Statistical Relational Learning (SRL) Models combine the powerful formalisms of probability theory and first-order logic to handle uncertainty in large, complex problems. While they provide a very effective representation paradigm due to their succinctness and parameter sharing, efficient learning is a significant problem in these models. First, I will discuss state-of-the-art learning method based on boosting that is representation independent. Our results demonstrate that learning multiple weak models can lead to a dramatic improvement in accuracy and efficiency.
The first milestone in the CIS Ph.D. program is the Directed Research Project (DRP). Each student must complete this milestone within their first two years in the program. Typically, students devote the summer after their first year in the program exclusively to their research projects. In this colloquium, seven of our Department's Ph.D. students will present summaries on the progress on their DRPs over this past summer.
Understanding videos of human actions, recorded in an uncontrolled setting, is an open problem in computer vision. Video surveillance, content retrieval, autonomous driving and sports analysis are examples of practical applications. We focus our research on efficiency and robustness of action recognition in real-world videos.
This week we continue our introduction to faculty research topics. We encourage all faculty members and PhD students to attend and hope these presentations help students meet faculty members and get exposed to the full range of the research portfolio in our department. We feature the following three speakers this week:
Presentation #1: "Research Overview for CDUX: Computing and Data Understanding at eXtreme Scale" by Associate Professor Hank Childs
Tor is a popular network for anonymous communication. The usage and operation of Tor is not well-understood, however, because its privacy goals make common measurement approaches ineffective or risky. We present PrivCount, a system for measuring the Tor network designed with user privacy as a primary goal. PrivCount securely aggregates measurements across Tor relays and over time to produce differentially private outputs. PrivCount improves on prior approaches by enabling flexible exploration of many diverse kinds of Tor measurements while maintaining accuracy and privacy for each.
Follow-up to mini town-hall meeting for graduate students.
The arrival of the Precision Medicine age brings tremendous opportunities to scientific discovery and quality improvement in medicine and healthcare. However, it also raises big challenges in dealing with large and massive healthcare data from heterogeneous sources.
With petabytes or even zettabytes of data that are continuously collected about various aspects of the Internet, how hard can it be to obtain an accurate picture of its traffic, its physical topology (i.e., router-level Internet), its logical overlays (e.g., the Web, online social networks), or its "dark" sides and associated activities (i.e., cyber crimes)? In this talk, I will first use some well-known examples of "big (Internet) data" to illustrate what this data does and doesn't tell us about the Internet's traffic and its physical topology.
Dialogue system (Chatbot) is becoming a very hot topic and fascinating technology trend both in acadmic community and industries. Google, Fackbook, Microsoft and many startups seem to believe that Chatbots will be a new generation of apps after websites and mobile apps. Many dialog systems have been implemented to provide a variety of services, such as call routing, flight booking, weather forecasting, and restaurants recommendation. Those goal-driven dialog systems enable a human user to acquire information and services by interacting with a computer agent using natural languages.
In the same way that the Internet has combined with web content and search engines to revolutionize every aspect of our lives, the scientific process is poised to undergo a radical transformation based on the ability to access, analyze, and merge large, complex data sets. Scientists will be able to combine their own data with that of other scientists to validate models, interpret experiments, re-use and re-analyze data, and make use of sophisticated mathematical analyses and simulations to drive the discovery of relationships across data sets. This “scientific web” will yield higher quality science, more insights per experiment, an increased democratization of science, and a higher impact from major investments in scientific instruments.