AI Innovation Institute

Current work includes: 

• network or software based attacks,
• detection and confinement, and
• Software Analysis and Debugging Tools to detect defects in complex systems.

Current work includes: 

• Content-Based Access Control: CBAC enables bloggers and social networkers to specify simple access-control policies that can be applied automatically to future posts, photographs.
• Static Analysis for Security: Develop tools to help programmers write better software.
• Secure Mobile Commerce: Cell-phones will soon replace credit cards in point-of-sale transactions. We propose an architecture for enabling secure commerce on an otherwise untrusted platform.
• Usable Password Management: We are developing a usable web password manager based on the principle that "a secure user interface should not only make doing the right thing easy, it should make doing the wrong thing hard."

N/A

Among the projects are: 

• FiST: A language for stackable file systems ported to Linux, Solaris, and FreeBSD. FiST allows incremental file system development, without requiring an in-depth knowledge of kernel internals. A Windows NT port is under way.
• CoSy: Compound System Calls, a method of aggregating multiple system calls into a single system call. This reduces data copies and context switches. Initial results show promising performance improvements.
• NCryptfs: A next-generation cryptographic file system, which balances security, convenience, and performance.
• Elastic Quotas: A novel disk-space management method, which allows users to temporarily exceed their storage quota.
• Versionfs: A stackable versioning file system, with support for powerful version storage and retention policies. Versionfs provides online backup and restore capabilities, with low overhead.

Current projects include: 

• Mobility Modeling: We work on the development of a data-driven mobility model, using statistical likelihoods for path selection and speed of movement. Our focus is on the development of a framework to support any mobility traces and for ease of evaulation of various communication protocols, deployments, etc under such models.
• Energy Efficient Wireless Protocols: We work on the development of wireless protocols which consider energy metrics for lifetime extension in sensor netowrks.
• Mobile Computing & Applications: We work on the development of socially influenced mobile phone applications for developing countries. Challenges include working with limited and out-dated hardware, limited infrastructure and cost models in deployed areas, and facilitating current computation models to the platform. Another focus is on the development of peer-to-peer mobile and social information networks.
• Architectures for Emerging Interconnect Technologies: With the emergance of new on-chip interconnect technologies, CAD is facing a drastic shift in optimization priorities, such as power, area, and bandwidth. We are developing new architectures and optimization techniques for emerging technologies, such as RF, optical, and/or plasmonics.

Members of the Lab are engaged in three major projects: 

• FLORA - a declarative object-oriented language for programming knowledge intensive applications.
• The LMC (Logic Programming-Based Model Checking)
• XSB - A high-performance logic programming and deductive database system.

Specific Projects: 
Network localization: We develop efficient localization algorithm in which sensors collaborate with each other to locate themselves and self-organize into a connected network.
Topology discovery: We work on distributed algorithms that discover and maintain high-order topological features (e.g., holes not covered by sensors) or the geometric shape of the sensor field.
Mobility: We work on heterogeneous sensor networks that integrate mobile nodes (robots, people holding cell phones) with static monitoring sensor nodes and foster novel applications.

Our research covers a range of topics in natural language processing. A current focus is using Deep Learning techniques to build concise representations of the meanings of words in all significant languages, and use these powerful features to recognize entities and measure sentiment and other properties of texts. Another focus involves analyzing Wikipedia to identify the fame and significance of historical figures as reported in our book Who's Bigger? and associated website. Our Lydia technology has been licensed by General Sentiment, a social media analysis startup.

Theoretical and experimental analysis of string, graph, and combinatorial algorithms, including 

• applications in computational biology and combinatorial computing,
• randomized algorithms, with applications to scheduling, and
• computational geometry and approximation algorithms, particularly with applications to computer graphics and manufacturing.