Research Scientist (Siemens) | 8+ Years Experience | PhD Scholar (University of Munich, LMU) Germany
Deep Learning | Machine Learning | Natural Language Processing (NLP) | Research-cum-Applications
CURRENT AFFILIATIONResearch Scholar @CIS University of Munich (LMU), GermanyResearch Scientist @Siemens AG, Munich
EXPERIENCEProfessional Hands-on (8.5+ years)Research (4.5+ years)
Research InterestNATURAL LANGUAGE PROCESSINGInformation Extraction, Entity-Relation Extraction, Static and Dynamic Topic Modeling, Representation Learning, Semantic Textual Similarity, Document Layout Analysis, etc.
MACHINE LEARNINGDeep Learning, Recurrent Neural Networks, Probabilistic Graphical Models, Unsupervised Representation Learning, Semi-supervised Bootstrapping, Transfer and Domain learning, Lifelong Learning, Multi-tasking, Explainable AI, etc.
ACTIVE RESEARCHRelation Extraction, Topic Modeling, Representation Learning.
Mr. Pankaj Gupta is a final-year PhD scholar in Computer Science / Computational Linguistics advised by Prof. Hinrich Schütze at University of Munich (LMU), Germany. Additionally, he is working as Research Scientist in an exciting Research Group Machine Intelligence with Dr. Ulli Waltinger at Siemens AG, Munich Germany, where Pankaj leads AI-driven NLP and applies his research outcomes into challenging industrial applications. Mr. Gupta also offers lectures about "Deep Learning and Artificial Intelligence" at LMU. During Oct 2016-Jan 2017, he did a PhD research internship at IBM Research Zurich with Dr. Angela Fahrni and Dr. Abdel Labbi.
Pankaj Gupta earned Masters (Computer Science) with Distinction from Technical University of Munich, Germany in Nov 2015. His Master thesis titled: "Deep Learning Methods for the Extraction of Relations in Natural Language Text " was supervised by Prof. Thomas Runkler, Dr. Heike Adel, Dr. Bernt Andrassy, Dr. Hans-Georg Zimmermann, and Prof. Hinrich Schütze. Before coming to Munich, he worked with Wipro and Aricent Technologies, India as Senior Software Developer (C programmer) during 2010-2013. In 2010, he obtained Bachelor of Technology in Information Technology with Distinction from Amity University, Noida India. His Bachelor thesis is "Summarizing text by ranking text units according to shallow linguistic features".
Pankaj Gupta is particularly interested in Deep learning methods to challenge Natural Language Processing (NLP) tasks, such as information extraction, entity-relation extraction, topic modeling, representation learning, semantic textual similarity, etc. Towards NLP progress, Mr. Gupta likes building neural computational models, for instance, composite neural language models at the intersection of the probabilistic graphical models and deep learning techniques, especially in unsupervised learning paradigm. In doing so, he aims at improving language representations by the fusion of global and local semantics, respectively using neural topic and language models in the realms of transfer as well as lifelong (or never-ending) learning settings. The explainable AI for analysing and interpreting neural networks for natural language tasks is one of his ongoing explorations.
Pankaj Gupta actively contributes in NLP and machine learning conferences and have peer-reviewed publications in the top tier conferences, such as AAAI, ICLR, NAACL, EMNLP, COLING, etc. He has served as program committee (PC) member for several conferences, for instance, ACL, EMNLP, etc. If interested, please checkout his talks and lectures here.
Pankaj Gupta often looks for good master and PhD students in research and applications. If interested in working student, inter-disciplinary project, master thesis, intern or PhD position, please write him.
Checkout Mr. Gupta's current/past students here.
- [Mar-April, 2019] TWO awesome candidates (Usama Yaseen and Yatin Chaudhary) starting their PhD @Siemens, advised by me
- [Jan 30, 2019] TWO oral talks ("iDocNADEe" and "iDepNN") and TWO posters ("iDocNADEe" and "iDepNN") presented at AAAI-2019 at Honolulu, Hawaii USA
- [Jan, 2019] 2nd Patent published. Title: "Method and system for automatic discovery of topics and trends over time " (No: EP 3432155)
- [Nov 29, 2018] Offering guest lecture about "Representation and Distributional Learning" at LMU, Germany
- [Nov 24, 2018] One long paper ("textTOvec") accepted in ICLR-2019 conference. Acceptance rate: 31.4% (500/1591)
- [Nov 14, 2018] Offering guest lecture about "Recurrent Neural Networks" at LMU, Germany
- [Oct 31, 2018] TWO long papers ("iDocNADEe" and "iDepNN") accepted in AAAI-2019 conference. Acceptance rate: 16.2% (1150/7095)
- [Oct 31, 2018] Poster presentation for our paper ("LISA") at EMNLP-2018 BlackBoxNLP workshop at Brussels, Belgium
- [Aug 19, 2018] Oral talk for our paper ("Replicated Siamese LSTM") delivered at COLING-2018 at New Mexico, USA
- [Jun 07, 2018] 1st Patent published. Title: "Device and Method for Natural Language Processing" (No: US 15/498517)
- [Jun 01, 2018] TWO oral talks ("RNN-RSM" and "JBM") delivered at NAACL-HLT-2018 at New Orleans, USA. Talk1 Talk2
- [Feb 14, 2018] TWO long papers ("RNN-RSM" and "JBM") accepted in NAACL HLT-2018 conference. Acceptance rate: 32.0% (207/647)
- [Mar-Oct, 2017] Paper rejections. Working on scientific writings, advised by Prof. Hinrich Schütze and Dr. Florian Büttner
- [Sep 25-27, 2017] Attending Google's 2nd Natural Language Processing Summit at Google Zurich, Switzerland
- [May, 2017] Reviewing papers for EMNLP-2017
- [May, 2017] Pitching/Invited Talk at Young Research Forum at Siemens AG
- [Dec 11-16, 2016] Oral talk for our paper ("TF-MTRNN") delivered at COLING-2016 at New Mexico, USA
- [Oct 2016-Jan 2017] PhD research internship at IBM Research Zurich
- [Sept 20, 2016] One long paper ("TF-MTRNN") accepted in COLING-2016 conference. Acceptance rate: 32.4% (337/1039)
- [March, 2016] One short paper ("Ranking-RNN") accepted in ICASSP-2016 conference. Acceptance rate: 47.0% (1265/2682)
- [March, 2016] One short paper accepted in NAACL HLT-2016 conference. Acceptance Rate: 28.9% (82/284)
- [Dec 2015] Starting PhD, advised by Prof. Hinrich Schütze at University of Munich (LMU) and Research Group Machine Intelligence at Siemens AG
- [Nov 2015] Submitting Master Thesis titled: "Deep Learning Methods for the Extraction of Relations in Natural Language Text " supervised by Prof. Thomas Runkler, Dr. Heike Adel, Dr. Bernt Andrassy, Dr. Hans-Georg Zimmermann, and Prof. Hinrich Schütze