Summer Research Internship, IBM Research Lab

May 2022 - August 2022

Gurgaon, Haryana

Explaining Neural Response Generation using Summarization

• The main aim of this project is to explain how the transformer-based pretrained model generate response using summary of the dialogues and to explain why summary is useful to generate response.
• Performed data analysis on 5 different dialogues based datasets.
• Trained a transformer-based pretrained BART model to generate next response using summary of the dialogues.
• Using Reinforcement learning technique, we trained pre-trained summarizer further and generate response and update the parameters of the summarizer based on the reward generated

Guide: Dinesh Raghu¹, Sachindra Joshi², Senior Research Scientist¹, Senior Manager², at IBM Research, India

Software Engineer, Cognam Tech.

Aug 2020 - July 2021

Dwarka, New Delhi

Cognam Technologies Pvt. Ltd.

• Worked with ML Team on different financial-based projects like Face Recognition, Forecasting, etc.
• Worked on the model performance and built a robust system.
• R&D new machine learning approach and methods to implement it on financial datasets.

Winter Research Internship, TCS Research & Innovation

Dec 2019 - Jan 2020

Kolkata, West Bengal

Gestures Recognition for Human-Robot Interaction

• Build a monocular vision based robust gesture identification network for natural human-robot interaction.
• Prepared own datasets of commonly used gestures while interaction between human-human.
• Designed a system which recognizes gesture as an input and as per identified gesture command will be given to robot for its movement.

Guide: Dr. Chayan Sarkar, Scientist at TCS Research & Innovation

Summer Research Internship, IIT Bombay

May 2019 - July 2019

Powai, Mumbai

Multi-Task Learning For Automated Essay Scoring

• Designed a deep multi-task learning model for Automated Essay Scoring.
• Enhanced the baseline model with multi-task learning, using the attribute scores of the essays like language fluency, vocabulary, structure, organization, content, etc.
• The multi-task learning model performed well in all 8 prompts of the ASAP++ competition datasets.

Guide: Prof. Pushpak Bhattacharyya, Department of Computer Science and Engineering, IIT Bombay

Master of Technology (Computer Science & Engineering)

2021 - Present

IIT Kanpur

Indian Institute of Technology Kanpur is a public technical and research university located in Kanpur, Uttar Pradesh, India. It was declared to be an Institute of National Importance by the Government of India under the Institutes of Technology Act.

Bachelor of Technology (Electrical Engineering)

2016 - 2020

IIT Patna

Indian Institute of Technology Patna is an autonomous institute of education and research in science, engineering and technology located in Patna, India. It is recognized as an Institute of National Importance by the Government of India.

Student Mentor

July 2018 - July 2019

Students' Gymkhana, IIT Patna

Student Mentorship Program (SMP) - A 1-year program which has been launched with the aim of providing a platform for 1st-year students to interact with seniors from the same department and to motivate students to excel in both academic and non-academic fields and to make the most of their life at IIT Patna and beyond.

Sub-Coordinator

April 2017 - April 2018

Sparkonics, Electrical Club at IIT Patna

Sparkonics - An official electric club where different electrical projects are taken and motivate juniors to participate based on interests and complete them within limited time period. I was in the Learning division under Sparkonics. At its core, I personally believe on being a mentor is being a trusted advisor. I involved myself available to support and advise my juniors when they need it , and always keeping their best interests in mind.

Many Hands Make Light Work: Using Essay Traits to Automatically Score Essays

NAACL'22: North American Chapter of the Association for Computational Linguistics

Rahul Kumar, Sandeep Mathias, Sriparna Saha, Pushpak Bhattacharyya

July 2022

Most research in the area of automatic essay grading (AEG) is geared towards scoring the essay holistically while there has also been some work done on scoring individual essay traits. In this paper, we describe a way to score essays holistically using a multi-task learning (MTL) approach, where scoring the essay holistically is the primary task, and scoring the essay traits is the auxiliary task. We compare our results with a single-task learning (STL) approach, using both LSTMs and BiLSTMs. We also compare our results of the auxiliary task with such tasks done in other AEG systems. To find out which traits work best for different types of essays, we conduct ablation tests for each of the essay traits. We also report the runtime and number of training parameters for each system. We find that MTL-based BiLSTM system gives the best results for scoring the essay holistically, as well as performing well on scoring the essay traits. The MTL systems also give a speed-up of between 2.30 to 3.70 times the speed of the STL system, when it comes to scoring the essay and all the traits.

Multi-objective Cuckoo Search based Streaming Feature Selection for Multi-label Dataset

ACM Transactions on Knowledge Discovery from Data

Rahul Kumar, Dipanjyoti Paul, Sriparna Saha and Jimson Mathew

January 2021

In this paper, we develop streaming feature selection methods for multi-label data where the multiple labels are reduced to a lower-dimensional space. The similar labels are grouped together before performing the selection method to improve the selection quality and to make the model time efficient. The multi-objective version of the cuckoo search-based approach is used to select the optimal feature set. The proposed method develops two versions of the streaming feature selection method: ) when the features arrive individually and ) when the features arrive in the form of a batch. Various multi-label datasets from various domains such as text, biology, and audio have been used to test the developed streaming feature selection methods. The proposed methods are compared with many previous feature selection methods and from the comparison, the superiority of using multiple objectives and label co-relation in the feature selection process can be established.

Online Feature Selection for Multi-label Classification in Multiobjective Optimization Framework

ASONAM '19: Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

Rahul Kumar, Dipanjyoti Paul, Sriparna Saha and Jimson Mathew

August 2019

This paper addresses the online feature selection problem in multi-label classification framework where multi-labelled data with features arriving in an online fashion is considered as input. The proposed approach works in two phases, in the first phase, the best subset of features is selected from the initial available set of features using a multi-objective optimization (MOO) based feature selection technique. In the second phase of the proposed feature selection technique, a newly arrived feature is accepted or rejected based on redundancy with respect to the already selected set of features and relevancy of the arrived feature with respect to the class labels. In order to show the efficacy of the proposed algorithm, it is tested on 7 various types of multi-label datasets of different domains such as text, biology, and audio. The obtained results outperform the results obtained by state-of-the-art approaches in majority of the cases.

Data Structure & Algorithm

C

C++

Java

Machine Learning Tools

Python

Keras

TensorFlow

PyTorch

Scikit-learn

Other Tools

Git & GitHub

Linux OS

Postman

Matlab

Web Development

HTML

CSS

Bootstrap

React JS

JavaScript

Backend Development

Node.js

Express.js

Databases

PostgreSQL