Tanya Berger-Wolf - US grants

New technologies for collecting genotypic data from natural populations open the possibilities of investigating many fundamental biological phenomena, including behavior, mating systems, heritabilities of adaptive traits, kin selection, and dispersal patterns. Mining the emerging genotype data for ecological and evolutionary information is one of the most challenging problems in modern biology. Yet full utilization of the genotypic data is only possible if statistical and computational approaches keep pace with our ability to sample organisms and obtain their genotypes. The power and potential of genotypic information often rests in our ability to reconstruct genealogical relationships among individuals. Current computational methods for kinship (lower order pedigree) reconstruction have been developed mainly in the context of human populations. Natural populations pose unique computational and scientific challenges for genetic research: data collection is often limited to a demographic subgroup, such as juveniles; test data for the population under study is rarely available; the number of used genetic markers is relatively small, and typical family sizes can be orders of magnitude larger than in humans. Almost all currently available kinship reconstruction methods are statistical and thus are sensitive to noisy and incomplete data and require a priori knowledge about various parameter distributions, a difficult condition to satisfy in natural populations. The goal of the proposed research is to develop a robust computational method for reconstructing kinship relationships from microsatellite genetic data. The proposed method uses the fundamental genetic laws of inheritance to limit the genetic configurations of possible kinship relationships and powerful optimization techniques to find among those the most parsimonious. The resulting familial reconstruction method requires sampling a minimal number of generations, uses few assumptions about the structure of the data, and relies on little prior knowledge about the sampled population. The diverse tasks of this project include biological modeling, algorithm design and implementation, optimization integration, and experimental validation, many of which may be of use beyond the scope of genetics. The research team will leverage diverse expertise of its members in molecular genetics, mathematical modeling, experimental and theoretical computer sciences to develop accurate and effective methods for familial relationships reconstruction. The proposed interdisciplinary research will have broader impacts on diverse research communities. Improved methods of analysis and inference of kinship relationships open the door to asking new biological questions. The combined advantages of the proposed approach would be applicable to and useful not only for population biology but to various areas of the life sciences, including conservation and management of endangered species, animal behavior, evolutionary genetics, human genealogy, forensics, and epidemiology, any time familial relationships must be inferred from genetic data. The research and software resulting from the proposed project will be disseminated both in computational and biological communities and enhanced by cross-disciplinary training activities. The diverse scientific tasks that comprised the proposed research are suitable for a wide range of students in biology and computer science and will serve to train a new generation of interdisciplinary scientists.

The goal of the proposed research is to create analytical and computational tools that explicitly address the time and order of social interactions between individuals. The proposed approach combines ideas from social network analysis, Internet computing, distributed computing, and machine learning to solve problems in population biology. The diverse computational tasks of this project include design of algorithmic techniques to identify social entities such as a communities, leaders, and followers, and to use these structures to predict social response patterns to danger or disturbances. Nowhere is the impact of social structure likely to be greater than when species come in contact with predators. Thus, the accuracy and predictive power of the proposed computational tools will be tested by characterizing the social structure of horses and zebras (equids) both before and after human- or predator-induced perturbations to the social network. The proposed interdisciplinary research will have broader impacts on a wide range of research communities. New methods for analysis of social interactions in animal populations will be useful for behavioral biologists in such diverse fields as behavioral ecology, animal husbandry, conservation biology, and disease ecology. The machine learning algorithms that will be develop are relevant to many studies in which researchers need to classify temporal interaction data. The proposed network methods have broader relevance to human societies: disease transmission, dissemination of ideas, and social response to crises are all dynamic processes occurring via social networks. Further, through teaching and participation in outreach, students and school teachers will gain access to opportunities for hands-on, interdisciplinary experiences in a new area of computational biology. The research and software resulting from the proposed project will be disseminated both in computational and biological communities and enhanced by cross-disciplinary training activities and will serve to train a new generation of interdisciplinary scientists.

Loyola University, in collaboration with the University of Illinois at Chicago, the University of Illinois at Champaign-Urbana, proposes a planning project for an effort to broaden participation in computing among students in the densely populated Chicago metropolitan area. Proposed activities will focus on a coordinated effort at high school outreach: extensive metro-wide outreach to high schools growing out of a rich network of connections among college and high school students, faculty, and staff that includes an emphasis on interdisciplinary work as well as exposure to varied career paths, and community-building activities. Proposed activities are designed not only to cultivate a larger, stronger, and more diverse corps of computing professionals but also one that is more outward-looking and more service-oriented. This project is seen as the first step in building a larger Alliance.

Computation has fundamentally changed the way we study nature. Recent breakthroughs in data collection technology, such as GPS and other mobile sensors, gene sequencing, and microsatellite genotyping, are giving biologists access to data about wild populations, from genetic to social interactions, that are orders of magnitude richer than any previously collected. Such data offer the promise of answering some of the big questions in population biology: How do animals form social groups and how do genetic ties affect these processes? Which individuals are leaders and to what degree do they control the behavior of others? How do social interactions affect the survival of a species? Unfortunately,in this domain,our ability to analyze data lags substantially behind our ability to collect it. There are three major drawbacks with currently available techniques for analysis of both genotypic and social structure data. First, most traditional methods are aggregate and numeric, thus they are inappropriate for identifying infrequent yet critical events, such as response to predation. Second, the newer approaches focus on human populations and are not directly applicable in the context of wildlife biology. Finally, current analysis techniques are essentially static in that all information about the time and order of social interactions or the concurrency of gene expressions is discarded. Thus, they lack the expressive and computational power to answer the questions outlined above.

Intellectual Merit
The goal of this interdisciplinary research is to develop a robust and scalable computational framework for the emerging field of computational population biology. Ultimately, this research will enable biologists in their scientific inquiry to take advantage of new data by focusing on its underlying qualitative (rather than numerical) and explicitly dynamic structure. This research will use combinatorial techniques to extract that structure. In the scope of this project the following will be developed:
1. Techniques for inferring genetic relationships in wildlife populations and using them to predict genetic diversity.
2. Novel computational methodologies and tools for analyzing dynamic social interactions, focusing on prediction of interaction patterns and dynamic processes within populations.
3. Techniques for combining the genetic and the social structures of a population and across species to identify global ecological processes.

Broader Impacts
Many students, especially female, turn away from computer science in part because of the perceived lack of its applicability to real-world issues and impact on the society. This project has the potential to attract those who would otherwise be lost to computing by providing the view of its larger impact and connection to science. A comprehensive interdisciplinary education and outreach plan will be developed which bridges the traditional pipeline from K-12 to graduate education. The standard views of mathematics and computing will be broadened to include "puzzle-solving" combinatorial thinking by introducing hands-on outreach activities. The unique conflation of wild life biology and computing will continue to be presented at various forums aimed at attracting minorities and girls to science and computer science. Finally, through introduction of biological motivation in computer science courses and the computational methodology in biology courses, this research will provide the students in both disciplines with experiences in asking and answering biological questions by developing new applications of computer science. The methodologies, concepts, and tools developed as part of this interdisciplinary research will be useful to scientists in diverse fields such as behavioral ecology, conservation biology, and disease ecology. Techniques for analysis of social structures have broader relevance to human societies, especially in the context of epidemiology, dissemination of ideas, and crisis management.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

The project consists of an upgrade to part of the campus network and to the campus' network connection to the Starlight facility. This upgrade is designed to support computer science research in networking, visualization and data-mining, primarily at the University's National Center for Data Mining and Electronic Visualization Laboratory. To that end, the upgraded connection to Starlight will be implemented with 100 Gigabit-per-second networking infrastructure.

The renovation will facilitate research on the development and use of cloud computing and cloud data storage. Developing data management and computing services that scale to very large datasets is a fundamental research problem, as is developing services for visualizing and collaboratively analyzing these datasets. The renovated network will also provide researchers on campus with remote access to a new magnetic resonance imaging system as well as facilitating participation in the analysis of data from the Large Hadron Collider.

In addition to providing infrastructure for research, the renovation is likely to enable students from under-represented groups to participate in advanced research, since approximately half of the National Center for Data Mining's Ph.D. students come from underrepresented groups and almost half of the supported research assistants in the Electronic Visualization Laboratory come from underrepresented groups.

This project will address the "frontier problem-solving" aspect missing in interdisciplinary research, and create a model of a course where experts from several disciplines collaborate on a complex project, particularly a field project. Today's science is increasingly an interdisciplinary endeavor, with researchers from several disciplines collaborating on a project. Yes, current graduate (and undergraduate) education is discipline-centered, with few opportunities for team projects, especially from outside the field. As a result, students are unprepared for the realities of cutting edge research. This problem is starting to be addressed by many institutions in a variety of ways. This work will design a highly integrated interdisciplinary field course, centered around computational and field biology research. The proposed course will be offered by the University of Illinois - Chicago, Princeton, and University of Nairobi where graduate students in biology (primarily ecology and evolutionary biology) and engineering (primarily in computer science and bioinformatics) work with faculty in both disciplines to learn how to ask questions, frame hypotheses and understand how and why the disciplines and cultures do this differently. Fieldwork will be conducted in Kenya.

This exploratory project, involving The University of Illinois at Chicago and Northwestern University, investigates the use of sensor-based technologies and general engineering approaches by fourth and fifth grade elementary students, and the effects of that use on how the students formulate research questions in environmental science and biological science and develop domain specific knowledge and concepts.

Teachers, students, and researchers are partnering with the research team to contribute to an iterative process that ensures a diversity of inputs to the approach and design, as they explore opportunities and challenges of using the sensing technologies while learning science. Several research questions are considered in this process and include: Which scientific characteristics are appropriate for elementary school students to grapple with, and which do they struggle with on a conceptual level? Which concepts or processes are more motivating for students? How can an already rich bounty of software technologies for gathering, storing, visualizing, and working with data in instrumented investigations of animal behavior be leveraged? What kinds of new tools are needed to extend those capabilities? How can activities be structured by educators to engage student interest and connect classroom work to field investigations? How can educators and technologists design instruction, materials, and learning technologies in ways that foster students' abilities to formulate scientific questions, choose measures, and plan effective investigations? What pragmatic and content area concerns need to be addressed for teachers to support engineering-enhanced ecological research? The research questions and analysis include observations of small group and classroom discourse, student work products, and reflective grounded interviews to investigate aspects of practice, operationalization of research questions, and examination of research designs, evidence-based argumentation, and explanatory processes.

For sensing technologies and their impact on learning to be fully understood, there are design factors that must be considered. This research is providing the field of learning sciences with some much needed information on design factors that involve sensor-based technologies and domain-based knowledge on scientific practices and engineering approaches to student learning. This interdisciplinary project makes contributions to the fields of learning technologies, engineering education, and biological sciences.

Scalable kinship inference in wild populations across years and generations

A cornerstone of research in molecular ecology is the reconstruction of family groups (kinship analysis).
Understanding how individuals in free-living populations are related to each other provides the best
opportunity to study many important biological processes, ranging from sexual selection to patterns
of dispersal and recruitment. Recent advances in molecular DNA technologies and computational
methods have made these studies possible. However, many conceptual and computational challenges
remain and need to be addressed in order to advance these studies. To date, existing research work
on kinship analysis has primarily focused on computational methods that address a single relationship, such as parentage assignment or reconstruction of full sib groups. Inclusion of multiple objectives, such as half-sib reconstruction with minimum parentage assignment, or hierarchy over multiple generations, makes formulation of the underlying computational problem extremely challenging, and simple extensions of previous methods do not address in a practical, scalable, and robust manner the problem of kinship reconstruction for data sets that include multiple generations of species or involve multiple optimization functions.

The goal of the proposed research is to design robust, parsimonious, and versatile computational
approaches for inferring multi-generation kinship relationships in wild populations from multiallelic
markers. Parsimony assumption is fundamental to these approaches as it requires no prior knowledge,
assumptions about sampling methodology, or existence of models, which is the case for most free-living
populations. The diverse tasks of this project include formulating computational kinship inference
problems based on existing biological studies, analyzing computational complexity of and providing
solutions to the resulting combinatorial optimization problems, and designing robust, scalable and
efficient high performance implementations. The resulting computational methods will be evaluated
on datasets collected from existing biological studies and will be deployed to the biological community
through the Kinalyzer web-based service, currently actively used for sibship inference only.

The research proposed in this project will greatly impact diverse application areas including funda-
mental research in combinatorial optimization and data mining, and within biology, areas as diverse as
behavioral ecology, evolutionary genetics, conservation, forensics, and epidemiology. The multidisci-
plinary nature of the project and the research team will enhance curriculum design of related areas and
introduce new cross-disciplinary courses. This cohesive, multidisciplinary project will provide training
opportunities in biology, operation research, algorithms analysis, bioinformatics and high performance
computing, within a single application framework. The project will leverage the diverse scientific ex-
pertise and extensive mentoring experience of the team to foster a true interdisciplinary collaboration
and to provide a thriving environment for a new generation of interdisciplinary scientists.

This CREATIV award is partially funded by the Networking Technologies and Systems (NeTS) program in the Division of Computer Networks and Systems in the Directorate of Computer & Information Science & Engineering, the Animal Behavior program through the the Divisions of Integrative Organismal Systems and Emerging Frontiers in the Directorate of Biological Sciences, the Office of the Division Director in CISE/CNS and the Office of the Assistant Director in CISE. Despite many years of research on animal interactions, our information about the details of those interactions, especially in large groups of animals, is limited. Fortunately, today's technology can be used to extend our ability to track the social interactions of animals to a much larger scale. While collaborations between social networking and sensor networking researchers have started in the right direction, current approaches have not provided the depth of proximity and orientation information necessary to take the next logical step and infer social interactions between animals. The main challenge lies in the need to balance the accuracy of information about the animal interactions with the energy consumed by the devices themselves, with the ultimate goal of an effective, long running system. To this end, we have designed Mingle, an adaptive sensor-based systems that tracks social interactions between animals. The novelty of Mingle comes from the observation that such social interactions can be tracked by monitoring the animals' relative orientation and relative distance to each other. By relying on local information, Mingle optimizes energy efficiency by integrating local collaborative sensing with the judicious use of infrastructure-based solutions based on observations about the mobility of the animals. Finally, Mingle integrates real application constraints to ultimately drive energy-efficient data collection.

Mingle has the potential to change education, science, and how we view our own society. The ability to see the very detailed social interactions about an entire population of animals will change how we understand and study them. Automating the process of the collection of information about animal, and human, interactions will free behavioral scientists from the collection process while providing data at the level of detail and magnitude never before possible. Moreover, entirely new educational curricula can be designed that engage children in scientific inquiry in fundamentally novel ways. For example, students will have the ability to "become the animals", enacting herding and foraging strategies. Additionally, information about the children's own social interactions will change educational research, enabling our understanding of how children learn in a group and through interactions. While we focus on social interactions between animals in this proposal, the results from this research can be taken into the human social networking domain, where many people already carry sensor-rich smartphones, enabling new and exciting social networking applications for interactions between people, exposing social networking information based on actual social interactions, or measuring social interactions to research social behavior and social patterns.

This project, developing SENSEI (SENSor Environment Imaging), an instrument that targets a broad range of big data science and engineering challenges, promises a unique sensor platform for 2D and 3D recording within dynamic environments. SENSEI contributes to a broad range of data-driven application domains. These range from fundamental research in instrument design and development to data processing, fusion and synthesis, enabling not only the creation of the instrument, but also its use as a resource for multi-domain science and engineering on the ground, in the air, and underwater.

Specifically, SENSEI is a spherical, (ultra) high-resolution (9-times-IMAX resolution and ~terapixel/minute flood of imagery), vision-based capture system capable of video-rate data-acquisition. The instrument will address domain challenges in science, engineering, medicine, and beyond by enabling investigation of big data acquisition, streaming, processing, archiving and access, visualization, and analytics.

The broader significance of this project will be felt in a variety of image-intensive scientific disciplines. The areas of environmental monitoring, remote sensing, situational awareness, homeland security, and mechanical and structural engineering can greatly benefit from the proposed instrument. The instrument will be designed for replication by the global community of researchers and the graduate students. The technology will be communicated through classes, projects, theses, publications, as well as museum exhibits and conferences. Special attention has been paid to broadening participation through the Minority Serving Institutions Cyber-Infrastructure Empowerment Coalition (MSI-CIEC).

This grant provides international and domestic travel support for U.S. based graduate student participants to attend the 20th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2014), which will be held in New York NY, from August 24-27, 2014. KDD is the world's premier research conference in data mining. It is an interdisciplinary conference that brings together researchers and practitioners from all aspects of data mining, knowledge discovery, and large-scale data analytics. This year, the conference has a special theme, Data Mining for Social Good, which will highlight how the work of data analytics researchers and practitioners in contributing towards social good as well as how these high impact, social problems provide a rich set of challenges for KDD researchers. Conference proceedings are published by ACM. Besides the technical program, the conference features workshops, tutorials, panels, demonstrations, exhibits, and a data mining contest (KDD Cup). In 2013 the conference also included a Broadening Participation in Data Mining workshop. A strong representation of U.S. researchers, particularly students, at the conference is useful in maintaining U.S. competitiveness in this important area and also contributes to the career development of the students. This grant will partially support the travel costs for up to 20 U.S. based graduate student participants to attend the KDD conference. The award will be advertised and the results will be announced at the KDD 2014 conference website.

Images are rapidly becoming the most abundant, widely available, and cheapest source of information about the natural world. Images taken by field scientists, tourists, and incidental photographers, and gathered from camera traps and autonomous vehicles provide rich data with the promise of addressing big ecological questions at high resolution and at fine-grained scale. Realizing this potential requires building a large autonomous computational system that starts from image collections and progresses all the way to answering ecological queries, such as population sizes, species distributions and interactions, and movement patterns. The system must have methods of extracting the relevant ecological information from the images and of integrating with other ecological data sources, with minimal human interaction, using state-of-the art information management, computer vision, and data analytics technologies. Such a system will advance computer systems and simultaneously enable ecology to develop as a science of connections across spatial, temporal, and biological scales, as well as provide data- and scientifically-grounded support for ecological decisions.

This work aims to build a prototype of an Image-Based Ecological Information Software System (IBEIS) that relies on a proliferation of images collected daily on a single facility from many different sources, both human and automatic, to determine both the species as well as recognition of distinct individuals. The system will allow for tracking location and movement while providing a data management system that will allow scientists to better understand, and at finer granularity, behaviors and motivations. The system will include: (1) an infrastructure and a mechanism for collecting images from tourists and other sources; (2) a (cloud) infrastructure and a data management system for storing, accessing, and manipulating the images and the derived data; (3) computer vision techniques for extracting information from the images about the identity of individual units, as well as techniques for combining that information with other relevant data to derive information about meaningful ecological units; and (4) statistical techniques and query structures to support ecological queries of the data, such as population sizes and dynamics, movement history and home ranges, and species interactions.

This work will advance computer systems including information management, computer vision, and data analytics technologies, all the while increasing public engagement in science and ecology.

The National Ecological Observatory Network (NEON) is coming online and will provide atmospheric and ecological data locally, regionally and continent wide. At the same time, images are rapidly becoming the most abundant, widely available, and cheapest source of information about the natural world, especially about animals. This project will extend NEON's data, scientific, and citizen science capacity with image-based animal sighting data to scalably collect, manage, and analyze data for individually identifiable wildlife using the Image-Based Ecological Information System (IBEIS) prototype recently developed under another NSF award. Combined with other ecological data, the image data offer the promise of addressing big questions about animal ecology, behavior, and conservation - who? where? when? what? and why? - at high resolution and at fine-grained scale, across landscapes and ecosystems, from an individual animal to regional and global systems. As part of this project, undergraduate and graduate students from ecology and computer science at four institutions will produce and test the application interface, and will develop a suite of companion applications and training tools to allow greater involvement of citizen scientists.

These tools will allow NEON to connect its database to data derived from large volumes of animal photographic images. Although this is primarily a proof of concept proposal focused on connecting whale shark images to NEONs atmospheric data, it will provide the means to be able to apply IBEIS algorithms and databases on images of distinctly marked North American species such as tortoises, monarch butterflies, salamanders, spotted skunk, bobcat, lynx, and humpback whales, thereby connecting these to NEON?s other data streams related to organisms, land use, hydrology and biogeochemistry. The proposed suite of tools includes: 1. an infrastructure and a mechanism for collecting images from scientists, automated remote cameras, citizen scientists and other sources; 2. a data management system for storing, accessing and manipulating images and derived data; 3. computer vision techniques for extracting information from the images about the identity of species and individual animals, as well as techniques for combining that information with other relevant data to derive information about ecological units such as animals, populations, species, and habitats; 4. a software application-program interface integrating the image and derived data with and within NEON; 5. a framework for engaging citizen scientists in data collection, derived science, and interaction with nature. Previous funding from NSF allowed building and testing of an IBEIS prototype. This project will focus on the detection and identification methods for the identifiable US species, on integrating the system with NEON, and on scaling the system to many thousands of daily images from a variety of sources.

Massive amounts of brain imaging data open an unprecedented window into the structure and function of the brain, yet the tools to aid the understanding of the data are lagging. The main goal of this research project is to understand the dynamics of brain function, particularly in the auditory system, through global yet very high spatial and temporal resolution imaging techniques and using the state-of-the-art analytical tools developed for analysis of dynamic social networks. The interdisciplinary neuroscience and computational team, building on promising initial results, will work to adapt these tools developed for understanding human and animal behavior to the context of brain networks and the processes that happen over them. Using this innovative approach, the team will study a particular brain pathway that connects two brain regions that are critical for normal hearing. The project will not only lead to a greater understanding of brain function, but will also bring a new technique to the neuroscience toolbox which may help other investigators to study network properties of the brain. Graduate students and postdocs in computer science and neuroscience will collaborate across disciplinary boundaries, building new scientific approaches and insights.

Top-down projections are ubiquitous in sensory systems and are poorly understood. In the current proposal, a model descending system, the auditory corticothalamic projection in the mouse, will be examined. The research team will take advantage of recent methodological developments in the study of this system and ask: What is the impact of corticothalamic projections on network interactions across populations of thalamic neurons? To answer this, a novel dynamic network analysis method known as Community Dynamic Analysis, or CommDy, will be used to analyze imaging data from a brain slice preparation that retains connectivity between the auditory cortex, auditory thalamus, and other related structures in the mouse. Both calcium imaging data and flavoprotein autofluoresence imaging data will be used for this analysis. Since this study represents the first use of CommDy in neuroscience, validation studies will be done in a simplified brain slice preparation containing bilateral motor cortices and the corpus callosum.

Recent technological advances in location tracking, video and photo capture, accelerometers, and other mobile sensors provide massive amounts of low-level data on the behavior of animals and humans. Analysis of this data can teach us much about individual and group behavior, but analytical techniques that lead to insight about that behavior are still in their infancy. In particular, these new data can provide an unprecedented window into the lives of wild animals, augmenting the traditional time-consuming first-hand observations from field biologists. Unfortunately, the interpretation of low-level (i.e., unprocessed) data from animal-borne electronic sensors still poses a significant bottleneck in leveraging all of the available data to better understand the individual, pairwise, and group behavior of animal populations. This project will develop tools for scaling the expert knowledge needed to interpret high-level behaviors from low-level sensor data using tools from statistical machine learning and network analysis. These data and analytical tools promise to fundamentally change our understanding why animals do what they do, at high resolution and across multiple scales, from individuals to entire populations. The results of the project will be applicable in many settings where massive sensor data is overwhelming traditional insight derived from observational approaches. As part of the project, unique data on primate behavior that will bridge the low-level data and expert knowledge will be collected at Mpala Research Centre, Kenya. Undergraduate, graduate, and postdoctoral students from computer science and animal behavior will collaborate across continental and disciplinary boundaries.

The technical aims of this project include developing structured prediction methods that improve behavior recognition at multiple levels (individual, pair-wise, and group), using network properties to improve the identification of group activities, and advancing active learning in the structured prediction setting so that "expensive" expert knowledge and supplemental data collection will be judiciously utilized for maximum benefit in learning behavior recognition models. Recognizing animal behavior from low-level sensor data is hierarchical in this approach, with individual activities recognized directly from data and the context of these data, the inferred individual activities informing pair-wise behavior recognition, and inferred pair-wise behavior informing group-level activity recognition. The benefits of improving the accuracy of individual and pair-wise behavior for recognizing group-level behavior will enable expert annotations to be requested that improve behavior recognition the most across all levels. These advances will enable field-biologists to investigate new hypotheses about fundamental evolutionary, ecological, and population processes at scale without the burdens of complete manual annotation of collected data. The methods will be applicable beyond field biology to understanding the hierarchy of behavior from individual entities to groups, from humans to cells, in scientific, educational, and business contexts. The team will leverage the interdisciplinary and international nature of the project to continue its ongoing work to increase participation of women and minorities in STEM research at undergraduate and graduate levels.

This project will create an introductory Computer Science course with a Biology theme, by adapting and modifying an existing course developed at a small private college, for use at the University of Illinois at Chicago. The goal of the project is to attract more women to computer science and enable biology students to use the power of computing and computational thinking. Implementing this course at a large scale, in a different academic setting, will result in a version of the course that will be more widely adoptable. A robust dissemination plan will raise awareness of the work, and will result in broader adoption of the course at a national level.

The research component of this project will make significant contributions to the knowledge base about what attracts women and under-represented groups to computing and whether or not the results that were obtained when this course was taught in a small college setting can be replicated at a large public institution. The project will measure student attitudes and capabilities in computing using a variety of methods including the Computer Science Attitudes Survey. The result will be a measure of student confidence learning computer science, their attitude toward success in computer science, their view of computer science as a male domain, their view of the usefulness of computer science, and their motivation in computer science

The project creates a collaborative research institute combining aspects of mathematics, statistics, computer science, and engineering to study the foundations of data science at the University of Illinois at Chicago (UIC). The institute will be a collaboration between three departments: Computer Science (CS), Mathematics, Statistics, and Computer Science (MSCS), and Electrical and Computer Engineering (ECE). The institute will leverage the wide range of expertise among the investigators on this project in the three departments to bring the theoretical foundations of data science closer to the practice of data science. This involves studying idealized models of data, understanding inherent computational limits associated to these idealized models, and then developing models and methods that are robust to realistic models of uncertainty. The institute will also focus on training the next generation of researchers and will leverage the diversity of UIC, a large urban public research-intensive university with one of the most diverse student bodies in the country.

The research aims to push the boundaries of the theory of data science by both gaining deeper understanding of idealized models and by building a theory around realistic models of data and computation. The themes pursued by this institute will include 1) the representation and structure of data; 2) machine learning and complexity; and 3) robustness and privacy. These themes will serve to link the theory and application of data science and to provide opportunities for the investigators to pool their expertise across the three disciplines of theoretical computer science, mathematical sciences, and electrical engineering. The specific activities of the research institute will include hosting themed research workshops, developing the UIC data science curriculum across the three departments, and fostering regional and industrial collaborations through partnerships with the Midwest Big Data Hub and the Discovery Partners Institute. Broader impacts of the institute will include applications of the proposed research to practical data science problems, the development of interdisciplinary data science courses spanning multiple departments, and increasing participation, especially of underrepresented groups, by broadly recruiting students from UIC's diverse community to study data science.

This project is part of the National Science Foundation's Harnessing the Data Revolution (HDR) Big Idea activity.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

The traits that characterize living organisms, in particular, their morphology, physiology, behavior and genetic make-up, enable them to cope with forces of the physical as well as the biological and social environments that impinge on them. Moreover, since function follows form, traits provide the raw material upon which natural selection operates, thus shaping evolutionary trajectories and the history of life. Interestingly, most living organisms, from microscopic microbes to charismatic megafauna, reveal themselves visually and are routinely captured in copious images taken by humans from all walks of life. The resulting massive amount of image data has the potential to further understanding of how multifaceted traits of organisms shape the behavior of individuals, collectives, populations, and the ecological communities they live in, as well as the evolutionary trajectories of the species they comprise. Images are increasingly the currency for documenting the details of life on the planet, and yet traits of organisms, known or novel, cannot be readily extracted from them. Just like with genomic data two decades ago, our ability to collect data far outstrips our ability to extract biological insight from it. The Institute will establish a new field of Imageomics, in which biologists utilize machine learning (ML) algorithms to analyze vast stores of existing image data—especially publicly funded digital collections from national centers, field stations, museums and individual laboratories—to characterize patterns and gain novel insights on how function follows form in all areas of biology to expand our understanding of the rules of life on Earth and how it evolves.

This Institute will introduce structured knowledge from the biological sciences to guide and structure ML algorithms to enable biological trait discovery from images, establishing the field of Imageomics. With images captured and annotated by scientists and the public serving as the basis for the work, the Institute’s convergent approach uses structured biological knowledge to provide scientifically validated inductive biases and rich supervision for ML, and ML will in turn enrich the body of biological knowledge. The resulting ML models and tools will help to make what was hidden visible, so that scientists from a wide range of biological communities can discover and infer the traits of organisms; assess shared similarities and differences between individuals, populations, and species; and come to see the world in new ways. Imageomics will accelerate and transform the biomedical, agricultural, and basic biological sciences as they seek to understand and control genes that relate to specific phenotypes and enable an overarching understanding of how the genome evolved in tandem with the organismal phenome. Because traits are the essential links between genes and the environment, using ML to help characterize them will lead to emergent understandings of how they function. Harnessing the insights that arise from these new visualizations will stimulate the use of new genetic technologies, such as CRISPR gene editing, and more nuanced ecological practices, such as modified land use schemes that emerge from better understanding the connections between individual decision-making within species and their impact on their population dynamics. With the emergence of new and better targeted practices that generate fewer unintended consequences, the new linkages resulting from a better understanding of traits and their consequences will bolster the nation’s bioeconomy. In addition, by leveraging and expanding existing diverse, inclusive and intellectually wide-ranging collaborative networks, the Institute will also educate the next generation of scientists and engage the broader public in scientific inquiry and knowledge discovery so that Imageomics can transform and democratize science for public good.

This project is part of the National Science Foundation's Big Idea activities in Harnessing the Data Revolution (HDR). This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Division of Biological Infrastructure within the NSF Directorate for Biological Sciences, and by the Division of Information and Intelligent Systems within the Directorate for Computer and Information Science and Engineering.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.