Our research is focused on advancing the methods (e.g., improving the quality and speed). Key research areas include: (i) cluster analysis (ii) anomaly detection (iii) intrinsic dimensionality and data complexity (iv) topic modeling (v) event detection (vi) similarity search in databases.

Cluster Analysis

Unsupervised discovery of typical patterns in unlabeled data.

Anomaly Detection

Unsupervised discovery of atypical instances in unlabeled data

Intrinsic Dimensionality

A measure of data complexity related to the expansion rate and smoothness of data.

Topic Modeling

Unsupervised modeling of textual data in topical patterns, along with a topic explanation.

Event Detection

Detection of change in behavior of temporal sequences such as textual data streams.

Similarity Search

Efficient search of similar objects in high-dimensional data sets.