All Seminars
| Title: Honors Defense |
|---|
| Defense: Computer Science |
| Speaker: Yilin Dong, |
| Contact: TBA |
| Date: 2020-03-30 at 3:00PM |
| Venue: https://emory.zoom.us/j/785410306 |
 Download Flyer Add to Calendar |
| Abstract: Classification algorithms build models that can classify new observations. While they require a training set of samples' features and labels for training, in reality, many unstructured datasets do not meet the requirement. Since having experts to give out manual labels has a high cost, many industries adopted crowdsourcing, which enables a group of people to contribute to the same labeling task. However, multiple annotations cannot apply to classification algorithms because they assume that labels are single and consensus. In this paper, we use truth inference methods to estimate single labels given different annotations from multiple annotators. While the Expectation-Maximization method provides the best accuracy, our empirical results suggest that better predictive performance can be achieved by accounting for disagreements. Thus, we propose Medaboost, a new predictive model, that considers the degree of disagreements between annotators to improve predictive performance. Medaboost outperforms AdaBoost on both synthetic dataset and MIMIC-III dataset under different sets of simulated nurses’. |
| Title: Autonomic Formation of Large Scale Wireless Mesh Networks |
|---|
| Defense: Computer Science |
| Speaker: Sergio Luis Dias Lima Gramacho, Emory University |
| Contact: Dr. Avani Wildani, agadani@gmail.com |
| Date: 2020-03-27 at 2:00PM |
| Venue: https://emory.zoom.us/j/274041886 |
| Download Flyer Add to Calendar |
| Abstract: A Wireless Mesh Network (WMN) is an appealing network architecture for low-cost and wide geographical coverage. It serves as a potential alternative solution to improve worldwide connectivity in low to high-income countries. Theoretical studies predict, however, insufficient capacity for such network architecture at larger scales. Moreover, the inherently distributed nature of WMNs and their typical distributed network control mechanisms turned them hardened and inflexible to adapt to specific and varying control customization demands. We propose the modernization of the WMN architecture by allowing the general applicability of Software-Defined Networking (SDN) on the implementation of WMN control planes for increased control flexibility while also enforcing frequency diversity to promote throughput capacity. To achieve this, we devised autonomic agents that induce the formation of WMN topologies as a set of interconnected partitions, supporting a cooperative, multi-domain SDN-based WMN control plane able to operate at large-scales and low-cost for increased control flexibility. Moreover, the nature of this autonomic network based on WMN partitions also allows the enforcement of frequency diversity at low-cost and low-complexity for improved throughput capacity. The partitioned topology format is the result of the concurrent and distributed operation of our autonomic agents that manipulate the formation of the WMN using local information, without relying on any central controlling entity, characterizing a scalable and resilient solution. Partitions hold as invariants their maximum diameter and their maximum per node interface degree. These two induce an additional invariant: the maximum partition size in mesh nodes. Finally, the three properties permit limiting control latency and workload on SDN control plane domains. Our agents have different objectives such as organize, heal, optimize; thus, they cooperate and compete to determine final WMN topologies. We show that the competitive/cooperative behavior of these agents converge to stable formations in bounded time even under extreme mesh node churn conditions. The solution relies on an in-network leader election and stochastic delays to achieve the eventual stabilization of formed WMN topologies." |
| Title: On the Analysis of Entanglement Distribution in a Quantum Network |
|---|
| Seminar: Computer Science |
| Speaker: Gayane Vardoyan, |
| Contact: TBA |
| Date: 2020-03-25 at 11:30AM |
| Venue: https://emory.zoom.us/j/702927079 |
| Download Flyer Add to Calendar |
| Abstract: Entanglement is an essential component of quantum computation, information, and communication. Its applications range from quantum key distribution and secret sharing to quantum sensing. These applications drive the increasing need for a quantum switching network that can supply end-to-end entangled states to groups of endpoints that request them. To this end, I study a quantum switch that distributes entanglement to users in a star topology. I will present models for variants of this system and derive expressions for switch capacity and the expected number of qubits stored in memory at the switch. Much of this work focuses on bipartite entanglement switching. For this case, I will discuss how performance metrics are affected by decoherence and link heterogeneity. In this talk, I will also discuss a work wherein we explore a set of switching policies for a switch that can serve both bipartite and tripartite entangled states. I will conclude the talk with a discussion of future research directions and a long-term vision of leveraging tools from performance evaluation to analyze and help guide the design of future quantum networks. |
| Title: How to Program Your Quantum Computer and Get It Right! |
|---|
| Seminar: Computer Science |
| Speaker: Robert Rand, |
| Contact: TBA |
| Date: 2020-03-24 at 10:00AM |
| Venue: https://emory.zoom.us/j/449253098 |
| Download Flyer Add to Calendar |
| Abstract: Quantum programs are hard to write, hard to test and hard to run. In this talk, we show how techniques from programming languages, formal verification and compilation allow us to write quantum programs that are as reliable as the given hardware allows. This provides a path towards writing reliable software for quantum computers, both as we envision them in twenty years and as they exist today. |
| Title: Deep Learning Approaches Towards Computerized Drug Discovery |
|---|
| Defense: Computer Science |
| Speaker: Bonggun Shin, Emory University |
| Contact: Dorian Arnold, dorian.arnold@emory.edu |
| Date: 2020-03-16 at 1:00PM |
| Venue: PAIS Room 235 |
| Download Flyer Add to Calendar |
| Abstract: Proposing a new drug candidate is an essential part of the drug discovery process, consisting of many sub-tasks. Traditionally, these tasks have been tackled by chemistry and pharmaceutical experts and take years to design. Therefore, this thesis aims to accelerate drug discovery by proposing deep-learning models that accomplishes these tasks effectively and quickly. For the target identification problem, we propose new feature selection methods for both disease-related and prognosis-related features. Next, we propose a new drug-target interaction model to perform the drug re-purposing task. In this model, we present a new molecule representation to overcome the limitation of the current models. We also propose a novel drug generation model that can modify an existing drug to meet given molecule properties. For each project, we present an empirical evaluation to show the competency of the proposed approaches. In addition, we also provide analyses or case studies to demonstrate the practicality of our approaches. |
| Title: As We Are: Detecting and Mitigating Human Bias in Visual Analytics |
|---|
| Seminar: Computer Science |
| Speaker: Emily Wall, Georgia Tech |
| Contact: Vaidy Sunderam, vss@emory.edu |
| Date: 2020-02-19 at 11:30AM |
| Venue: Atwood 215 |
| Download Flyer Add to Calendar |
| Abstract: Visual Analytics combines the complementary strengths of humans (perception and sensemaking capabilities) and machines (fast and accurate information processing). However, people are susceptible to inherent limitations and biases, including cognitive biases (e.g., anchoring bias), social biases borne of cultural stereotypes and prejudices (e.g., gender bias), and perceptual biases (e.g., illusions). These biases can impact decision making in critical ways, leading to inaccurate or inefficient choices, or even propagating long-standing institutional and systemic biases. Given our knowledge of these biases and the increased use of data visualization for decision making, the goal of this research is to detect and mitigate human biases in visual data analysis. In this talk, I describe (1) which types of bias are particularly relevant in the process of visual data analysis, (2) how user interactions with data can be used to approximate human biases, and (3) how visualization systems can be designed to increase user awareness of potentially unconscious or implicit biases. By creating systems that promote real-time awareness of bias, people can reflect on their behavior and decision making and ultimately engage in a less-biased decision making process. Bio: Emily Wall is a Computer Science PhD candidate in the School of Interactive Computing at Georgia Tech, where she is advised by Dr.Alex Endert. Her research interests lie at the intersection of cognitive science and data visualization. Particularly, her research has focused on increasing awareness of unconscious and implicit human biases through the design and evaluation of (1) computational approaches to quantify bias from user interaction and (2) interfaces to support visual data analysis. Her research has been supported by NSF and Pacific Northwest National Laboratory. She has been awarded fellowships including Siemens FutureMaker Fellowship, Graduate Fellowship for STEM Diversity, and GA Tech GVU Foley Scholarship, among others. |
| Title: Distinguished Lecturer Series - Bias on the Web |
|---|
| Seminar: Computer Science |
| Speaker: Dr. Ricardo Baeza-Yates, Northeastern University |
| Contact: Eugene Agichtein, eugene.agichtein@emory.edu |
| Date: 2020-02-18 at 10:00AM |
| Venue: White Hall 112 |
| Add to Calendar |
| Abstract: The Web is the most powerful communication medium and the largest public data repository that humankind has created. Its content ranges from great reference sources such as Wikipedia to ugly fake news. Indeed, social (digital) media is just an amplifying mirror of ourselves. Hence, the main challenge of search engines and other websites that rely on web data is to assess the quality of such data. However, as all people has their own biases, web content as well as our web interactions are tainted with many biases. Data bias includes redundancy and spam, while interaction bias includes activity and presentation bias. In addition, sometimes algorithms add bias, particularly in the context of search and recommendation systems. As bias generates bias, we stress the importance of debiasing data as well as using the context and other techniques such as explore & exploit, to break the filter bubble. The main goal of this talk is to make people aware of the different biases that affect all of us on the Web. Awareness is the first step to be able to fight and reduce the vicious cycle of web bias. For more details see the article of same title in Communications of ACM, June 2018. Bio:Ricardo Baeza-Yates is currently CTO of NTENT, a search technology company based in Carlsbad, California, since June 2016; as well as Director of Graduate Data Science Programs (part-time) of Northeastern University, Silicon Valley campus, since January 2018. Previously, he was VP of Research at Yahoo Labs, based in Barcelona, Spain, and later in Sunnyvale, California, from January 2006 to February 2016. Between 2008 and 2012 he also supervised Yahoo Labs Haifa and between 2012 and 2014 Yahoo Labs London. Until 2005 he was the director of the Center for Web Research at the Department of Computer Science of the Engineering School of the University of Chile; and ICREA Professor and founder of the Web Science and Social Computing Research Group (formerly Web Research Group) at the Dept. of Information and Communication Technologies of Universitat Pompeu Fabra in Barcelona, Spain. He maintains ties with both mentioned universities as a part-time professor. Finally, he is also an adjunct professor at the CS department of the University of Waterloo, Canada. He is an author of one of the most popular books on information retrieval and web search: Modern Information Retrieval, co-authored with Berthier Ribeiro-Neto: http://grupoweb.upf.es/mir2ed/ Ricardo is ACM Fellow and IEEE Fellow. More information is available at http://www.baeza.cl and from Wikipedia: https://en.wikipedia.org/wiki/Ricardo_Baeza-Yates |
| Title: Towards Human-Centric Intelligent Systems: An Interactive Online Learning Approach. |
|---|
| Seminar: Computer Science |
| Speaker: Qingyun Wu, University of Virginia |
| Contact: Vaidy Sunderam, vss@emory.edu |
| Date: 2020-02-17 at 11:30AM |
| Venue: MSC W307C |
| Download Flyer Add to Calendar |
| Abstract: The past several years have witnessed a growing need for intelligent systems, such as recommender systems, and smart control systems, that work in real-time to satisfy people's various needs. The rapid appearance of new information, together with the ever-changing nature of the real-world environment, urges us to move from the passive learning paradigm to a more interactive and proactive one. In this talk, I will talk about a new interactive online learning paradigm based on multi-armed bandits and contextual bandits for human-centered intelligent systems. Interactive online learning solutions explore the unknowns by sequentially collect individual user's feedback, which helps address the notorious explore/exploit dilemma during sequential decision making. I will introduce our most recent development in collaborative multi-armed bandits and non-stationary bandits, which enable efficient interactive online learning in dynamically changing and potentially collaborative or structured real-world environments. Bio: Qingyun Wu is a Ph.D. candidate in the Department of Computer Science, University of Virginia. Her research interests are in interactive online learning, including multi-armed bandit, reinforcement learning, and their applications to intelligent systems such as recommender systems, online learning to rank. Qingyun has received multiple prestigious awards from the University of Virginia, including the Virginia Engineering Foundation Fellowship and the Graduate Student Award for Outstanding Research at the Computer Science Department. Her recent work on dueling bandit based online learning to rank won the Best Paper Award of SIGIR 2019. She was also selected as one of the Rising Stars in EECS 2019. |
| Title: Online Credibility, Conspiracies, and Extremism: Understanding Problematic Content on Social Media Platforms |
|---|
| Colloquium: Computer Science |
| Speaker: Tanu Mitra, Virginia Institute of Technology |
| Contact: Vaidy Sunderam, vss@emory.edu |
| Date: 2020-02-14 at 10:30AM |
| Venue: MSC W201 |
| Download Flyer Add to Calendar |
| Abstract: Online social media platforms have brought numerous positive changes, including access to vast amounts of news and information. Yet, those very opportunities have created new challenges—our information ecosystem is now rife with problematic content, ranging from misinformation, conspiracy theories, to hateful and incendiary propaganda. My research introduces computational methods and systems to understand and design defenses against such problematic online content. In this talk, I will focus on three aspects of problematic online information: 1) non-credible content, 2) conspiracy theories, and 3) extremist propaganda. First, I will present the development and analysis of a large-scale, systematic social media credibility corpus, called CREDBANK. With CREDBANK's 66M tweets nested in 1,377 real-world events, I will show that temporal and linguistic regularities can differentiate credible and non-credible information. Second, leveraging 10 years of discussion data spanning millions of conspiratorial discussion posts on Reddit, I will present scalable methods to automatically detect the recurring elements underlying these discussions and ways to unravel what causes users to join conspiratorial communities. Third, I will dive into a special type of problematic content: narratives of extremists hate groups. Merging framing theory from social movement research with big data analyses, I will discuss the ecosystem of cross-platform communication by hate groups. Finally, I will close by previewing important new opportunities I see my lab tackling in the next few years to address some of these problems, including conducting social audits to defend against algorithmically generated misinformation and designing socio-technical interventions and systems to promote online tolerance, civility, and trust. |
| Title: Connectivity in Complex Networks: Measures, Inference and Optimization |
|---|
| Seminar: Computer Science |
| Speaker: Chen Chen, PhD, Google |
| Contact: Vaidy Sunderam, vss@emory.edu |
| Date: 2020-02-13 at 10:00AM |
| Venue: MSC W507 |
| Download Flyer Add to Calendar |
| Abstract: Networks naturally appear in many high-impact applications, ranging from epidemic studies, social network mining to infrastructure analysis. The simplest model of networks is single-layered network, where nodes are from the same domain and links are of the same type. However, as the world is becoming increasingly connected and coupled, nodes from different application domains tend to be interdependent on each other, forming a more complex network model called multi-layered networks. Among the various aspects of network studies, network connectivity is the one that plays a foundational role in a myriad of tasks (e.g. information dissemination, robustness analysis, and community detection). In this talk, I will present my research about the connectivity measures, inference and optimization problems in complex networks. Specifically, I will introduce (1) a unified framework to measure the connectivity in complex network systems; (2) effective connectivity inference methods for networks under dynamic and incomplete settings; and (3) theoretical analysis and approximation algorithms for the generalized connectivity optimization problems. |