Amarnath Pathak

Mr Amarnath Pathak

University: National Institute of Technology (NIT) Mizoram



Research ID Profile URL:


Part-Of-Speech (POS) tagging of code-mixed content is mandatory prior to its productive utilization in NLP application domains. To this end, a Hidden Markov Model (HMM) based POS tagger has been implemented for POS tagging of code-mixed Indian Social Media Text. One of my co-authored papers [6], describing implementation of HMM based POS tagger, fetched Third Best Paper Award at International Conference on Social Transformation-Digital Way , Science City, Kolkata, Jan 19-21, 2018. Moreover, during my postgraduate research (2013-15), he researched on the role of evolutionary algorithms (particularly, Ant Colony Optimisation) in extracting exceptions along with the classification rules, from the given training data [7-10]. Discovery of exceptions adds interestingness to the rules and it provides an opportunity to amend one’s decision under exceptional circumstances. Mr Pathak has peer reviewed 10 publications.

Mr Amarnath Pathak has been carrying out research in the domain of Natural Language Processing (NLP) and its applications since the time (01/11/2016) of his selection as Junior Research Fellow, in the Dept. of CSE, NIT Mizoram, for working on DST-SERB sponsored project entitled “An Application of Textual Entailment and Semantic Textual Similarity in Scientific Document Retrieval Systems”. Project falls under ICT research domain and it concerns retrieval of mathematical formulae from scientific document, a fascinating research topic which has grabbed researchers’ attention worldwide. He contributed to the domain in form of a novel substitution tree-based indexing technique and a math formula embedding based retrieval system which offer improved performance over conventional text-search based systems. Proposed methodologies and the system outcomes have been published in reputed journal and proceedings of reputed conferences.