Ara Institute of Technology - Research Week - Session 3

 This week, Ara research office convenes 'Research Week', an opportunity for our researchers to share their scholarship, research practices and approaches.

Having been away the last for the lunch time sessions on Monday and Tuesday, it was good to be able to attend two - both around teaching and learning (ako).

First up, Dr. Amit Sarkar shares an update on some of the AI@Ara projects. AI@Ara:building capability for both kaiako and ākonga in tertiary education. 

Amit has been working in computing for over 20 years and the field is continually moving. Defined AI as 'computer softward that mimics how humans think in order to perform tasks such as reasoning, learning, and analysing information. Ran through the arrival of Gen AI in November 2022. Then shared the implication on teaching and learning (ako). Reviewed the differences between the predictive machine learning model and Generative AI models - from labelling data, to producing 'new' content. Open AI's ChatGPT had 11 million users within 2 months.

Overviewed AI @Ara - projects from 2023 to present. From pilots, to chatbots, to agents. Summarised the number of ākonga (students), kaiako (teachers), computing tutors, and educational developers. Learned from the projects as they progressed. Need to workout how computer science knowledge applied to aligning to project objectives. Shared the steps to be taken to develop something that supports, rather than replaces learning. Context engineering includes the original quesion, students' current working, earliest attemps, identified misconception, hints already provided. A directed linear RAG workflow is required. Explained concept fo agentic state graph execution to tighten agents. Shared the many technical learnings from the projects. Projects used LangChain and LangGraph. Within the Ara system, Microsoft Copilot studion and Cogniti.

Next, Graeme Harris with 'research with a motor-racing flavour: simulation-driven optimization of a methanol-fueld performance SI engine as part of teaching engineering'. Provided rational for the current project. Currently works with ākonga on a Year 3 engineering course on systems analysis. Has access to a simulator (EngMod4) for ākonga to run data through to optimise how an engine will run. Detailed some of the technical details,processes and challenges of a engine-inlet with methanol injection. The problem solving requires ākonga to apply their learning. Summarised how the simulation runs. Closed with future goals.