Artificial Intelligence

Learning with less dependency on labeled data for cyberbullying detection and image classification tasks

Thursday, November 05, 2020, 10:00am - 11:00am PSTiCal
This event is open to the public.
AI Seminar
Elaheh Raisi (Brown University)

My research interests span in the areas of machine learning, data mining, and computational social science. As a postdoc at Brown University, I am working on designing algorithms to learn when we do not have enough labeled data or information about the label of data. For problems such as fine-grained visual categorization that suffer from lack of labeled data, we introduced a multi-task learning based model to jointly train the target and other related tasks simultaneously. To construct the auxiliary data, we leverage a knowledge graph to query for semantically related concepts that are grounded in labeled images (accepted in VL3 workshop at CVPR 2020). I am extending this work to include more related auxiliary data to the target task to improve the performance.

I received my Ph.D. from the computer department at Virginia Tech in May 2019. My primary research objective was to address the computational challenges associated with designing machine learning approaches for harassment-based cyberbullying detection. We developed a weakly supervised framework, co-trained ensemble, in which two learning algorithms co train one another, seeking consensus on whether examples in unlabeled data are cases of cyberbullying or not. For this research, I won a best paper award at the IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM) 2017 and a best paper award at the Learning with Limited Labeled data (LLD) workshop at NIPS, 2017. Additionally, I received an honor award from the Computer Science department at Virginia Tech for good work in 2017, and best graduate student poster presentation award at the Virginia Tech SAIC Integrated Security Colloquium, 2018. 

« Return to Events