ResearchMachine Learning Projects
Group Behavior Analysis in Crowd
Automatic video analysis gets more attention due to an increase in the demand for public safety and physical security in crowd. This research focuses on analysis and recognition of group behaviors in a crowd. For each frame in the video stream, group features are extracted by tracking individuals in a crowd and analyzing their relationships in terms of the distance, direction and speed. Since a video is a sequence of frames/images, we propose an approach based on analyzing the extracted features in sequential structure.
Accelerating the Local Outlier Factor Algorithm (LOF) with Graphic Processing Units (GPGPUs) for Intrusion Detection System
The Local Outlier Factor (LOF) is a very powerful anomaly detection method available in machine learning and classification. The algorithm defines the notion of local outlier in which the degree to which an object is outlying is dependent on the density of its local neighborhood, and each object can be assigned an LOF which represents the likelihood of that object being an outlier. Although this concept of a local outlier is a useful one, the computation of LOF values for every data object requires a large number of k-nearest neighbor queries. This overhead can limit the use of LOF due to the computational overhead involved. Due to the growing popularity of Graphics Processing Units (GPU) in general-purpose computing domains, and equipped with a high-level programming language designed specifically for general-purpose applications (e.g., CUDA), we look to apply this parallel computing approach to accelerate LOF. I explored utilizing a CUDA-based GPU implementation of the k-nearest neighbor algorithm to accelerate LOF classification. I achieved more than a 100X speedup over a multi-threaded dual-core CPU implementation. I also considered the impact of input data set size, the neighborhood size (i.e., the value of k) and the feature space dimension, and report on their impact on execution time.
Machine Learning for Big Data Applications using GPU Cluster
Research focuses on developing machine learning algorithms on huge data using GPU nodes in a cluster. Work on a recommendation algorithm used in Apache Mahout using parallel algorithms on GPUs to process distributed huge data.