Talk Time: 1100 to 1200 Hrs
Venue: Seminar Hall, KICS, UET, Lahore
Speaker: Ms. Amara Tariq
Multimedia datasets are growing rapidly with the expansion of the World Wide Web. Systems involving collaborative knowledge mining across heterogeneous data types are gaining importance. Image and text are two very common and important data types. Intelligent search, retrieval and organization systems often require pattern mining and relation extraction within and across these two data types. Large image and text datasets hold key background semantic or context information for various elements of the datasets such as images, words, or named entities mentioned in the text. Effective extraction and incorporation of this context is necessary to build practical information search and retrieval systems. In this talk, I shall present ongoing work in various image- and text-related applications such as automatic annotation of images with text and dynamic relation extraction among named entities. These applications deal with various types of data simultaneously. In most cases, background semantics for instances of a certain modality of data are defined over instances of another modality of data. I shall discuss the use of various image and language processing techniques to establish contextual relationship across heterogeneous data types and their effective incorporation in practical systems.
About the Speaker:
Amara Tariq is a doctoral student at Computational Imaging Lab. in University of Central Florida (UCF), USA. She received M.S. Computer Science degree from UCF in 2014. She also received B.E. degree from University of Engineering and Technology (UET), Lahore, in 2008, and M.S. Computer Engineering degree from Lahore University of Management Sciences (LUMS) in 2011. She was awarded Fulbright scholarship for doctoral studies in 2011. Her research interests include computer vision, natural language processing, machine learning and artificial intelligence.