Keynote Speakers



Prof. Girish Nath Jha

ICBSLP 2019 is honored to have Professor Girish Nath Jha as a keynote speaker. He will speak on "LT Resource Crunch - challenges and opportunities for South Asian languages"

Prof. Girish Nath Jha teaches Computational Linguistics at the School of Sanskrit & Indic Studies in Jawaharlal Nehru University (JNU) and is currently the Dean of the school. He also holds concurrent appointments in JNU’s Center of Linguistics, Special Center of E-Learning and is an Associated Faculty in the ABV School of Management and Entrepreneurship. Prof Jha was previously the director of JNU’s International Collaboration during 2016-18.

His research interests include Indian languages corpora and standards, Sanskrit and Hindi linguistics, Science & Technology in ancient texts, Lexicography, Machine Translation, e-learning, web based technologies, RDBMS, software design and localization. Details on his work can be obtained from http://sanskrit.jnu.ac.in . Prof Jha has done collaborative research with the Center for Indic Studies, University of Massachusetts, Dartmouth, MA, USA as "Mukesh and Priti Chatter Distinguished Professor of History of Science" during 2009-12, has been a visiting professor at the Yogyakarta State University, Indonesia in 2013. He was awarded DAAD fellowships twice in 2014 and 2016 to teach Computational Linguistics in the Digital Humanities department at University of Würzburg, Germany and has been a visiting Professor at the University of Florence in the summer of 2016.

Prof. Jha did his M.A., M.Phil. and Ph.D. in Linguistics (Computational Linguistics) from JNU and then got another masters degree in Linguistics (specializing in Natural Language Interface) from University of Illinois, Urbana Champaign, USA in 1999. Since then he worked as software engineer and software development specialist in USA before joining JNU in 2002. Prof Jha has books published from publishers like Springer Verlag, Cambridge Scholar Publishing and has over 133 research papers/presentations/publications and over 178 invited talks. Prof Jha has had several consultancies including those from Nuance, Swiftkey, Microsoft Research USA, Microsoft Research India, Microsoft Corporation, Linguistic Data Consortium (University of Pennsylvania), University of Massachusetts Dartmouth, EZDI among others. Prof Jha has completed several sponsored projects for Indian language technology development and has led a consortium of 17 Indian universities/institutes for developing corpora and standards for Indian languages sponsored by Ministry of Electronics and IT (MEITY), Govt. of India.

Prof Jha has been chair/co-chair for at least 13 international seminars/conferences and has been nominated member of more than 30 committees. He was nominated to the editorial board of a leading journal from Springer and has been a reviewer of many leading journals and proceedings in the area of NLP. He has supervised 42 M.Phil. and 50 Ph.D. students. Prof Jha's efforts in collaboration with software industry has led to the development of key technologies for Indian languages including English-Urdu MT for Microsoft Bing Translator, predictive keyboards for several Indian languages by Swiftkey. His awards include Datta Peetha award for Sanskrit linguistics (2017), KECSS Felicitation award for promotion of Sharada script (2016)



Title: Formal language and cross linguistic generalizations

Dr. Niladri Sekhar Dash

Dr. Niladri Sekhar Dash works as Associate Professor in Linguistic Research Unit, Indian Statistical Institute, Kolkata. For the last 25 years, he is working in the area of Corpus Linguistics, Language Technology, Language Documentation and Digitization, Computational Lexicography, Computer Assisted Language Teaching, and Manual and Machine Translation.

Dr. Dash received Visiting Fellowship of the British Academy, UK (2018). To his credit, he has published 17 research monographs and 260 research papers in peer-reviewed international and national journals, anthologies, and conference proceedings.

Dr. Dash acts as Guest Faculty of School of Languages and Linguistics, Jadavpur University, Kolkata; Guest Faculty of Dept. of Linguistics, University of Calcutta, Kolkata; and Visiting Research Fellow of the School of Psychology and Clinical Language Science, University of Reading, UK (2018-2021).

As an invited speaker, Dr. Dash has delivered lectures at more than 40 Universities and institutes in India and abroad. He acts as a Consultant for several multinational organizations working on Language Technology and Natural Language Processing. Dr. Dash has acted as Principal Investigator for 15 projects funded by DeitY, MeitY, & MoSPI, Govt. of India. He is the Editor-in-Chief of Journal of Advanced Linguistic Studies – a peer-reviewed international journal of linguistics and Editorial Board Member for 5 international journals. He is Member of several Linguistic Associations across the world and a regular Ph.D. Thesis Examiner of several Indian and foreign Universities.

At present Dr. Dash is working on Digital Pronunciation Dictionary, Contextualized Word Sense Cognition, Digital Lexical Profile, Parallel Translation Corpus, Bilingual Lexical Database, POS Tagging, Computer Assisted Language Teaching, Endangered Language Documentation and Digitization, and Human and Machine Translation, etc.

Details of Dr. Dash are at: https://sites.google.com/site/nsdashisi/home/



Title: The AIMS Method of Using Contextualised and Decontextualized Information in Word Sense Disambiguation: Some Studies from Bangla

Niladri Sekhar Dash
Linguistic Research Unit
Indian Statistical Institute, Kolkata
Email: ns_dash@yahoo.com

Abstract:

Every natural language has a large set of words which are semantically quite productive. When these words are used in a piece of text, they vary in meaning. It is noted that the context, where these words are used, plays an explicit (rarely implicit) and active (rarely passive) role to influence these forms to generate new context-based functions and meanings. The newly acquired meanings may vary from the original meaning which they usually derive from the source of their origin (i.e., etymology). This phenomenon of words in a natural language creates tremendous problems relating to understanding and describing meanings of words in general; cognition of semantic-cum-lexical functions of words in texts; representation of word meanings in machine learning algorithms; deciphering actual sense of polysemous words in information retrieval and machine translation; training the language learners in the act of learning word meanings; and in presenting the word meanings in dictionaries. In essence, a word of a natural language is a bundle of linguistic information relating to phonology, etymology, morphology, morphophonemics, syntax, lexicology, semantics, morphosyntax, grammar, text, metaphor, discourse, pragmatics, world knowledge and others (Pinker 1995: 344). Therefore, it is not easy to capture all types of information just by looking at its surface form or at its orthography. We require a versatile system along with our native language intuition to decipher all possible explicit and implicit senses of words used in a piece of text. A systematic study of meaning variation of words may help us to establish the notion of semantic indeterminacy and sense gradience (Leech, Francis, & Xu 1994) in the broader domain of language cognition. All these issues are addressed in this present talk with a focus on the Bangla language. It also proposes a cognitive-cum- computational method for understanding the actual contextualized meaning of words with reference to their usage in texts. This method (AIMS) is experimented and validated with a large amount of data retrieved from a Bangla written text corpus.

Fuad Rahman, Ph.D.

ICBSLP 2019 is honored to have Fuad Rahman as a keynote speaker. He will speak on "Building resources for ‘Bangla’— a low-resource language: Challenges and Perspectives"

Fuad Rahman is a Silicon Valley technologist with deep experience in building cutting-edge Big Data analytics solutions in the healthcare and financial sectors. He has extensive background in Big Data modeling and analytics, Natural Language Processing (NLP), computational linguistics, and pattern recognition. His early R&D life was spent working on SBIR and ATP/NIST projects – during which time he worked with all three branches of armed forces: US Air Force, US Navy and US Army. For the last ten years, Fuad has been an entrepreneur with four successful exits under his belt. He is widely published with 90+ publications in peer reviewed journals and conferences. He is an Adjunct Professor at Department of Biomedical Engineering at the University of Arizona.

Fuad Rahman, Ph.D., is an entrepreneur with extensive background in building companies in the area of software technology.

He is the Founder and CEO of Apurba Technologies Inc., a software company based in the Silicon Valley, USA. He is also the Founder of Apurba Technologies Ltd., based in Dhaka, USA and Apurba Technologies Sdn Bhd, based in Kuala Lumpur, Malaysia. Apurba specializes in building software, specifically in the healthcare and financial sectors.

Fuad has founded and co-founded many companies and had four successful exits in the last seven years.

Fuad is an Adjunct Professor at Department of Biomedical Engineering (BME) at the University of Arizona. BME is the newest Engineering department at University of Arizona that reflects a tradition of bleeding-edge interdisciplinary research at the intersection of medicine, science and engineering. This position is aimed to enhance the exchange of information and ideas between academia and industry, to facilitate biomedical engineering research collaborations, to develop funding opportunities, and specially to support undergraduate education in biomedical engineering at the university.

Before founding Apurba, Fuad was the Vice President of Product Development at KnewCo, a leader in semantics based knowledge discovery and In-Text advertising solutions. Before that, Fuad worked as the Research Team Lead at QuickSchools Inc., a leading innovator of cloud-based agile Software as a Service (SaaS) development platforms, where he was responsible for creating the product vision, identifying critical priorities, and establishing a customer-facing platform for its rapidly growing product set. Prior to that, he was a Research Lead and Product Manager at BCL Technologies Inc., where he led the R&D team, and won grants from US Army, US Navy, and US Air Force on topics related to NLP, information retrieval, text mining, and summarization. He has also served as an Algorithm Lead at Handisplay Inc., leading cutting-edge research projects focusing on core algorithm development that has commercial applications in document layout analysis, web page analysis, web page clustering and semantic clustering of free-format natural language. Early in his career, he designed algorithms for mapping web page content to cell phone browsers at WiredPocket Inc. as a Research Engineer.

Fuad has published widely in the scientific journals. His published body of work comprises of over 90 technical peer reviewed articles in the scientific literature including 6 book chapters and 24 journal papers. He has worked on the editorial boards of technical journals, acted as referee to many journals, and worked as technical committee member in many international conferences.

Fuad is an accomplished public speaker. He has made numerous presentations not only to technical peer-reviewed conferences on core and applied research, but also in trade shows showcasing software, connecting to technical and non-technical audience. He also worked on multiple defense projects and presented his work to US Navy, Army and Air Force, visiting a lot of US military facilities.

Fuad was a member of the Domain Steering Committee of the XBRL.US for a 2-year term between 2013-2014 that oversees the development of taxonomies to meet the business reporting needs of key markets in the United States. In the area of regulatory compliance, this is a singular honor, because this puts him in a very select group of industry leaders in the US and the world.

Fuad has a Ph.D. in Ubiquitous Computer Vision and Pattern Recognition from University of Kent in England.



Building resources for ‘Bangla’— a low-resource language: Challenges and Perspectives

Fuad Rahman, PhD

Abstract:

Although a top ten ‘global’ language in terms of number of people speaking it, resources for Bangla language processing is surprisingly hard to find. This talk will focus on what few resources are available and what is not. It will then offer a case study on Bangla stemmers and the challenges to build even such a primary resource. The talk will end with a broader perspective on the state of the art of these resources and the ways to address this in an organized and efficient way.