Chandan R. Akiti

State College, PA

I am a deep learning engineer passionate about building robust intelligent systems. I am pursuing M.S. in Computer Science at Penn State University, focusing on Deep Learning, NLP and Computer Vision.

My thesis advisor is Prof. Arzoo Katiyar. I am working on Encoder-Decoder model-based Few-Shot Learning for Named Entity Recognition.

Currently, I am also a Graduate Research Assistant working with Prof. Sarah Rajtmajer and Prof. Anna Squicciarini. I work on natural langauge understanding, especially Frame Semantics-based Information Extraction models for analysis of Information Disclosure of Twitter users during Pandemic.

I also have four years of work experience in Samsung Research India as SDE, where I worked on exciting research projects in deep learning and remote-healthcare for 2 years. I completed my B.Tech in Computer Science and Engineering at Indian Institute of Technology, Madras.

news

Apr 19, 2021	Student reviewer for College of Engineering Research Symposium 2021.
Apr 19, 2021	Placed 445th in Codechef April Long Challenge 2021, username chandan047. I am back to Division 1 on Codechef!
Apr 12, 2021	Placed 597th in Google Kickstart 2021 Round B, username chandan047.
Mar 14, 2021	Placed 5th in CodePSU 2021, username sippy_coder - $125 cashprize .
Sep 15, 2020	Our paper A Semantics-based Approach to Disclosure Classification in User-Generated Online Content is accepted to the Findings of EMNLP 2020. Conference to be held on Nov 16th, 2020.
Feb 7, 2020	Our preliminary work on Contextual Representation of Self-Disclosure and Supportiveness in Short Text is published in Proceedings of AAAI-20 Workshop on Affective Content Analysis

selected publications

EMNLP

A Semantics-based Approach to Disclosure Classification in User-Generated Online Content

Akiti, Chandan, Squicciarini, Anna, and Rajtmajer, Sarah

In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: Findings 2020

Abs PDF

As users engage in public discourse, the rate of voluntarily disclosed personal information has seen a steep increase. So-called self-disclosure can result in a number of privacy concerns. Users are often unaware of the sheer amount of personal information they share across online forums, commentaries, and social networks, as well as the power of modern AI to synthesize and gain insights from this data. This paper presents an approach to detect emotional and informational self-disclosure in natural language. We hypothesize that identifying frame semantics can meaningfully support this task. Specifically, we use Semantic Role Labeling to identify the lexical units and their semantic roles that signal self-disclosure. Experimental results on Reddit data show the performance gain of our method when compared to standard text classification methods based on BiLSTM, and BERT. In addition to improved performance, our approach provides insights into the drivers of disclosure behaviors.