Deep learning provides new fundamental tools, such as contextualised word embeddings and seq2seq models, that let us build new kinds of Natural Language Understanding apps faster, better and cheaper than ever before. The advanced pattern-matching capabilities of deep learning enable a new approach to app development where the system's behaviour is learnt from training data, dramatically reducing the need for manual scripting. This talk describes how we are using this technology in the Oracle Digital Assistant, focusing especially on Conversational AI. The talk ends with a discussion of how research advances in areas such as explainability, few-shot learning, data augmentation and transfer learning can help this technology achieve its full potential.
Mark Johnson is Chief AI Scientist, Oracle Digital Assistant at Oracle Corporation, and a Professor of Language Sciences in the Department of Computing, Macquarie University. He’s also an Editor in Chief for the Transactions of the Association for Computational Linguistics. Mark has worked on a wide range of topics in computational linguistics, but his main area of research is natural language understanding, especially syntactic parsing and semantic analysis, and their applications to text and speech processing.
Nicholas Evans is Director of the ARC Centre of Excellence for the Dynamics of Language (CoEDL), a Distinguished Professor of Linguistics at the College of the Asia Pacific, ANU and an ARC Laureate Professor. His contributions to linguistics include documentation of fragile and little-known languages in Australia and New Guinea based on over six years of fieldwork, studying the implications of little-known languages for linguistics and creating a framework for linguistic typology to enable systematic comparison of languages. He has also done applied work as a linguist, anthropologist and interpreter in areas ranging from Native Title, traditional ecological knowledge, promotion of Indigenous art in Australia, and vernacular education in Australia, PNG and elsewhere
Ben Foley is project manager of CoEDL's Transcription Acceleration Project (TAP). TAP brings cutting-edge language technology within reach of people working with some of the worlds oldest languages. Ben's previous experience with Aboriginal and Torres Strait Islander language resource development has resulted in apps and websites galore, including Iltyem-Iltyem and Gambay First Languages Map.