Research in the Computer Science department concentrates on foundational problems with real-life applications. Our faculty actively engages in and lead large-scale interdisciplinary research projects in analysis of networked information, data science and stewardship, bioinformatics, ecology and environmental conservation. We seek to develop principled solutions to problems and illustrate our solutions with working prototypes.
Network Science and Social & Cognitive Networks
Research in network science and social & cognitive networks links together top social scientists, neuroscientists, and cognitive scientists with leading physicists, computer scientists, mathematicians, and engineers in the search to uncover, model, understand, and foresee the complex social interactions that take place in today's society. All aspects of social networks, from the origins of adversarial networks to gauging the level of trust within vast social networks, are investigated. More generally, network dynamics, network synchronization, resilience of infrastructure networks and networks in general are also studied.
Semantic Web and Data Infrastructure
This research focuses on the emerging area of Web Science and specifically the evolving web and related semantic technologies. Active areas of research include: Semantic Web technology, knowledge graphs and cognitive computing, provenance and explanation; data integration and analytics; cyberinfrastucture for data integration, particularly for science data; ontology development; and the ethical, policy, and social aspects of Web use and usability. Research in this area also focuses more broadly on data cyberinfrastructure, including policy, social and technical infrastructure needed for data stewardship and preservation.
Machine Learning, Data Mining, Artificial Intelligence
Our research tackles the theoretical and applied sides of extracting knowledge from data. Within the Big Data arena, we emphasize efficient, scalable, and parallel algorithms for data mining and data management tasks (association rules, classification, clustering, sequence mining, etc.). For small data sets, the emphasis is on robust learning systems (supervised, unsupervised and reinforcement). Prior and current projects touch several application areas, which include: computational biology (bioinformatics, computational genomics); biomedical engineering; public health informatics; cheminformatics, web mining; geographic information systems; computational finance; natural language processing, and, multi-agent social data aggregation into informed actions (such as in recommendation systems like Yelp and TripAdvisor).
Theory of Computation and Algorithms
Research in this area provides the foundation needed for effective computing. The theory group at Rensselaer's Computer Science Department brings together researchers in many areas of Computer Science to develop novel approaches and solutions to computational problems. Our research at RPI includes networks of all kinds (computer networks, social networks, information networks, sensor networks); machine learning and data mining; distributed algorithms; graph algorithms; algorithmic game theory and computational economics; independent and strategic agents; voting and social choice; computational finance; security; and bioinformatics.
Research in computer vision in the Department of Computer Science has taken a new direction. Professor Charles Stewart and his students, both graduate and undergraduate, are working on applications of computer vision to problems in environmental monitoring, with the largest domain being oceanography. A wide range of problems arise, including illumination modeling and color correction, registration and 3d reconstruction, motion analysis, and recognition. Practical issues of high-volume throughput and large-scale software system development are also under consideration.
Pervasive Computing and Distributed Systems
Researchers investigate computer networks and their protocols, with a focus on wireless and sensor networks through the International Technology Alliance, a new 10-year research consortium led by the IBM Research Division and funded jointly by the US and UK Governments with participation of the leading researchers in the world. The focus is on sensor information processing and delivery, improvement of the quality of information obtained from sensor networks and adaptation of sensor networks to the dynamically changing user demands. Another area of activity is the security of computers, networks, and sensors. Secruity concerns are quickly becoming a significant barrier to the wide-spread acceptance of pervasive computing. The research tackles such issues as trust in Internet communications, identity of groups on the Internet, cryptographic and systemic challenges in sensor networks. Finally, in the area of high-performance pervasive computing, the focus is on computational environments in which task allocation, migration, and fault tolerance are supported automatically and on application of such environments to computations relevant to different scientific disciplines.
Cybersecurity is called upon to answer some of the most pressing questions of our digital society. Researchers in the CyberSecurity group are interested in security problems both from a theory perspective and at the systems level. Specific problems of interest include malware analysis and reverse engineering, program obfuscation, covert channels, discovering hidden networks in social networks; network camouflaging; privacy protection in data mining systems; foundations of cryptography and provable security, secure (multi-party) computation, fault-tolerant distributed protocols, intrusion detection, game-theoretic cryptography, cryptocurrencies and e-voting.
Programming Languages and Software Engineering
The Programming Languages and Software Engineering research group develops programming models, languages, and tools to improve software reliability, performance, security, and programmer productivity. Recent efforts have focused on Android security, concurrent software verification, data analytics in cloud environments, and open source software practices.
Computational Science and Engineering
Students and faculty members work on computational approaches and algorithms to solve large-scale problems that arise in natural science and engineering. Current research includes massively parallel computing methods, adaptive methods for solving partial differential equations, multiscale computations, scientific software libraries, algorithms for medical imaging and tomography, high-performance matrix algorithms, computational biology, and parallel adaptive unstructured mesh methods.
The faculty and students in the Computer Graphics Research Group are interested in a wide variety of rendering, geometry, simulation, and visualization problems motivated by computer games, special effects in movies, architectural design & pre-visualization, and many other exciting applications. We study topics including physically-based digital sculpting, efficient high-quality photo-realistic rendering, new data representations and algorithms, and the use of modern graphics hardware for interactive applications. Other topics include modeling terrain and compressing large datasets in computational cartography and geographic information science.
Bioinformatics is the science of managing, retrieving, analyzing, and interpreting biological data. Research is being carried out on topics such as sequence assembly, protein and RNA structure prediction, sequence/structure/motifs, comparative genomics, and the gene regulatory networks. Research also spans emerging areas like microarray data analysis, protein design, high dimensional indexing, database support, information integration, and data mining.
Computational Cognitive Modeling
At the CogWorks Lab we are interested in basic and applied research in the area of immediate interactive behavior. On the basic side, we are working to understand the interplay of cognition, perception, and action in routine interactive behavior. These interests entail understanding top-down versus bottom-up control of behavior, the role of implicit versus explicit knowledge, internal versus external representations, and knowledge in-the-head versus knowledge in-the-world. On the applied side, we specialize in the field of Cognitive Engineering (cognitive science theory applied to human factors issues). Our research methods include behavioral and performance measures (including eye-tracking), brain-based measures (EEG), and computational cognitive modeling (usingACT-R, SanLab, and closed form modeling).
Logic-Based Artificial Intelligence
Researchers in the RAIR Lab design and build intelligent agents, software, robots, etc. on the basis of formal logic. R&D has been and is sponsored by NSF, ARDA/DTO, AFOSR, etc. PhD students need to have some background in logic, AI, and corresponding programming paradigms.
This research area deals with the efficient and effective methods for storing, querying and maintaining data from possibly disparate and heterogeneous resources. Data is used in many different applications from scientific data sets, sensor data, images, video and audio to hypertext documents, and data on stock market behavior. Research focuses on methods for caching data, querying large and distributed databases and supporting applications such as computer-aided design and manufacturing and collaborative engineering.
Current research in computational geometry has two themes. The first concentrates on algorithms for the reconstruction of smooth geometric objects from their samples. Problems of interest include characterizing the conditions on sampling density, which allow a curve to be reconstructed from its samples. The reconstruction is homeomorphic and sufficiently close to the original and the algorithms developed to achieve the reconstruction. Also involved are the dependence of such algorithms on the dimension of the embedding space, related algorithms for the reconstruction of surfaces and manifolds, and finding the most concise representation of a manifold in terms of its samples. A second research track focuses on applications of computational geometry, particularly in robotic motion planning. The second computational geometry theme emphasizes small, simple, and fast geometric data structures and algorithms. Note that efficiency in both space and time can become more important as machines get faster. This research is applicable to computational cartography, computer graphics, computational geometry, and geographic information science. GeoStar, a recently concluded DARPA-funded project in this theme, modelled terrain to compress it and to site observers and then to plot motion paths to avoid those observers. A current NSF-funded project is modeling how levees erode as floodwaters overtop them.