Continual Learning | Federated Learning | Representation Learning

Natural Language Processing | Deep Learning

CURRENT AFFILIATIONS

  • PhD Student @ CIS, LMU
  • Applied AI-NLP Scientist @ DRIMCo GmbH

EMAIL

firstname.lastname@drimco.net 

Yatin Chaudhary

I am a PhD student at the Center for Information and Language Processing (CIS) at the Ludwig Maximilian University of Munich (LMU). I am also currently working (full-time) as an Applied AI-NLP Scientist at DRIMCo GmbH, Munich in the area of Natural Language Processing (NLP).  I am co-supervised by Prof. Dr. Hinrich Schütze, CIS, LMU and Dr. Pankaj Gupta, DRIMCO GmbH. I completed my Master's thesis under the co-supervision of Prof. Dr. Thomas Runkler and Dr. Pankaj Gupta at Machine Intelligence (MIC), Siemens Corporate Technology. My research focuses on developing Representation Learning methods/algorithms (extending to Continual/Lifelong Learning and Federated Learning paradigms) using deep neural topic and language models for natural language text understanding under high and low-resource data settings. I work on research as well as applied projects and one of the research goals is the "easy integration and effective results" in AI-based products. 

EDUCATIONAL BACKGROUND

PUBLICATIONS


Title:             Federated Continual Learning for Text Classification via Selective Inter-client Transfer

Authors:     Yatin Chaudhary, Pranav Rai, Matthias Schubert, Hinrich Schütze, Pankaj Gupta

Venue:         Published as a conference paper EMNLP 2022   |   [Paper]   [Code]


Title:             Multi-source Neural Topic Modeling in Multi-view Embedding Spaces 

Authors:     *Pankaj Gupta, *Yatin Chaudhary, Hinrich Schütze 

Venue:         Published as a conference paper NAACL 2021   |   [Paper]   [Code]


Title:             TopicBERT for Energy Efficient Document Classification

Authors:     *Yatin Chaudhary, *Pankaj Gupta, Khushbu Saxena, Vivek Kulkarni, Thomas Runkler, Hinrich Schütze

Venue:         Published as a conference paper at EMNLP 2020   |   [Paper]   [Code]


Title:             Explainable and Discourse Topic-aware Neural Language Understanding

Authors:     Yatin Chaudhary, Hinrich Schütze, Pankaj Gupta

Venue:         Published as a conference paper at ICML 2020   |   [Paper]   [Code]


Title:             Neural Topic Modeling with Continual Lifelong Learning

Authors:     Pankaj Gupta, Yatin Chaudhary, Hinrich Schütze 

Venue:         Published as a conference paper at ICML 2020   |   [Paper]   [Code]


Title:            Lifelong Neural Topic Learning in Contextualized Autoregressive Topic Models of Language via Informative Transfers

Authors:    Yatin Chaudhary, Pankaj Gupta, Thomas Runkler 

Venue:         Master's Thesis (Technical University Munich) 2019   |   [Thesis]


Title:            BioNLP-OST 2019 RDoC Tasks: Multi-grain Neural Relevance Ranking Using Topics and Attention Based Query-Document-Sentence Interactions

Authors:    *Yatin Chaudhary, *Pankaj Gupta, Hinrich Schütze

Venue:         Published as a workshop paper at EMNLP-IJCNLP 2019    |   [Paper]   [Code]


Title:            textTOvec: Deep Contextualized Neural Autoregressive Topic Models of Language with Distributed Compositional Prior

Authors:    Pankaj Gupta, Yatin Chaudhary, Florian Buettner, Hinrich Schütze

Venue:         Published as a conference paper at ICLR 2019   |   [Paper]   [Code]


Title:            Document Informed Neural Autoregressive Topic Models with Distributional Prior

Authors:    Pankaj Gupta, Yatin Chaudhary, Florian Buettner, Hinrich Schütze

Venue:         Published as a conference paper at AAAI 2019    |   [Paper]   [Code]