How ChatGPT and Related AI Tools work

Click here to watch the recording of the seminar.
ChatGPT as a multi-turn conversational agent has stunned end-users world-wide from all trades with its near human language capabilities. The abilities in generating executable code and writing with coherent arguments for practically any topic take the public imaginations to even start contemplating the future of AI embedded workforce.
This workshop provides the first forum for our young academics, from Dept. of Computer Science and Software Engineering, researching into the areas of deep learning for Language and Vision, to exchange ideas and thoughts, with the objectives to

• ground the discussions to focus on the machine learning models and techniques underpinning ChatGPT and related AI tools;
• build some basic understandings in the wider community of how data-driven AI is obtained through deep learning;
• unveil the technicalities, challenges and limitations to identify future directions of research.

• AI and ML researchers who would like to join the discussion on what makes ChatGPT so successful (at least for near perceivable future) and what we can learn from it from multiple angles: human computer interface, transfer learning, deep reinforcement learning from human preference ..., just to name a few.
• General audience in the University community who are interested in knowing more about the under-the-hood technologies of ChatGPT and related AI tools.

• Dr Caren Han: Transformer, GPT, and ChatGPT Explained
• Sirui Li: The Journey of OpenAI GPT Models
• Dr Naveed Akhtar: Fooling Deep Learning in Computer Vision and Addressing it with XAI
• Dr Naeha Sharif: From Images to Text to Text to Images
• Prof. Zach Aman: Fine-tuning GPT-3 on textbooks to build a tutorial bot for a level 2 class at UWA

Knowledge Extraction and Visualization an Enablers of Digital Content Value Chains

Click here to watch the recording of the seminar.
Seminar Abstract
Opinions about current and future events are reflected in stakeholder Web sites and the public debate across news outlets, participatory platforms, social media, etc. Extracting emerging stories from these channels provides a real-time window into consumer and stakeholder perceptions. This enables content producers and communication professionals to react quickly and incorporate the latest trends into their decision-making - by forecasting future topics, for example, visualizing the relevance of these topics for specific products or services, or identifying opinion leaders who might help promote a product or amplify a message.
The talk will present results from ongoing European research initiatives, including a live demonstration of a visual analytics dashboard with a focus on sustainability and brand communication. It will discuss how Artificial Intelligence and semantic technologies can support the entire digital content value chain, from the data-driven production to the distribution, promotion and analysis of content assets.
Short Bio
Prof Dr Arno Scharl is the Managing Partner of Storypact and webLyzard technology, and a Professor of Information Systems at the New Media Technology Research Center of Modul University Vienna. Previously, he held professorships at the University of Western Australia and Graz University of Technology, and was a Visiting Fellow at Curtin University and the University of California at Berkeley. After completing his doctoral research and habilitation at the Vienna University of Economics and Business, Prof. Scharl has authored 190 refereed publications and edited two books in Springer‘s Advanced Information and Knowledge Processing Series. In various roles including Coordinator, R&D Lead and System Integration Lead, he helped acquire and manage more than 30 European and Austrian research projects. His current research interests focus on Web intelligence and visual analytics, opinion mining, AI-based communication success metrics, information visualization and the integration of semantic and geospatial technology.

Workshop on ChatGPT in Learning, Teaching and the Workplace

Click here to watch the recording of the seminar.
Please join Dr David Glance and guests for a workshop on ChatGPT and its impact on Learning, Teaching and the Workplace.
Please reserve a ticket for the event.
Please see this PDF for details of the event including the outcomes.
Overview
The workshop will seek to provide a overview of how ChatGPT and other AI tools can be used to enhance teaching and learning in higher education and how it will be used in the workplace.
This will be a workshop with 2 * 45 min sessions covering:

1. How does ChatGPT work? What are its strengths and limitations?
2. Teaching focus: how can ChatGPT be used to assist teaching:
   • Production of content: courses, lectures, tutorials, assessment, FAQs
   • Production of outcomes, rubrics
   • Marking of assessment and giving feedback
   • Tutoring
3. Student focus: how can students legitimately use ChatGPT for their learning outcomes:
   • Using it for learning
   • Editing work
   • Role in accessibility for disabled students
   • Learning how to incorporate these tools in future workplace
   • Understanding its limitations
   • Understanding the ethical and legal aspects of its use
4. Workplace focus: how will/is AI and ChatGPT being used in professional work? How can we prepare students for a workplace that becomes reliant on these types of assistive technologies?
5. Assessment: In the short term, how do we adapt assessment so that the use of tools such as ChatGPT have minimal impact or are integrated into the process.
6. Ethics and legal aspects of ChatGPT.