Toyota Technological Institute At Chicago

NEW COMPUTATIONAL METHODS FOR DATA-DRIVEN PROTEIN STRUCTURE PREDICTION

THE PURPOSE OF THIS AGREEMENT IS TO FUND RESEARCH SUPPORTING THE DEFENSE ADVANCED RESEARCH PROJECTS AGENCY DARPA GUARANTEEING AI ROBUSTNESS AGAINST DECEPTION GARD PROGRAM. THE TERM OF THE BASE PERIOD FOR THIS AGREEMENT COMMENCES ON THE EFFECTIVE DATE OF THE AWARD AND CONTINUES THROUGH TWELVE 12 MONTHS THEREAFTER.

INSTITUTE FOR DATA, ECONOMETRICS, ALGORITHMS AND LEARNING (IDEAL) -THE INSTITUTE FOR DATA, ECONOMETRICS, ALGORITHMS, AND LEARNING (IDEAL) WILL CONSOLIDATE AND AMPLIFY RESEARCH DEVOTED TO THE FOUNDATIONS OF DATA SCIENCE ACROSS ALL THE MAJOR RESEARCH-FOCUSED EDUCATIONAL INSTITUTIONS IN THE GREATER CHICAGO AREA: THE UNIVERSITY OF ILLINOIS AT CHICAGO, NORTHWESTERN UNIVERSITY, THE TOYOTA TECHNOLOGICAL INSTITUTE AT CHICAGO, THE UNIVERSITY OF CHICAGO, AND THE ILLINOIS INSTITUTE OF TECHNOLOGY. THIS TRANSDISCIPLINARY INSTITUTE INVOLVES OVER 50 RESEARCHERS WORKING ON KEY ASPECTS OF THE FOUNDATIONS OF DATA SCIENCE ACROSS COMPUTER SCIENCE, ELECTRICAL ENGINEERING, MATHEMATICS, STATISTICS, AND SEVERAL RELATED FIELDS LIKE ECONOMICS, OPERATIONS RESEARCH, AND LAW, AND THEY ARE COMPLEMENTED BY MEMBERS OF GOOGLE?S LEARNING THEORY TEAM. ITS RESEARCH GOALS RANGE FROM THE CORE FOUNDATIONS OF DATA SCIENCE TO ITS INTERFACES WITH OTHER DISCIPLINES: 1) TACKLING IMPORTANT CHALLENGES RELATED TO FOUNDATIONS OF MACHINE LEARNING AND OPTIMIZATION, 2) ADDRESSING STATISTICAL, ALGORITHMIC AND MATHEMATICAL CHALLENGES IN DEALING WITH HIGH-DIMENSIONAL DATA, AND 3) EXPLORING THE FOUNDATIONS OF ASPECTS OF DATA SCIENCE THAT INTERACT WITH SOCIETY. THE INSTITUTE WILL FOSTER STRONG CONNECTIONS WITH THE COMMUNITY AND LOCAL HIGH SCHOOLS, BROADEN PARTICIPATION IN DATA SCIENCE LOCALLY AND NATIONALLY, AND BUILD LASTING RESEARCH AND EDUCATIONAL INFRASTRUCTURE THROUGH ITS ACTIVITIES. INSTITUTE ACTIVITIES WILL INCLUDE WORKSHOPS FOR UNDERGRADUATE STUDENTS, HIGH SCHOOL TEACHER WORKSHOPS, PUBLIC LECTURES, AND MUSEUM EXHIBIT DESIGNS. THESE WILL BUILD NEW PATHWAYS FOR UNDERGRADUATE STUDENTS, HIGH SCHOOL STUDENTS, AND THE BROADER PUBLIC FROM DIVERSE AND UNDERREPRESENTED BACKGROUNDS, TO INCREASE PARTICIPATION AND ENGAGEMENT WITH SCIENTIFIC FIELDS RELATED TO DATA SCIENCE. THE RESEARCH THRUSTS OF THE INSTITUTE WILL CENTER AROUND THE FOUNDATIONS OF MACHINE LEARNING, HIGH-DIMENSIONAL DATA ANALYSIS AND INFERENCE, AND DATA SCIENCE AND SOCIETY. SPECIFIC TOPICS INCLUDE FOUNDATIONS OF DEEP LEARNING, REINFORCEMENT LEARNING, MACHINE LEARNING AND LOGIC, NETWORK INFERENCE, HIGH-DIMENSIONAL DATA ANALYSIS, TRUSTWORTHINESS & RELIABILITY, FAIRNESS, AND DATA SCIENCE WITH STRATEGIC AGENTS. THE RESEARCH ACTIVITIES ARE DESIGNED TO FACILITATE COLLABORATION BETWEEN THE DIFFERENT DISCIPLINES AND ACROSS THE FIVE CHICAGO-AREA INSTITUTIONS, AND THEY BUILD ON THE EXTENSIVE EXPERIENCE FROM PREVIOUS EFFORTS OF THE PARTICIPATING UNIVERSITIES. THE ACTIVITIES INCLUDE TOPICAL SPECIAL PROGRAMS, POSTDOCTORAL FELLOWS, CO-MENTORED PHD STUDENTS, WORKSHOPS, COORDINATED GRADUATE COURSES, VISITING FELLOWS, RESEARCH MEETINGS, AND BRAINSTORMING SESSIONS. THE PROPOSED RESEARCH WILL LEAD TO NEW THEORETICAL FRAMEWORKS, MODELS, MATHEMATICAL TOOLS AND ALGORITHMS FOR ANALYZING HIGH-DIMENSIONAL DATA, INFERENCE AND LEARNING. SUCCESSFUL OUTCOMES WILL ALSO LEAD TO A BETTER UNDERSTANDING OF THE FOUNDATIONS OF DATA SCIENCE AND MACHINE LEARNING IN BOTH STRATEGIC AND NON-STRATEGIC ENVIRONMENTS ? INCLUDING EMERGING CONCERNS LIKE RELIABILITY, FAIRNESS, PRIVACY AND INTERPRETABILITY AS DATA SCIENCE INTERACTS WITH SOCIETY IN VARIOUS WAYS. THE INSTITUTE WILL ALSO HAVE BROADER IMPACTS OF STRENGTHENING RESEARCH AND EDUCATIONAL INFRASTRUCTURE, DEVELOPING HUMAN RESOURCES, BROADENING PARTICIPATION FROM UNDERREPRESENTED GROUPS, AND BY CONNECTING THEORY TO SCIENCE AND INDUSTRY. THE INSTITUTE WILL ORGANIZE ACTIVITIES TO ENGAGE THE COMMUNITY AND A DIVERSE GROUP OF STUDENTS AT ALL LEVELS, INCLUDING INTRODUCTORY WORKSHOPS FOR UNDERGRADUATE RESEARCH PARTICIPANTS, HIGH SCHOOL STUDENT AND TEACHER OUTREACH (THROUGH A PARTNERSHIP WITH THE MATH CIRCLES OF CHICAGO), AND PUBLIC LECTURES AS PART OF BOTH OUR RESEARCH PROGRAM AND A PARTNERSHIP WITH THE MUSEUM OF SCIENCE AND INDUSTRY. THE CHICAGO PUBLIC INSTITUTIONS THAT WE ENGAGE SERVE A VERY DIVERSE POPULATION, SO THE OUTREACH, RECRUITMENT, AND TRAINING ACTIVITIES WILL BROADEN PARTICIPATION FROM UNDERREPRESENTED GROUPS. FINALLY, THE INSTITUTE WILL HAVE DIRECT ENGAGEMENT WITH APPLICATIONS AND INDUSTRY THROUGH ITS ACTIVITIES INVOLVING GOOGLE, OTHER INDUSTRY PARTNERS IN THE BROADER CHICAGO AREA, AND APPLIED DATA SCIENCE INSTITUTES. 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.

RI: AF: MEDIUM: LEARNING AND MATRIX RECONSTRUCTION WITH THE MAX-NORM AND RELATED FACTORIZATION NORMS

RI: MEDIUM: COLLABORATIVE RESEARCH: MODELS OF HANDSHAPE ARTICULATORY PHONOLOGY FOR RECOGNITION AND ANALYSIS OF AMERICAN SIGN LANGUAGE

AF: SMALL: FOUNDATIONS FOR SOCIETAL MACHINE LEARNING -MACHINE LEARNING HAS BECOME A HIGHLY SUCCESSFUL AND PRACTICAL TOOL FOR UNDERSTANDING DATA, ENABLING NEW TECHNOLOGIES, AND AIDING HUMAN DECISION-MAKING. HOWEVER, ITS INCREASED USE IN APPLICATIONS THAT IMPACT PEOPLE HAS ALSO LED TO A NUMBER OF CONCERNS. THESE INCLUDE CONCERNS ABOUT THE FAIRNESS OF DECISIONS MADE, CONCERNS ABOUT INCENTIVES GENERATED AND THE EFFECT OF STRATEGIC BEHAVIOR ON ACCURACY OF THESE SYSTEMS, AND CONCERNS ABOUT THE IMPACT OF CLASSIFICATION DECISIONS ON SOCIETAL WELFARE. THIS PROJECT AIMS TO DEVELOP THEORETICAL FRAMEWORKS THAT ADVANCE THE FOUNDATIONS FOR MACHINE LEARNING SYSTEMS THAT ADDRESS THESE CONCERNS. IN PARTICULAR, THE HIGH-LEVEL GOAL OF THIS WORK IS TO BE ABLE TO PROVIDE CLEAN GUARANTEES BOTH TO THOSE USING THESE SYSTEMS AND TO THOSE AFFECTED BY THE DECISIONS THEY MAKE. SPECIFICALLY, THIS PROJECT IS CENTERED AROUND THREE MAIN RESEARCH DIRECTIONS. THE FIRST IS TO ADVANCE THE UNDERSTANDING OF FAIRNESS IN MACHINE-LEARNING AND ALGORITHMIC CONTEXTS, WITH EMPHASIS ON THE INTERACTION BETWEEN FAIRNESS CONDITIONS AND BIASED TRAINING DATA, AND ON IMPLEMENTING FAIRNESS CONDITIONS IN MULTI-STAGE DECISION SYSTEMS. THE SECOND DIRECTION INVOLVES STRATEGIC CLASSIFICATION, WHICH IS THE PROBLEM OF MAKING CLASSIFICATION DECISIONS ON AGENTS THAT HAVE THE ABILITY TO MODIFY THEIR OBSERVABLE FEATURES TO A LIMITED EXTENT, AND WHO MAY DO SO IF IT LEADS TO A DECISION THEY PREFER. THIS WORK WILL TACKLE A NUMBER OF FUNDAMENTAL PROBLEMS IN THE DESIGN OF ALGORITHMS WITH PROVABLE ACCURACY GUARANTEES IN SUCH SETTINGS, ESPECIALLY FOR THE CHALLENGING CASE OF ONLINE SEQUENTIAL DECISION-MAKING. THE THIRD DIRECTION INVOLVES IMPACTS ON SOCIETAL WELFARE, AND DEVELOPMENT OF LEARNING ALGORITHMS THAT COMBINE CLASSIC ACCURACY GOALS WITH GOALS THAT INVOLVE INCENTIVIZING SOCIETALLY-BENEFICIAL BEHAVIORS. 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.

ABI DEVELOPMENT: DEVELOPING RAPTORX WEB PORTAL FOR PROTEIN STRUCTURE AND FUNCTIONAL STUDY

ABI DEVELOPMENT: CONTINUED DEVELOPMENT OF RAPTORX SERVER FOR PROTEIN STRUCTURE AND FUNCTIONAL PREDICTION

AF: RI: MEDIUM: COLLABORATIVE RESEARCH: UNDERSTANDING AND IMPROVING OPTIMIZATION IN DEEP AND RECURRENT NETWORKS

CAREER: UNDERSTANDING POLYNOMIAL STRUCTURE ANALYTICALLY AND ALGORITHMICALLY

HDR TRIPODS: COLLABORATIVE RESEARCH: INSTITUTE FOR DATA, ECONOMETRICS, ALGORITHMS AND LEARNING

CAREER: EXACT AND APPROXIMATE ALGORITHMS FOR 3D STRUCTURE MODELING OF PROTEIN-PROTEIN INTERACTIONS

CAREER: METRIC GEOMETRY TECHNIQUES FOR APPROXIMATION ALGORITHMS

AF: SMALL: PARALLELS IN APPROXIMABILITY OF DISCRETE AND CONTINUOUS OPTIMIZATION PROBLEMS

COLLABORATIVE RESEARCH: AF: MEDIUM: DESIGN AND ANALYSIS OF MODELS AND ALGORITHMS FOR REAL-LIFE PROBLEMS

AF: SMALL: ALGORITHMS FOR GRAPH ROUTING, DRAWING AND PARTITIONING

CAREER: APPROXIMATION ALGORITHMS AND HARDNESS OF NETWORK OPTIMIZATION PROBLEMS

AF: SMALL: ALGORITHMS FOR SOLVING REAL-LIFE INSTANCES OF OPTIMIZATION AND CLUSTERING PROBLEMS

RI: SMALL: FROM ACOUSTICS TO SEMANTICS: EMBEDDING SPEECH FOR A HIERARCHY OF TASKS

AF: SMALL: GRAPH ROUTING, VERTEX SPARSIFIERS, AND CONNECTIONS TO GRAPH THEORY

RI: SMALL: MULTI-VIEW LEARNING OF ACOUSTIC FEATURES FOR SPEECH RECOGNITION USING ARTICULATORY MEASUREMENTS

COLLABORATIVE RESEARCH: AF: MEDIUM: FAST COMBINATORIAL ALGORITHMS FOR (DYNAMIC) MATCHINGS AND SHORTEST PATHS -A GRAPH IS A COLLECTION OF VERTICES (POINTS OR OBJECTS), AND A COLLECTION OF EDGES (LINKS OR LINES), THAT CONNECT PAIRS OF VERTICES. GRAPHS ARE A CENTRAL AND AN EXTENSIVELY STUDIED TYPE OF MATHEMATICAL OBJECT, AND THEY ARE COMMONLY USED TO MODEL VARIOUS PROBLEMS IN MANY DIFFERENT REAL WORLD SCENARIOS AND APPLICATIONS. FOR EXAMPLE, IT IS NATURAL TO MODEL A ROAD NETWORK IN A CITY, OR A COMPUTER NETWORK, OR FRIENDSHIP RELATIONSHIPS IN A SOCIAL NETWORK AS A GRAPH. THERE ARE COUNTLESS OTHER SCENARIOS WHERE A PROBLEM ONE NEEDS TO SOLVE, OR AN OBJECT ONE DESIRES TO STUDY, CAN BE NATURALLY ABSTRACTED BY A GRAPH. AS A CONSEQUENCE, THE DESIGN OF EFFICIENT ALGORITHMS FOR CENTRAL GRAPH PROBLEMS IS FUNDAMENTAL TO COMPUTER SCIENCE AND BEYOND, AND HAS A SIGNIFICANT IMPACT ON MANY ASPECTS OF COMPUTATION. AS THE AMOUNT OF DATA THAT APPLICATIONS NEED TO DEAL WITH GROWS, IT IS INCREASINGLY IMPORTANT TO ENSURE THAT SUCH ALGORITHMS ARE VERY FAST. IN THIS PROJECT, THE INVESTIGATORS WILL STUDY SEVERAL CENTRAL GRAPH PROBLEMS, SUCH AS MAXIMUM MATCHING, MAXIMUM FLOW, AND SHORTEST PATHS, IN TWO BASIC SETTINGS. THE FIRST IS THE STANDARD MODEL WHERE THE INPUT GRAPH IS KNOWN IN ADVANCE, AND THE GOAL IS TO DESIGN A FAST ALGORITHM FOR THE PROBLEM, WITH RUNNING TIME NOT SIGNIFICANTLY HIGHER THAN THE TIME REQUIRED TO READ THE INPUT, WHICH IS CLOSE TO THE FASTEST POSSIBLE RUNNING TIME. THE SECOND IS THE MODEL OF DYNAMIC ALGORITHMS, WHERE THE GRAPH CHANGES OVER TIME (FOR EXAMPLE, CONSIDER A ROAD NETWORK, WHERE THE COMPUTATION HAS TO ACCOUNT FOR ROADS BECOMING MORE OR LESS CONGESTED WITH TRAFFIC), AND THE GOAL IS TO QUICKLY SUPPORT QUERIES ABOUT THE GRAPH, SUCH AS, FOR EXAMPLE, COMPUTING A SHORT PATH BETWEEN TWO GIVEN VERTICES. THIS PROJECT IS ORGANIZED ALONG FOUR MAIN INTERCONNECTED THRUSTS. THE FIRST THRUST FOCUSES ON THE DESIGN OF ALGORITHMS FOR DYNAMIC ALL-PAIRS SHORTEST PATHS (APSP), THAT CAN WITHSTAND AN ADAPTIVE ADVERSARY, AND THAT SIGNIFICANTLY IMPROVE UPON THE CURRENTLY KNOWN TRADEOFFS BETWEEN THE APPROXIMATION QUALITY AND THE RUNNING TIME, IN BOTH DIRECTED AND UNDIRECTED GRAPHS. ALGORITHMS FOR APSP AND ITS VARIANTS ARE OFTEN USED IN COMBINATION WITH THE MULTIPLICATIVE WEIGHTS UPDATE FRAMEWORK TO EFFICIENTLY SOLVE VARIOUS FLOW AND CUT PROBLEMS IN GRAPHS, AND THUS PROVIDE A VALUABLE AND POWERFUL ALGORITHMIC TOOLKIT. THE SECOND THRUST IS DIRECTED TOWARDS IMPROVING AND EXTENDING KNOWN EXPANDER-RELATED TOOLS THAT ARE OFTEN USED IN THE DESIGN OF FAST ALGORITHMS FOR VARIOUS GRAPH PROBLEMS. EXPANDERS ARE PLAYING AN INCREASINGLY CENTRAL ROLE IN GRAPH ALGORITHMS, AND THESE TOOLS CAN SERVE AS BUILDING BLOCKS FOR MANY OTHER GRAPH PROBLEMS. THE THIRD THRUST FOCUSES ON THE MAXIMUM MATCHING PROBLEM. USING TECHNIQUES INSPIRED BY ALGORITHMS FOR DYNAMIC SHORTEST PATH IN DIRECTED GRAPHS, THE GOAL OF THIS PART OF THE PROJECT IS TO DEVELOP FAST COMBINATORIAL ALGORITHMS FOR BOTH THE BIPARTITE AND THE GENERAL VERSION OF THE PROBLEM. THE FINAL THRUST FOCUSES ON DESIGNING IMPROVED ALGORITHMS FOR MAINTAINING NEAR-OPTIMAL MATCHINGS IN DYNAMIC GRAPHS, BUILDING ON INSIGHTS AND ALGORITHMS DEVELOPED FOR THE SECOND AND THE THIRD THRUSTS. 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.- SUBAWARDS ARE NOT PLANNED FOR THIS AWARD.

RI: MEDIUM: COLLABORATIVE RESEARCH: EXPLICIT ARTICULATORY MODELS OF SPOKEN LANGUAGE, WITH APPLICATION TO AUTOMATIC SPEECH RECOGNITION

ALGORITHM AND WEB SERVER FOR LOW-HOMOLOGY PROTEIN THREADING

BIGDATA: COLLABORATIVE RESEARCH: F: STOCHASTIC APPROXIMATION FOR SUBSPACE AND MULTIVIEW REPRESENTATION LEARNING

AF: SMALL: GRAPH THEORY AND ITS USES IN ALGORITHMS AND BEYOND

CAREER: AN ARCHITECTURE-AWARE OPTIMIZATION THEORY FOR DEEP LEARNING: NON-EUCLIDEAN DESCENT, STRUCTURED PRECONDITIONING, AND SCALE INVARIANCE -TRAINING MODERN ARTIFICIAL INTELLIGENCE SYSTEMS REQUIRES LARGE AMOUNTS OF COMPUTING TIME, ENERGY, AND MONEY. MANY OF THE OPTIMIZATION METHODS USED TO TRAIN NEURAL NETWORKS ARE STILL CHOSEN LARGELY THROUGH TRIAL AND ERROR BECAUSE EXISTING THEORY DOES NOT ADEQUATELY EXPLAIN WHY SOME METHODS WORK BETTER THAN OTHERS ON DIFFERENT MODEL ARCHITECTURES. THIS PROJECT WILL DEVELOP A SCIENTIFIC FOUNDATION FOR MAKING TRAINING FASTER, MORE RELIABLE, AND MORE RESOURCE EFFICIENT BY LINKING OPTIMIZATION METHODS TO THE STRUCTURE OF THE NEURAL NETWORKS THEY ARE USED TO TRAIN. THE PROJECT CAN REDUCE THE COST AND ENERGY USE OF MODEL TRAINING, PROVIDE MORE DEPENDABLE GUIDANCE FOR PRACTITIONERS, AND SUPPORT EFFICIENT AND RELIABLE ARTIFICIAL INTELLIGENCE DEVELOPMENT. IT WILL ALSO SUPPORT GRADUATE EDUCATION IN OPTIMIZATION FOR DEEP LEARNING, RESEARCH-PREPARATION ACTIVITIES FOR UNDERGRADUATES, AND HANDS-ON ARTIFICIAL INTELLIGENCE LEARNING MODULES FOR LOCAL HIGH SCHOOL STUDENTS, WITH PARTICIPATION IN PROJECT ACTIVITIES OPEN TO ALL. THIS PROJECT STUDIES HOW NEURAL NETWORK ARCHITECTURE INFLUENCES OPTIMIZATION THROUGH THREE COMPLEMENTARY DIRECTIONS: STRUCTURED PRECONDITIONING, OPTIMIZATION METHODS ADAPTED TO DIFFERENT NOTIONS OF DISTANCE, AND SCALE INVARIANCE INDUCED BY NORMALIZATION LAYERS. THE RESEARCH WILL ANALYZE REPRESENTATIVE MODEL COMPONENTS SUCH AS MULTILAYER PERCEPTRONS, ATTENTION MODULES, EMBEDDING PARAMETERS, AND LAYERS PRECEDING NORMALIZATION. IT WILL DEVELOP THEORY EXPLAINING WHEN OPTIMIZATION METHODS MATCHED TO THE MODEL ARCHITECTURE IMPROVE TRAINING EFFICIENCY, CHARACTERIZE THEIR BEHAVIOR ON LOSSES WHOSE LANDSCAPE GEOMETRY IS SHAPED BY THE NETWORK ARCHITECTURE, AND STUDY HOW OPTIMIZATION AFFECTS THE QUALITY OF LEARNED SOLUTIONS BEYOND TRAINING LOSS ALONE, INCLUDING DOWNSTREAM TASKS AND PERFORMANCE WHEN DATA DIFFER FROM THE TRAINING DISTRIBUTION. THESE IDEAS WILL BE TESTED THROUGH CONTROLLED EXPERIMENTS ON REPRESENTATIVE NEURAL NETWORK ARCHITECTURES AND LARGER-SCALE VALIDATION ON PRODUCTION-RELEVANT TRAINING PIPELINES. THE PROJECT WILL RELEASE OPEN-SOURCE SOFTWARE, REPRODUCIBLE EXPERIMENTAL CONFIGURATIONS, BENCHMARK RESULTS, AND PUBLIC DOCUMENTATION. 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.- SUBAWARDS ARE NOT PLANNED FOR THIS AWARD.

NRI: COLLABORATIVE RESEARCH: LEARNING ADAPTIVE REPRESENTATIONS FOR ROBUST MOBILE ROBOT NAVIGATION FROM MULTI-MODAL INTERACTIONS

AF: SMALL: FOUNDATIONS FOR COLLABORATIVE AND INFORMATION-LIMITED MACHINE LEARNING

NRI: INT: COLLAB: SHARED AUTONOMY FOR UNSTRUCTURED UNDERWATER ENVIRONMENTS THROUGH VISION AND LANGUAGE

AF:III: SMALL: CONVEX OPTIMIZATION FOR PROTEIN-PROTEIN INTERACTION NETWORK ALIGNMENT

CCF-BSF: AF: SMALL: CONVEX AND NON-CONVEX DISTRIBUTED LEARNING

EAGER: COLLABORATIVE RESEARCH: WORLD MODELING FOR NATURAL LANGUAGE UNDERSTANDING

COLLABORATIVE PROPOSAL: DISCRIMINATIVE LATENT VARIABLE OBJECT DETECTION

AF: SMALL: UNDERSTANDING EXPANSION PHENOMENA: GRAPHICAL, HYPERGRAPHICAL, GEOMETRIC, AND QUANTUM -A BROAD VARIETY OF PHENOMENA IN COMPUTATION CAN BE VIEWED AS DIFFERENT FORMS OF ?EXPANSION?, WHICH ENSURE THAT LOCAL PROPERTIES WHICH CAN BE OBSERVED BY LOOKING AT SMALL PARTS OF AN OBJECT, CAN BE USED TO INFLUENCE AND UNDERSTAND GLOBAL PROPERTIES EXHIBITED AT A MUCH LARGER SCALE. THIS IS AN IMPORTANT DESIGN REQUIREMENT IN SEVERAL APPLICATIONS, SUCH AS (1) CLASSICAL AND QUANTUM ERROR-CORRECTION, WHERE ONE WANTS ERRORS TO BE EASILY DETECTABLE BY LOCAL CHECKS, (2) OPTIMIZATION PROBLEMS, WHERE ONE WANTS LOCAL CHOICES TO PUSH THE GLOBAL SOLUTION TOWARDS OPTIMALITY, AND (3) GEOMETRIC EMBEDDINGS OF HIGH-DIMENSIONAL DATA, WHERE ONE WANTS TO USE LOCAL (LOW-DIMENSIONAL) CONDITIONS TO INFLUENCE HIGH-DIMENSIONAL BEHAVIOR. IN THE PAST FEW YEARS, SEVERAL NEW CONCEPTS AND TECHNIQUES HAVE EMERGED TO STUDY EXPANSION PHENOMENA IN DIFFERENT CONTEXTS. THIS PROJECT AIMS TO STUDY SEVERAL DIFFERENT FORMS OF EXPANSION PHENOMENA IN A UNIFIED WAY, WITH AN EMPHASIS ON APPLICATIONS IN THE AREAS OF ERROR-CORRECTING CODES AND (APPROXIMATE) OPTIMIZATION. THIS RESEARCH IS LIKELY TO LEAD TO NEW CONNECTIONS BETWEEN MULTIPLE AREAS WHERE SUCH PHENOMENA ARE USEFUL. THE MATERIAL GENERATED AS PART OF THIS RESEARCH WILL ALSO BE DISSEMINATED THROUGH SURVEYS AND A SERIES OF EXPOSITORY VIDEOS. THIS PROJECT AIMS TO OBTAIN A UNIFIED VIEW OF THE FOLLOWING DIFFERENT FORMS AND APPLICATIONS OF EXPANSION PHENOMENA: - CLASSICAL NOTIONS OF GRAPH EXPANSION AND NOVEL NOTIONS OF HIGH-DIMENSIONAL EXPANSION FOR HYPERGRAPHS, AND THEIR CONNECTIONS TO RECENT ADVANCES IN CODING THEORY. - APPLICATIONS OF CLASSICAL EXPANSION PHENOMENA TO QUANTUM CODES, AS WELL AS QUANTUM EXTENSIONS OF CLASSICAL EXPANSION PHENOMENA. - CONNECTIONS OF HIGH-DIMENSIONAL EXPANSION TO THE STUDY AND APPROXIMABILITY OF EXPANSION PHENOMENA IN GEOMETRIC SPACES, AND RELATED PROBLEMS ABOUT FINE-GRAINED GRAPH EXPANSION. THE RESEARCH DIRECTIONS PURSUED IN THIS PROJECT AIM TO INTRODUCE NEW TECHNIQUES IN ALGORITHMIC CODING THEORY AND IN THE STUDY OF APPROXIMABILITY OF DISCRETE AND CONTINUOUS OPTIMIZATION PROBLEMS. THE PROJECT CONSIDERS SEVERAL PROBLEMS THAT HAVE PROVED TO BE BOTTLENECKS FOR CURRENT ALGORITHMIC AND ANALYTIC TECHNIQUES, EXPLORES NEW APPROACHES ARISING FROM THE STUDY OF EXPANSION IN A DIFFERENT CONTEXT. THE PROJECT AIMS TO APPLY THESE IDEAS FOR THE DESIGN OF NEW ERROR-CORRECTING CODES, AND NEW ALGORITHMS FOR EXISTING CODES, TOWARDS THE DESIGN OF NEW PSEUDORANDOM OBJECTS, AND ALSO NEW FAMILIES OF COMBINATORIAL AND GEOMETRIC INSTANCES FOR PROVING INAPPROXIMABILITY RESULTS. 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.

GRADUATE RESEARCH FELLOWSHIP PROGRAM (GRFP)

GRADUATE RESEARCH FELLOWSHIP PROGRAM (GRFP) -THE NATIONAL SCIENCE FOUNDATION (NSF) GRADUATE RESEARCH FELLOWSHIP PROGRAM (GRFP) IS A HIGHLY COMPETITIVE, FEDERAL FELLOWSHIP PROGRAM. GRFP HELPS ENSURE THE VITALITY AND DIVERSITY OF THE SCIENTIFIC AND ENGINEERING WORKFORCE OF THE UNITED STATES. THE PROGRAM RECOGNIZES AND SUPPORTS OUTSTANDING GRADUATE STUDENTS WHO ARE PURSUING RESEARCH-BASED MASTER'S AND DOCTORAL DEGREES IN SCIENCE, TECHNOLOGY, ENGINEERING, AND MATHEMATICS (STEM) AND IN STEM EDUCATION. THE GRFP PROVIDES THREE YEARS OF FINANCIAL SUPPORT FOR THE GRADUATE EDUCATION OF INDIVIDUALS WHO HAVE DEMONSTRATED THEIR POTENTIAL FOR SIGNIFICANT RESEARCH ACHIEVEMENTS IN STEM AND STEM EDUCATION. THIS AWARD SUPPORTS THE NSF GRADUATE FELLOWS PURSUING GRADUATE EDUCATION AT THIS GRFP INSTITUTION. 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.

CRII: AF: STREAMING APPROXIMABILITY OF MAXIMUM DIRECTED CUT AND OTHER CONSTRAINT SATISFACTION PROBLEMS -CONSTRAINT SATISFACTION PROBLEMS (CSPS) ARE UBIQUITOUS AND ENCOMPASS SEVERAL IMPORTANT OPTIMIZATION PROBLEMS STUDIED IN DIVERSE AREAS IN COMPUTER SCIENCE. SOME OF THEIR MOST POPULAR APPLICATIONS INCLUDE DATA CLUSTERING, RESOURCE PLANNING, AND CHIP DESIGN. IN THEORETICAL COMPUTER SCIENCE, SEVERAL BROADLY-APPLICABLE ALGORITHMIC AND COMPLEXITY THEORETIC TOOLS HAVE ORIGINATED FROM THE STUDY OF CSPS. WHILE TRADITIONAL RESEARCH ON CSPS ASSUMES THAT THE ENTIRE INPUT IS AVAILABLE TO THE ALGORITHM, THE BIG DATA BOOM HAS NECESSITATED STUDYING THESE PROBLEMS IN NEWER MODELS OF COMPUTATION THAT ARE WELL-SUITED FOR PROCESSING VERY LARGE DATASETS THAT CANNOT ENTIRELY FIT IN THE ALGORITHM?S MEMORY. ONE SUCH WELL-STUDIED MODEL OF COMPUTATION IS THE STREAMING MODEL WHERE THE INPUT TO THE ALGORITHM IS PROVIDED AS A STREAM OF DATA, AND THE STREAMING ALGORITHM MUST PERFORM ALL ITS COMPUTATIONS USING LIMITED MEMORY, MUCH SMALLER THAN THE SIZE OF THE STREAM. A PARTICULAR CSP OF INTEREST IS THE MAXIMUM DIRECTED CUT (MAX-DICUT) PROBLEM, WHICH HAS EMERGED AS A CENTRAL PROBLEM IN THE STUDY OF CSPS IN THE STREAMING SETTING. DESPITE SIGNIFICANT ATTENTION, MANY FUNDAMENTAL QUESTIONS ABOUT MAX-DICUT AND OTHER CSPS REMAIN UNANSWERED IN THE STREAMING MODEL. THIS PROJECT AIMS TO TACKLE SOME OF THESE FOUNDATIONAL PROBLEMS AND PROVIDE SIGNIFICANT INSIGHTS INTO THE CAPABILITIES AND THE LIMITATIONS OF STREAMING ALGORITHMS IN SOLVING CSPS. THIS PROJECT ALSO PROVIDES RESEARCH OPPORTUNITIES FOR GRADUATE AND UNDERGRADUATE STUDENTS THROUGH A NEW COURSE IN ADVANCED ALGORITHMS THAT WILL INTEGRATE TOPICS FROM THESE RESEARCH DIRECTIONS. THE RESEARCH AND COURSE MATERIALS PRODUCED AS A RESULT WILL BE MADE ACCESSIBLE TO A GENERAL AUDIENCE. IN TERMS OF RESEARCH DIRECTIONS, MOST PREVIOUS WORKS FOCUSED ON ?SINGLE-PASS? STREAMING ALGORITHMS, WHERE THE ALGORITHM IS ALLOWED ONLY ONE PASS THROUGH THE STREAM AND THE ORDER IN WHICH THE DATA ARRIVES IS DECIDED BY A MALICIOUS ADVERSARY. WHILE THIS WORKS AS AN EXCELLENT MODEL IN THEORY, IN PRACTICE, THE ALGORITHMS OFTEN HAVE ADDITIONAL ?HELP?. FOR EXAMPLE, THEY MAY BE ALLOWED MULTIPLE PASSES OVER THE INPUT OR REQUIRED TO PERFORM WELL ONLY ON INPUTS DRAWN FROM A CERTAIN DISTRIBUTION. THEY MAY ALSO HAVE ACCESS TO QUANTUM BITS OR A MACHINE LEARNING ORACLE THAT CAN PREDICT THE REST OF THE STREAM. IN SUCH SCENARIOS, THERE MAY BE ALGORITHMS THAT ARE EXPONENTIALLY MORE SPACE EFFICIENT THAN THE BEST SINGLE-PASS ALGORITHMS. THIS PROJECT AIMS TO DESIGN STREAMING ALGORITHMS FOR MAX-DICUT AND OTHER CSPS IN SUCH MORE GENERAL SETTINGS. 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.- SUBAWARDS ARE NOT PLANNED FOR THIS AWARD.

RI: SMALL: PROBABILISTIC LATENT VARIABLE MODELS FOR SPARSE DATA

DEEP STRUCTURED LEARNING FOR SCENE UNDERSTANDING

MOBILE MANIPULATION PLATFORM FOR LANGUAGE-BASED SHARED AUTONOMY

RI: SMALL: DEBUGGING MACHINE VISUAL RECOGNITION VIA HUMANS IN THE LOOP

COLLABORATIVE RESEARCH: AF: SMALL: FINE- GRAINED COMPLEXITY OF APPROXIMATE PROBLEMS

EAGER: DISCOVERY OF SEGMENTAL SUB-WORD STRUCTURE IN SPEECH

AF: SMALL: NEW DIRECTIONS IN LEARNING THEORY

ATD: COLLABORATIVE RESEARCH: AUTOMATIC, ADAPTIVE DETECTION AND DESCRIPTION OF CHANGE IN TIME-LAPSE IMAGERY

CCF: EAGER: DIMENSION REDUCTION AND OPTIMIZATION METHODS FOR FLEXIBLE REFINEMENT OF PROTEIN DOCKING

AF: MEDIUM: COLLABORATIVE RESEARCH: CIRCUIT LOWER BOUNDS VIA PROJECTIONS

COLLABORATIVE RESEARCH: JOINT ANALYSIS OF CORRELATED DATA

COMPUTER AND INFORMATION SCIENCE AND ENGINEERING GRADUATE FELLOWSHIPS (CSGRAD4US) -THIS AWARD IS FUNDED IN WHOLE OR IN PART UNDER THE AMERICAN RESCUE PLAN ACT OF 2021 (PUBLIC LAW 117-2).? THE COMPUTER AND INFORMATION SCIENCE AND ENGINEERING GRADUATE FELLOWSHIPS (CSGRAD4US) IS INTENDED FOR INDIVIDUALS WHO HAVE SOME PRACTICAL EXPERIENCE FOLLOWING THEIR BACHELOR?S DEGREE AND ARE NOW INTERESTED IN PURSUING A RESEARCH-BASED DOCTORAL DEGREE. CSGRAD4US FELLOWSHIPS ARE A PART OF AN OVERALL STRATEGY BY NSF'S CISE DIRECTORATE TO DEVELOP THE WORKFORCE NECESSARY TO ENSURE THE NATION'S LEADERSHIP IN ADVANCING CISE RESEARCH AND INNOVATION. THE CSGRAD4US FELLOWSHIP PROVIDES THREE YEARS OF FINANCIAL SUPPORT FOR THE GRADUATE EDUCATION OF INDIVIDUALS WHO HAVE DEMONSTRATED THEIR POTENTIAL FOR SIGNIFICANT RESEARCH ACHIEVEMENTS IN CISE DISCIPLINES. THIS AWARD SUPPORTS THE CSGRAD4US FELLOWS PURSUING GRADUATE EDUCATION AT THIS CSGRAD4US INSTITUTION. 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.

RI: SMALL: COLLABORATIVE RESEARCH: STRUCTURED INFERENCE FOR LOW-LEVEL VISION

GRADUATE RESEARCH FELLOWSHIP PROGRAM (GRFP)

DOCTORAL CONSORTIUM AT THE 2018 INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA)

CIF: SMALL: COLLABORATIVE RESEARCH: INFERENCE BY SOCIAL SAMPLING

COLLABORATIVE RESEARCH: CPA-SEL: IMPLEMENTATION TECHNIQUES FOR HIGH-LEVEL PARALLEL LANGUAGES

CONFERENCE: 2025 MIDWEST ROBOTICS CONFERENCE -THE ROBOTICS FIELD IS EXPERIENCING TREMENDOUS GROWTH DUE TO ALGORITHMIC AND TECHNOLOGICAL ADVANCES, THE AVAILABILITY OF COMMON LOW-COST SENSORS AND PLATFORMS, THE STANDARDIZATION OF OPEN-SOURCE DEVELOPMENT, AND LARGE-SCALE INDUSTRIAL INITIATIVES. THESE ADVANCEMENTS, COUPLED WITH A RAPIDLY GROWING COHORT OF TALENTED, HIGHLY SKILLED ROBOTICISTS, COMBINE TO ACCELERATE THE TRANSITION OF ROBOTICS FROM RESEARCH LABS TO REAL-WORLD DEPLOYMENT---RESHAPING INDUSTRIES, ENHANCING DAILY LIFE, AND TRANSFORMING THE WAY WE INTERACT WITH TECHNOLOGY IN OUR HOMES, WORKPLACES, AND PUBLIC SPACES. A KEY DRIVER OF THIS PROGRESS IS THE PRESENCE OF CLOSE-KNIT COMMUNITIES OF ROBOTICS RESEARCHERS AND PRACTITIONERS FROM ACADEMIA AND INDUSTRY, WHO FOSTER COLLABORATION, EXCHANGE CUTTING-EDGE IDEAS, MENTOR THE NEXT GENERATION OF ROBOTICISTS, AND BRIDGE THE GAP BETWEEN FUNDAMENTAL RESEARCH AND REAL-WORLD APPLICATIONS. THE MIDWEST ROBOTICS WORKSHOP (MWRW) BRINGS TOGETHER ROBOTICISTS FROM IN AND AROUND THE MIDWEST, PROVIDING A UNIQUE OPPORTUNITY FOR RESEARCHERS AND PRACTITIONERS TO SHARE THEIR WORK WITH OTHERS AND TO NETWORK IN A FOCUSED SETTING, WITH THE GOAL OF CREATING A MORE COHESIVE AND VIBRANT ROBOTICS COMMUNITY IN THE MIDWEST. A DEFINING FEATURE OF MWRW IS ITS INVITED TALKS THAT SHOWCASE SENIOR PH.D. STUDENTS AND EARLY CAREER (I.E., PRE-TENURE) FACULTY, PROVIDING THEM WITH VALUABLE EXPOSURE, INCREASING THE VISIBILITY OF THEIR WORK WITHIN THE ROBOTICS COMMUNITY, AND OFFERING OPPORTUNITIES TO RECEIVE INSIGHTFUL FEEDBACK FROM SENIOR RESEARCHERS. THIS AWARD SUPPORTS NON-LOCAL STUDENTS IN PARTICIPATING IN MWRW 2025, WHICH WILL TAKE PLACE JUNE 2?3, 2025 IN CHICAGO, IL. 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.- SUBAWARDS ARE NOT PLANNED FOR THIS AWARD.