CIC Visitor Seminar August 2018
Applied Text Analytics: the need for human information interaction and cognitive approaches in learning contexts
Presented by Dr Andrew Gibson
In the last decade there have been significant advances in Natural Language Processing (NLP). At the data level, cloud computing and techniques for processing Big Data have enabled us to engage with unstructured text in new ways. Advances in neural inspired algorithms have resulted in new approaches to constructing language models. Vector space approaches continue to produce many algorithms for working with semantics. The increasing work in Natural Language Understanding (NLU) has spawned a new breed of intelligent agents and chatbots.
However, amidst the incredible advances, there is a more sobering picture when NLP is applied in real human contexts. Many of the best approaches still fail to live up to human expectations when released in the complex contexts of humans interacting with language.
In this presentation Andrew Gibson will explore some of these issues in the application of NLP to learning contexts, and will provide provocations on how this area may be advanced. Drawing on his experience in Text Analytics for learning, Andrew will outline his opinion on why a simultaneously abductive and pragmatic approach is required, and what such an approach might look like. Supported with examples from recent work, he will argue that to make significant advances in future, the field will need to include both models of human cognition and new processes for the interaction between people and information.
Andrew Gibson is a Lecturer in Information Science at QUT. His primary research interest is in transdisciplinary understandings of fields where people and technology interact, in particular the field of Learning Analytics. He has developed this interest theoretically in Transepistemic Abduction, a specialised mode of reasoning, and practically in Reflective Writing Analytics. Andrew has initiated open source software projects to support his research work, notably GoingOK, and TAP (Text Analytics Pipeline). He has delivered workshops and talks locally and internationally on Text Analytics, and previously held a position of Research Fellow in Writing Analytics at the Connected Intelligence Centre, University of Technology Sydney (UTS). Originally a secondary school music teacher, specialising in music technology, Andrew has also held positions in IT management, creative production management, and management of a small tech hardware start-up. Andrew holds a PhD in Information Science, a Bachelor degree in Educational Studies, Postgraduate Diploma in Information Technology, and Diploma in Teaching.