Intelligent Systems encompasses the study of algorithms and architectures that enable effective decision making in complex environments. Researchers in this area include faculty, undergraduate and graduate students working on projects in computer vision, robotics, virtual theatre, sensor networks, data mining, document recognition, and the theoretical foundations of decision-making (e.g. Markov chains and the properties of voting protocols).
- Prof. Zack Butler
- Prof. Edith Hemaspaandra
- Prof. Hadi Hosseini
- Prof. Chris Homan
- Prof. Leon Reznik
- Prof. Lingwei Wang
- Prof. Richard Zanibbi
Selected Research Projects
Classification of the severity of Parkinson's disease symptoms
Parkinson's disease is a progressive disorder of the nervous system. The disease affects movement and the symptoms worsen over time. Medications can be used to reduce symptoms. The goal of this project is to automatically determine the severity of the symptoms, resulting in a possible change to the prescribed treatment. | This project is a collaborative effort with medical researchers at the University of Rochester Medical Center. Accelerometer and gyroscope sensor data have been collected from a number of patients with Parkinson's disease. | The data is from wrist, leg, head and torso locations. The current research is focused on developing a fuzzy rule based classifier that is optimized using a genetic algorithm that will determine the severity of the symptoms during drug treatment. |
Computational Social Choice
Elections are broadly used in both human and computational settings, including a rapidly expanding range of applications in multiagent systems. It has been known since the mid-1970s (the Gibbard-Satterthwaite Theorem) that every reasonable election system has instances on which voters have an incentive to vote strategically. Computational social choice seeks to sidestep that impossibility result by making manipulation not impossible but rather computationally prohibitive.
Computational social network analysis| (CSNA)
This work combines social media, network science, sociology, and advanced computational methods to study social networks and the roles they play in organizational behavior, motivated by the belief that a better understanding of those dynamics will lead to better health, productivity, and general social welfare.
Corobots for use across the CS Curriculum
This project is based around two important ideas: one, that using a real-world domain such as robotics can inspire and aid students in learning computing concepts; and two, that robots that live and work around humans ("corobots") have many potential applications that we are still in the process of discovering!
min math search interface
A math editor and search interface supporting math input using images, keyboard, and mouse/touch. The program works in web browsers on both iPads and desktop machines. Queries may include keywords and math, and be sent to various math-aware search engines including Wolfram Alpha and our own Tangent search engine.
Neural networks for cognitive sensor networks
This project puts forward a concept of cognitive sensor networks, which should be able to perform self-organizing and self-reconfiguring depending on the given goal. It investigates a feasibility of artificial neural networks application for its realization through the design of novel hierarchical configurations imitating the structural topology of brain-like architectures. They are composed from artificial neural networks distributed over network platforms with limited resources. The project examines a cognition idea based on its implementation through the signal change detection. The novel multilevel neural networks architectures are designed and tested in sensor networks built from Crossbow Inc. sensor kits. The results are compared against conventional multilayer perceptron structures in terms of both functional efficiency and resource consumption.
Tangent search engine
A search engine for math expressions in Wikipedia, using new indexing and retrieval algorithms developed by David Stalnaker for his MSc thesis. Search results may edited in min, and then copied to documents or used in other searches.
Tests used to prevent abuses of web-based programs often distinguish humans and computers by asking them to transcribe distorted text images. These are Completely Automated Public Turing test to tell Computer and Humans Apart, and for many users they are difficult to pass. For his MSc thesis, Kurt Kluever developed a new CAPTCHA where a user must provide tags describing a video based on its content, with 'correct' tags obtained by mining tags from related videos in YouTube, producing usable human success rates while maintaining usable security against attacks using common tags.