Challenges in Natural Language Processing Applications

Join the Buffalo Computer Society Chapter in welcoming Distinguished Visitor, Maiga Chang, on Monday January 29, at 2 PM CST/4 PM ET, when he gives a virtual presentation for our second installment on the Neural Networks of AI.

Prof. Maiga Chang’s research group uses Natural Language Processing techniques including n-grams and part-of-speech tags to analyze large datasets like CORD-19, DBpedia, and Google Books to build an online summary generation system (https://ask4summary.vipresearch.ca/) and Moodle plugin Ask4Summary (https://moodle.org/plugins/block_ask4summary) that are capable of reading articles regarding Coronavirus and course content and provide summary for users’ questions. This talk will explain to audience what and how we have done step-by-step to reach the goal; how the fundamental service like the valid n-grams identifier service (https://ngrampos.vipresearch.ca/) that can help developers and researchers; what are the challenges we have encountered; and what are the plans and potentional ways for investigating and overcoming those challenges.

At the end, Ask4Summary has reached 82.69% success rate for providing quick (0.766 seconds in average) responses of course relevant questions and is available for users in interactive web, Moodle plug-in form, and Python library. With the help of Statistical and Neural NLP, our Authorship Fingerprinting research is capable to correctly distinguish the works created by GPT 3.5, GPT 4, and human authors with recall rate 98.84% in our preliminary study.

(Photos: Maiga Chang at Ask4Summary, POS SNLP chart)

Speaker(s): Dr. Maiga Chang,

Virtual: https://events.vtools.ieee.org/m/405055