time-consuming task, in turn preventing scientists to quickly evaluate new algorithmic ideas and significantly increasing the development cost of new consumer applications. To alleviate this problem, efforts have been made to create specialized software and hardware stacks, which typically target a particular application domain (e.g., Deep Learning with systems such as TensorFlow and Torch).

In this talk I will present a short overview of the computing landscape and the challenges to generate high-performance implementations. We will discuss our recent efforts to build specialized software stacks for high-performance computing, including how data-specific compilation can be effectively deployed. Data-specific compilation involves optimizing programs not only as a function of the computation pattern implemented, but also as a function of the data on which the program will compute, to deliver additional performance. I will also briefly discuss my path to graduate school, and then to academia.

Bio: Louis-Noel Pouchet is an Associate Professor of Computer Science at Colorado State University, with a Joint Appointment in the Electrical and Computer Engineering department at CSU. He received his Ph.D. in Computer Science in 2010 from University of Paris-Sud 11 and INRIA, France, was a postdoctoral fellow (2010-2012) and then Research Assistant Professor (2014-2016) in the Computer Science and Engineering department at Ohio State University, and was a Visiting Assistant Professor at University of California Los Angeles (2012-2014). He works on a variety of topics for high-performance computing, including optimizing compilers for CPUs, GPUs and FPGAs, domain-specific languages and optimizations, machine learning, high-level synthesis, and distributed systems for large-scale scientific computing. Pouchet is a leading expert in polyhedral compilation, a mathematical framework for program optimization. His research is funded by the U.S. National Science Foundation, the U.S. Department of Energy, and Intel. He is the recipient of an NSF CAREER grant, the William J. McCalla IEEE best paper award at ICCAD’15, and has received several best paper awards at top conferences from the 70+ scientific articles he has co-authored. He is the main author of the PoCC polyhedral compiler, of the ROSE ⁄ PolyOpt software toolset (a complete polyhedral compilation framework for the LLNL ROSE compiler), and of the popular PolyBench ⁄ C benchmarking suite, which has been used in 200+ publications so far.


cs Computer Science Department Colloquium
Talk completed.

Speaker: .

When: 11:00AM ~ 11:50AM, Monday February 10, 2020
Where: CSB 130 map


cs Computer Science Department Colloquium
Creating Literacy and Complexity Profiles for Patients and Physicians from Secure Messages using NLP and machine learning

Speaker: Renu Balyan, Postdoctoral researcher and lecturer, Fulton Schools of Engineering, Arizona State University

When: 10:00AM ~ 10:50AM, Thursday February 13, 2020
Where: CSB 130 map

Abstract: Limited health literacy is a barrier to optimal healthcare delivery and is associated with untoward and costly health outcomes that contribute to health disparities. Given the time and personnel demands intrinsic to current health literacy instruments, combined with the sensitive nature of screening, measuring health literacy has historically been extremely challenging and controversial. Current measures requiring patients to self-report limitations are time-consuming and may be considered intrusive by some. This makes widespread classification of patient health literacy challenging.

In addition, poor communication exchange is an important mediator in the relationship between limited health literacy and health outcomes. Electronic patient portals are an increasingly popular channel for patients and providers to communicate and their secure messaging function provides researchers with the unprecedented opportunity to harness the data generated to create new automated tools to measure health literacy. Therefore, applying natural language processing and machine learning approaches to generate literacy profiles from patients’ and complexity profiles from physicians’ secure messages is a feasible, automated and scalable strategy to identify patients and sub-populations with limited health literacy. These literacy and complexity profiles provide a health IT tool to enable tailored communication support and other targeted interventions with the potential to reduce health literacy-related disparities.

Bio: Renu Balyan has a PhD in the area of natural language processing and holds a Masters in Computer Applications. She is a faculty in the Ira A. Fulton Schools of Engineering at Arizona State University and a research scholar with the Science of Learning and Educational Technology Lab. Dr. Balyan’s research interest focuses on developing and integrating natural language processing and machine learning techniques for applications such as machine translation, summarization, question asking, named entity recognition, sentiment analysis and automatic evaluations, mainly in health and education domain. Balyan is currently working on research projects funded by the National Institute’s of Health (NIH), Office of Naval Research (ONR), and Institute of Education Sciences.

Dr. Balyan has several years of experience working on NLP projects in private industry and government R&D organizations. In academia she has more than nine years of experience teaching computer science courses and mentoring undergraduate and graduate students for their Capstone projects. She has also designed and developed courses for Artificial Intelligence, Natural Language processing and Language Technologies. Balyan has over 30 publications in NLP, health and education journals, conferences and book chapters and serves as a reviewer for journals including Natural Language Engineering (NLE), International Journal of Artificial Intelligence in Education (IJAIED), American Medical Informatics Association (AMIA), and Behavior Research Methods (BRM).


cs Computer Science Department Colloquium
Situated and Interactive Approaches to AI Learning Challenges

Speaker: Nikhil Krishnaswamy, Postdoctoral researcher, Department of Computer Science, Brandeis University

When: 11:00AM ~ 11:50AM, Monday February 17, 2020
Where: CSB 130 map

Abstract: Modern, data-intensive AI, particularly natural language technology, displays a tension between approaches that favor greater amounts of data and approaches that favor better modeling of phenomena of interest. Large amounts of available data and compute power have resulted in systems that can demonstrate apparently sophisticated linguistic understanding or generation, but fail to transfer to situations they have not encountered before. In this talk, I argue that computational "situated grounding" provides a solution to these learning challenges, and to the tension between data and modeling, by creating situational representations that both serve as a formal model of the salient phenomena, and contain rich amounts of exploitable, task-appropriate data for training new, flexible computational models. I will discuss one such approach to situated grounding, created using 3D graphics, natural language processing, and computer vision, and demonstrate its application to create intelligent interactive agents capable of seeing, hearing, and understanding a human interlocutor, and interacting with them in real time. I will discuss how situated grounding can provide a novel approach to three AI learning challenges: learning object affordances, learning novel structures and configurations, and transferring such learned knowledge between domains.

Bio: Dr. Nikhil Krishnaswamy is a postdoctoral researcher in the Department of Computer Science at Brandeis University. His research focuses on the computational understanding of situational and context-dependent linguistic phenomena, including the modeling and representation of object and event properties. With a particular specialization in interactive, intelligent agents, he studies multimodal peer-to-peer communication between humans and machines, building computational implementations of theories of multimodal cognitive processing, implicating not only language, but also gesture, vision, and action. He obtained his Ph.D. in Computer Science at Brandeis researching the creation of visual simulation environments for demonstrating situated computational natural language understanding. He has worked in the defense and video gaming industries, and regularly serves on the program committees of many top-tier conferences in AI, NLP, and cognitive science, including AAAI, ACL, NAACL, EMNLP, and CogSci.


cs Computer Science Department Colloquium
Multi-Modal User Interaction: Gesture + Speech using Augmented Reality Headsets

Speaker: Francisco R. Ortega, Assistant Professor of Computer Science at Colorado State University, Director of the natural user interaction lab (NUILAB)

When: 11:00AM ~ 11:50AM, Monday February 24, 2020
Where: CSB 130 map

Abstract: Multi-modal interaction, in particular gesture and speech, for augmented reality headsets is essential as this technology becomes the future of interactive computing. It is possible that in the near future, augmented reality glasses become pervasive and the preferred device. This talk will concentrate on the motivation behind gesture and speech user interaction, a recent study, and future work. The first part of the talk will describe a study where we demonstrated early and essential findings of gesture and speech user interaction. Findings include types of the gestures performed, the timing between gesturing and speech when used in multi-modality (130 milliseconds), workload (using NASA TLX), and a series of design guidelines resulting from this study. I will also describe the future direction of this research and collaborative multi-modal gesture interaction.

Bio: Dr. Francisco R. Ortega is an Assistant Professor at Colorado State University and Director of the natural user interaction lab (NUILAB). Dr. Ortega earned his Ph.D. in Computer Science (CS) in the field of Human-Computer Interaction (HCI) and 3D User Interfaces (3DUI) from Florida International University (FIU). He also held a position of Post-Doc and Visiting Assistant Professor at FIU between February 2015 to July 2018. Broadly speaking, his research has focused on gesture interaction, which includes gesture recognition and elicitation. His main research area focuses on improving user interaction by (a) eliciting (hand and full-body) gesture sets by user elicitation, and (b) developing interactive gesture-recognition algorithms. His secondary research aims to discover how to increase interest for CS in non-CS entry-level college students via virtual and augmented reality games. His research has resulted in multiple peer-reviewed publications in venues such as ACM ISS, ACM SUI, and IEEE 3DUI, among others. He is the first-author of Interaction Design for 3D User Interfaces: The World of Modern Input Devices for Research, Applications, and Game Development book by CRC Press. Dr. Ortega serves as Vertically Integrated Projects coordinator that promotes applied research for undergraduate students across disciplines.


cs Computer Science Department Colloquium
Multimedia Engages with Database for Video Surveillance

Speaker: Dr. Jianquan Liu, Senior Researcher, the Biometrics Research Laboratories of NEC Corporation

When: 9:00AM ~ 9:50AM, Friday March 6, 2020
Where: CSB 130 map

Abstract: n this talk, Dr. Liu will give an overview of the research topics that are currently conducted at the Biometrics Research Laboratories of NEC Corporation, such as processing and analysis on multimedia big data. The talk will mainly focus on how NEC was ⁄ is engaging techniques of database and multimedia for large-scale video surveillance in the past, present, and future, based on the related researches conducted in NEC.

Representing the past status, he will introduce a commercial level demo system for surveillance video search, named Wally, which was exhibited at MM'14. Wally is a scalable distributed automated video surveillance system with rich search functionalities, and integrated with image processing products developed by NEC, such as NeoFace(R), FieldAnalyst, and StreamPro. Here, NeoFace(R) is one of the best face recognition technologies in the world, having highest recognition accuracy.

Subsequently, he will switch to the present status of video search. The current focus becomes that, the video search process can be triggered without giving any query objects, although the search can be performed automatically based on the analysis of a certain kind of pre-defined patterns. This part will be introduced with a series of work published at SIGGRAPH'16, MM'16, MM'17, ICMR'18, MIPR'19, CBMI'19, BigMM'19, and MM'19. Finally, Dr. Liu will pick up some challenging issues and share some future directions of video search for surveillance, such as the search process by integrating multiple features extracted from surveillance videos, and the key applications of surveillance video search, etc.

Bio: Dr. Jianquan Liu received the M.E. and Ph.D. degrees from the University of Tsukuba, Japan, in 2009 and 2012, respectively. He was a development engineer in Tencent Inc. from 2005 to 2006, and was a visiting researcher at the Chinese University of Hong Kong in 2010. He is currently a senior researcher at the Biometrics Research Laboratories of NEC Corporation, working on the topics of multimedia data processing. He is also an adjunct assistant professor at Graduate School of Science and Engineering, Hosei University, Japan, teaching courses related to data mining and database. His research interests include high-dimensional similarity search, multimedia databases, web data mining and information retrieval, cloud storage and computing, and social network analysis. He has published over 50 papers at major international ⁄ domestic conferences and journals, received over 20 international ⁄ domestic awards, and filed over 40 PCT patents. He also successfully transformed these technological contributions into commercial products in the industry. Currently, he is ⁄ was serving as the PC Co-chair of IEEE ICME 2020, AIVR 2019, BigMM 2019, ISM 2018, ICSC 2018, ISM 2017, ICSC 2017, IRC 2017, and BigMM 2016; the Workshop Co-chair of IEEE AKIE 2018 and ICSC 2016; the Demo Co-chair of IEEE MIPR 2019 and MIPR 2018. He is a member of ACM, IEEE, and the Database Society of Japan (DBSJ), a member of expert committee for IEICE Mathematical Systems Science and its Applications, and IEICE Data Engineering, and an associate editor of IEEE MultiMedia Magazine and the Journal of Information Processing (JIP).


cs Computer Science Department Colloquium
Writing and Reading DNA: Open Problems

Speaker: Jean Peccoud, Professor, Chemical & Biological Engineering, Colorado State University

When: 11:00AM ~ 11:50AM, Monday March 9, 2020
Where: CSB 130 map

Abstract: The possibility to chemically synthetize synthetic DNA molecules and edit existing genomes opens a new field of opportunities for computer scientists. The structure of DNA sequences can be represented using domain specific languages which could be used to build compilers to predict the behavior encoded in the sequence. The development of syntactic models of DNA sequences raises a number of interesting problems. In addition, there is a growing tension in the scientific community between advocates of semantic models derived from the molecular mechanisms regulating processes and advocates of predictive models that lack biological interpretations. Despite rapid progress in DNA sequencing technologies, reading DNA sequences is lagging behind our ability to write DNA sequences. More robust bioinformatics pipelines are needed to streamline the verification of genomic sequences and ensure the security of the research enterprise.

Bio: Dr. Peccoud’s research program combines computational and experimental efforts to develop predictive models of behaviors encoded in DNA sequences. He takes advantage CSU expertise in biomanufacturing to develop a rapid prototyping platform to manufacture biologic drugs, antibodies, and other proteins of commercial interest. Peccoud is also actively engaged in efforts to understand the security implications of synthetic biology.

Shortly after completing a graduate research project in molecular immunology, Jean Peccoud’s scientific interests shifted to computational biology. In 1989, he published one of the first articles describing a mathematical model of molecular noise in gene regulatory networks. In 1993, he recognized the importance of real-time PCR and developed new statistical techniques suitable to analyze this new type of data. In 2002, he observed with excitement the very early developments of synthetic biology. After exploring the potential applications of this new technology to plant biotechnology, he blazed a trail in synthetic biology informatics.

In 2016, Jean Peccoud joined the Department of Chemical & Biological Engineering at Colorado State University where he holds the Abell Chair in Synthetic Biology. He brought with him a diverse experience that includes working for Fortune 500 and start-up companies. He is the founding Editor-in-Chief of the open access journal Synthetic Biology published by Oxford University Press. He is also the CEO of GenoFAB, Inc. a company providing process automation solutions to the biotechnology industry.