Grants

Cognitive Models and Strategies for High-Performance Human-AI Teams

A key research challenge in human-AI systems is the current limitation in data, models and theories that explain their dynamic behavior, coordination, and performance. We do not yet fully understand the dominant socio-cognitive processes that determine the dynamic, adaptive, and learning behavior of human-AI teams.  Of especial interest are decision-making problems in intellective tasks with uncertainty and limited resources: what are the rational, efficient, or irrational strategies and heuristics that humans tend to adopt in such circumstances? Useful socio-cognitive models should inform the design of efficient AI agents that improve the overall human-AI team performance. In other words, empirically-validated models and theories are needed to model and build the human-AI teams of the future and to intervene when their performance deteriorates. The project's broad objective is the development and experimental validation of a theory of coordination of human-AI teams in complex intellective tasks.  We plan to combine fundamental insights and models of team behavior from social sciences with state-of-the-art machine learning and dynamical systems methods.  Specifically, our objective include: 1. modeling socio-cognitive structures in human-AI teams, including transactive memory systems, influence systems, and prospect theories; 2. identifying leading cognitive processes, heuristics and biases that underlie the formation of socio-cognitive structures and affect the accuracy of human-AI team decision making; 3. designing supervisory/coordinating AI agents in human-AI teams, based on concepts from applied psychology and machine learning, and testing/validating them in sequential, risky, uncertain decision making tasks; and 4. modeling how human-AI teams cope with limited training data acquired over short sessions, including how they react to various manipulations and intervention schemes.

Funded by: Army Research Office.

MURI: QUANTA: Quantitative Network-based Models of Adaptive Team Behavior

Multidisciplinary University Research Initiative (MURI) on the Network Science of Teams

The recent convergence of research in social and psychological sciences, dynamic and quantitative modeling, and network science has led to a re-examination of collective team behavior from a quantitative and systems-oriented viewpoint. Teams cannot be understood fully by studying their components (members) in isolation: team performance is not simply a sum of individual performances; and a diversity of opinions among members leads to better group outcomes. However, it is not yet understood how patterns of interactions and relationships among team members (i.e. team networks) impact performance. Understanding these patterns is critical, as the resolution of complex issues requires deliberative within-group interaction processes in which alternative courses of action are surfaced, evaluated, and acted upon. This project aims to build quantifiable informative models of teams as dynamical systems interacting over multiple networks, analyze dynamic team behavior by developing rigorous models that relate interaction patterns and network evolution to task performance, and break new ground in team design by scaling teams to solve complex tasks (i.e. teams of teams), and advancing social science theories of team performance. Besides UCSB, our team includes leading scientists from MIT, USC, UIUC, and Northwestern.

Funded by: Army Research Office, Award# W911NF-15-1-0577

Inferring Network Structure and Flow using Partial Observations

Predicting network structure and flows on edges is a precondition to effective planning and disaster response in critical infrastructure networks. Accordingly, we propose to develop algorithms inspired by network science, physical modeling, and machine learning to determine this information from incomplete observations. Using both physical domain-specific and data-driven approaches, these algorithms will address diverse problems including: the reconstruction of network topology and parameters, the estimation of network flows, the optimal selection and scheduling of multiple types of sensors, and the modeling of interdependencies in multilayer networks. Together these methods will develop a rigorous theory for solving the inverse problem of edge and flow determination from observations of critical infrastructure networks.

Funded by: Defense Threat Reduction Agency, Award# HDTRA1-19-1-0017

Modeling Heterogeneity in Dynamic Networks

The proposed research will develop novel methods for analyzing and modeling heterogeneous dynamic networked data. Network data arises in a number of application domains ranging from IoT, cloud computing, software analysis, neuroscience, biology, geography, to social sciences. Accordingly, network analysis has emerged as a major paradigm for exploring complex processes behind observed data. Compared to high dimensional data, analysis of network data is more challenging due to interdependencies between entities, the presence of attributes, and the natural evolution of networks over time. The goal of the project will be to understand and model the heterogeneity of behaviors in dynamic networks. The project will have a transformative impact on big data problems that are enabled by a network-centric approach to exploiting dynamic, heterogeneous data, such as brain networks.

The project’s focus is on heterogeneity in dynamic networks: heterogeneity of node behaviors across network structure and time, heterogeneity of the coupling of structure and attributes, and heterogeneity across networks. Against this backdrop, the project will consider the basic problems of clustering (partitioning), classification/regression, decomposition of networks into its basis elements, and the problem of explaining global network behaviors by small network fragments. These problems will be considered for a single network and for multiple networks. Within a network, heterogeneity is observed when nodes or clusters exhibit different behaviors, for instance due to hidden or missing data. Across networks, heterogeneity is observed in the diversity of subject populations or among network instances. The first research thrust will apply spectral theory for partitioning attributed and dynamic networks. The second research thrust will apply convex optimization to find clusters while tolerating heterogeneity across network structure and time. It will also develop methods for estimating graphical models for multiple dynamic networks. The final research thrust will focus on the discovery of succinct sub-networks that are predictive and that evolve concurrently with the underlying networks.

Funded by: National Science Foundation, IIS, Award# 1817046.

Central Coast Data Science Partnership: Training a New Generation of Data Scientists

Due to the societal and technological advances made possible by data-driven science, there is a strong demand for professionals versed in the tools and techniques needed for manipulating and understanding data. This project will develop an undergraduate curriculum in data science that spans and connects the three main public higher education systems in California: the research-driven University of California system, the practical and career-oriented California State University system, and the two-year California Community Colleges. The collaborative program will establish pathways for data science training through coursework and real-world projects. This project will impact students from diverse social, ethnic, cultural, and economic backgrounds and will improve the feeder pipelines from two-year colleges to four-year universities. This multi-campus approach to building a data science training program will foster collaborations for training a diverse workforce in data science. The resulting course materials and project outcomes will be made available so that other institutions can adopt best practices.

The partnership consists of four academic institutions on the West Coast: University of California, Santa Barbara (UCSB), California Polytechnic State University, San Luis Obispo (Cal Poly), Santa Barbara City College (SBCC), and California State University, San Bernardino (CSUSB). The alliance will expand training at UCSB and Cal Poly by building on existing strengths through a sequence of new capstone courses, as well as lay the groundwork for data science curriculum development at SBCC and CSUSB, whose students will participate in a summer internship program at UCSB. Over 100 undergraduate students will be supported by stipends during the course of the project. The developed courses will emphasize programming and data inference within the context of application domains that is critical to training in data science. Students will be taught the underlying principles of data science, including data-generating processes and the role of measurement, ethics and privacy, information-processing tools for harnessing the power of big data, and the oral and written communication skills necessary for pursuing effective professional careers in the field. The program will culminate in a year-long capstone course for seniors, who will synthesize and apply previously learned data science tools and techniques in a large-scale project in a chosen domain area.

Funded by: National Science Foundation, HDR DSC, Award# 1924205.