Home / Research Interests / Publications / Prospectve Students


What We Do

Many of the pressing challenges facing contemporary society concern sustainability and public health. For example, how can sustainable behaviors—such as reducing individual energy consumption—be encouraged? How can participation in activities that reduce overall healthcare costs—such as compliance with preventive care routines and leading healthy lifestyles—be supported? Common to these challenges is a fundamental question: how can we facilitate cooperative behavior adoption on a large scale?

Facilitating cooperation in large populations is qualitatively different from the problem of facilitating cooperation in small groups. The work of 2009 Nobel Laureate Elinor Ostrom has addressed the latter question: small homogenous groups with strong trust relationships can overcome “the tragedy of the commons” and manage themselves sustainably. However, the conditions for self-governance found in small groups do not apply in large populations. As a result, the question of how cooperation can be facilitated in large populations remains unanswered and is the focus of my work. Computing has a critical role to play in answering this question.

My research group is devoted to addressing issues of sustainability by identifying mechanisms to engender cooperative behavior adoption in very large groups. We develop computing artifacts—algorithms, integrative frameworks, efficient sensors and novel applications. We are developing an integrative framework that analyzes social signals (communication and information flow) across networks of people and senses their physical activity in the everyday through lightweight efficient sensors and smartphones. Analysis of communication patterns in networks of people helps us understand the crucial role played by social ties when people take decisions to adopt behaviors. We can begin also to observe and analyze the context in which the behaviors occur by attaching networked sensors to everyday objects. By coupling the analysis of communication with the analysis of actual behavior we derive insights into how to effectively shape and mold individual choices.

My group is actively engaged in addressing these challenges by solving fundamental problems in engineering and data science. The engineering projects include detecting small homogenous groups with algorithms grounded in social science (the first step towards facilitating cooperation within large heterogeneous populations is to unite similar individuals within the population into small homogenous groups); detecting the onset of large scale social coordination; developing smartphone applications for analysis of cooperation in the contexts of sustainability, productivity and health; developing approximate pattern analysis and sampling techniques for low-power sensors that achieve an order of magnitude power savings. In data science, we have designed new sampling methods for graphs to estimate behavior adoption (there is no Nyqvist criteria for sampling content on graphs); and we have developed a near real-time algorithm that detects rapid large scale changes to network structure (e.g. due to viral behavior adoption). We are breaking new ground by investigating the problem of large-scale behavior adoption through the lenses of information theory and the physics of materials.

We have also developed system-level control algorithms for mediated environments where we can teach people physical skills. We have a developed an experimental physical world, rich in sensors and media feedback, where we can observe and influence human activity. Stroke patient rehabilitation, is an example. In this work, the goal is to teach stroke survivors, how to reacquire body function, through interaction with the environment. In contrast to machine learning, where the goal is to develop algorithms that discover patterns from archived data, our framework can be termed as machine teaching, where the goal is to actively teach physical skills to human beings.

Our research has received several best paper and best student paper awards from the IEEE and the ACM. Our paper on tag recommendations based on social network data received the best student paper award at IEEE/ACM JCDL 2007. Our work on approximate transforms for low-power sensors was best student paper runner up at IEEE ICASSP 2006. Our work on system control applied to mediated rehabilitation was a best paper runner-up at the prestigious ACM Multimedia 2007 conference. My work on video summarization (best student paper, ACM Multimedia 2002), and video retrieval (best paper, IEEE Trans. On CSVT, 2000), has also been acclaimed.

Addressing societal challenges requires that multiple disciplines come together and inform one another. The social sciences, engineering and communication have informed my models. By coupling computing with insights from the social sciences, I ensure the validity of my models and endeavor to bring us closer to addressing these societal challenges. My active collaborators span the arts (Thanassis Rikakis, Todd Ingalls, David Tinapple), computing (Aviral Shrivastava, K. Selcuk Candan, Daragh Byrne, Pavan Turaga, Lexing Xie, Min-Yen Kan, Vikram Jandhyala, Ashutosh Sabharwal), design (Aisling Kelliher), innovation (Ronnie Chatterji) anthropology (Marco Janssen) and public policy (Erik Johnston). Along with David Tinapple and Daragh Byrne, I co-lead Reflective Living an interdiscplinary research group that weaves together ideas from art, design and computing.

Our work has been generously supported by the NSF and by Industry— IBM, Microsoft, NEC and Avaya. Below, I list active NSF grants and NSF grants that are to end in 2012.

Active NSF grants
  • SoCS: Tipping Collective Action in Social Networks,PI: Marco Janssen, Co-PI: Hari Sundaram, Allen Lee.
  • RanKloud: Data Partitioning and Resource Allocation Strategies for Scalable Multimedia and Social Media Analysis, PI: Selcuk Candan, Co-PI: Hari Sundaram, Maria-Luisa Sapino.
  • RAPID VOSS: Understanding the challenges inherent in the design, execution and participation in governance challenge platforms. PI: Erik Johnston. Other co-PI’s: Hari Sundaram, Spiro Maroulis, Marco Janssen, and John Anderies.
  • A virtual eXchange to support networks of creativity and innovation amongst Science, Engineering, Arts and Design (XSEAD), PI, Thanassis Rikakis, Co-PI: Hari Sundaram
  • IGERT: An Arts, Sciences and Engineering Research and Education Initiative for Experiential Media, PI: Thanassis Rikakis. Co-PI's: Hari Sundaram, Andreas Spanias, Jiping He, Michael McBeath, Wilhelmina Savenye. (Ends in 2012)
  • Design of Dense RFID Systems for Indexing in the Physical World across Space, Time, and Human Experience, PI: Hari Sundaram, Co-PI: Selcuk Candan, Vikram Jandhyala (UW) (Ends in 2012)
  • MiNC: NSDL Middleware for Network- and Context-aware Recommendations, PI: K. Selcuk Candan. Co-PI's: Hari Sundaram, Hasan Davulcu (Ends in 2012)