Staff

Speaker: Richard Mann

Classical methods for measuring image motion by computer have concentrated on the cases of optical flow in which the motion field is continuous, or layered motion in which the motion field is piecewise continuous. Here we introduce a third natural category which we call optical snow.

Speaker: Jane Morris

Most research on automated text understanding, including applications in information retrieval and text summarization, does not consider the potential subjectivity of human interpretation of text. In particular, the subjectivity of the linguistic theory of lexical cohesion that has been used in these applications has not been examined empirically.

Speaker: Zhenmei Gu

In information extraction (IE), training statistical models like HMMs (hidden Markov models) usually requires a considerably large set of labeled data. Such a requirement may not be easily met in a practical IE application. In this talk, we investigate how to adapt a fully supervised IE learner to a semi-supervised one so that the learner is able to make use of unlabeled data to help train a more robust model from very limited labeled data.

Speaker: Chris Eliasmith

I present a cognitive model, dubbed BioSLIE, that integrates and extends recent advances in:

• distributed, structure sensitive representation
• neurocomputational modeling
• our understanding of the neuroanatomy of linguistic inference

Speaker: Dana Wilkinson

Given a collection of data points annotated with action labels (obtained, for example, from a robot mounted with a web camera or a laser range finder) we wish to learn a representation of the data which can be used as a map (for planning, etc.).

Speaker: Shai Ben-David

Support Vector machines (SVM's) is one of the most useful and widely applicable machine learning techniques. Each concrete application of SVMs depends on a successful choice of a "kernel matrix". So far, most of the work in this area has focused on developing a variety of kernels and efficient algorithms for employing SVMs with these kernels. Relatively little research attention has been given to the question of how to pick a suitable kernel for any particular learning task at hand.

In this work, we analyze exactly that issue.

Speaker: Pascal Poupart

Reinforcement learning (RL) was originally proposed as a framework to allow agents to learn in an online fashion as they interact with their environment. Existing RL algorithms come short of achieving this goal because the amount of exploration required is often too costly and/or too time consuming for online learning. As a result, RL is mostly used for offline learning in simulated environments. We propose a new algorithm, called BEETLE, for effective online learning that is computationally efficient while minimizing the amount of exploration.

Speaker: David Pal

Stability is a common tool to verify the validity of sample based algorithms. In clustering it is widely used to tune the parameters of the algorithm, such as the number k of clusters. In spite of the popularity of stability in practical applications, there has been very little theoretical analysis of this notion. In this paper we provide a formal definition of stability and analyze some of its basic properties.

Speaker: Radoslav Radoulov

Automatic citation classification is a process that associates each citation reference in a scientific paper with one or more previously defined citation functions. Our goal is to build an automatic classifier that looks for features in the text around a citation reference and attempts to classify the citation based on those features. In this seminar I will present our project on automatic citation classification and a system that we will use to collect labeled citations.

Speaker: Zhenmei Gu

Hidden Markov models (HMMs) are powerful statistical models that have found successful applications in Information Extraction (IE). In current approaches to applying HMMs to IE, an HMM is used to model text at the document level, i.e., the entire document is modeled by an HMM. This modeling might cause undesired redundancy in extraction in the sense that more than one filler is identified and extracted.