Hi! I am Lahari. Welcome to my homepage. I work as an Applied Scientist in the NLP research team at Amazon Berlin. I obtained my PhD degree in Computer Science from National University of Singapore. My thesis focuses on analyzing user generated contents through the lens of opinion mining, and consensus modeling. My research interests include solving real world NLP/IR problems with Probabilistic Graphical Models, and Deep Learning.

Download Resume

Publications

Deploying a retrieval based response model for task oriented dialogues [pdf]
Lahari Poddar , Gyuri Szarvas, Cheng Wang, Georges Balazs, Pavel Danchenko, Patrick Ernst
EMNLP 2022 , Abu Dhabi, UAE

Calibrating imbalanced classifiers with focal loss: An empirical study [pdf]
Cheng Wang, Georges Balazs, Gyuri Szarvas, Patrick Ernst, Lahari Poddar , Pavel Danchenko
EMNLP 2022 , Abu Dhabi, UAE

DialAug: Mixing up dialogue contexts in contrastive learning for robust conversational modeling [pdf]
Lahari Poddar , Peiyao Wang, Julia Reinspach
COLING 2022 , Gyeongju, Republic of Korea

Attention Guided Dialogue State Tracking with Sparse Supervision [pdf]
Shuailong Liang, Lahari Poddar, György Szarvas
ArXiv Preprint 2021

Methods for Improving Usability of Online User Generated Content [pdf]
Lahari Poddar
2019, PhD Dissertation

Predicting User Reported Symptoms Using a Gated Neural Network [pdf]
Lahari Poddar , Wynne Hsu, Mong Li Lee
ICTAI 2019 , Portland, USA

Train One Get One Free: Partially Supervised Neural Network for Bug Report Duplicate Detection and Clustering [pdf]
Lahari Poddar, Leonardo Neves, William Brendel, Luis Marujo, Sergey Tulyakov, Pradeep Karuturi
NAACL 2019 , Minneapolis, USA

A Probabilistic Framework for Learning Domain Specific Hierarchical Word Embeddings [pdf]
Lahari Poddar, György Szarvas, Lea Frermann
ArXiv Preprint 2019

Predicting Stances in Twitter Conversations for Detecting Veracity of Rumors: a Neural Approach [pdf]
Lahari Poddar, Wynne Hsu, Mong Li Lee, Shruti Subramaniyam
ICTAI 2018 , Volos, Greece     Best Student Paper Award!

Cold Start Thread Recommendation as Extreme Multi-label Classification [pdf][code]
Kishaloy Halder, Lahari Poddar, Min-Yen Kan
XMLC for Social Media, WWW 2018 , Lyon, France

Author aware Aspect Topic Sentiment Model to Retrieve Supporting Opinions from Reviews [pdf] [code and data]
Lahari Poddar, Wynne Hsu, Mong Li Lee
EMNLP 2017 , Copenhagen, Denmark

Modeling Temporal Progression of Emotional Status in Mental Health Forum: A Recurrent Neural Net Approach [pdf] [code and data]
Kishaloy Halder, Lahari Poddar, Min-Yen Kan
WASSA, EMNLP 2017 , Copenhagen, Denmark

Quantifying Aspect Bias in Ordinal Ratings using a Bayesian Approach [pdf]
Lahari Poddar, Wynne Hsu, Mong Li Lee
IJCAI 2017 , Melbourne, Australia

IndoNet: A Multilingual Lexical Knowledge Network for Indian Languages [pdf]
Brijesh Bhatt, Lahari Poddar, Pushpak Bhattacharyya
ACL 2013 , Sofia, Bulgaria

Education

National University of Singapore (NUS)

PhD student in Computer Science • August 2014 - July 2019

Advised by : Prof. Wynne Hsu and Prof. Mong Li Lee

Indian Institute of Technology, Bombay (IIT Bombay)

M. Tech in Computer Science • July 2011 - July 2013

Advised by : Prof. Pushpak Bhattacharyya
Thesis Title: Multilingual Multiword Expressions [pdf]

West Bengal University of Technology (WBUT)

B. Tech in Computer Science • July 2007 - July 2011

Work Experience

Amazon

Applied Scientist • September 2019 - Now

NLP Research

Microsoft

Software Development Engineer • July 2013 - July 2014

Relevance and Ranking team, Bing Ads

Amazon

Applied Scientist Intern • May 2018 - July 2018

NLP research team

Snapchat

Research Intern • August 2018 - December 2018

NLP research team, Snap Research

Teaching Experience

  • • Teaching Assisstant for CS1020 Data Structures and Algorithms , Sem II 2014-2015
  • • Teaching Assisstant for CS3201 Software Engineering Project , Sem I 2015-2016, Sem II 2015-2016, Sem I 2016-2017

Get In Touch.

If you think there are intersections between your research interests and mine, feel free to get in touch! Thanks for reading this far... Have a good day!