Latest events

AIML Research Seminar: Sensing and Lasing with Multimode Fibre Optics

Date: Tue, 6 May 2025, 10:30 am - 11:30 am
Location: AIML

When light propagates in multimode optical fibres, it is confined in many orthogonal modes whose transverse profiles are propagation-invariant. This high dimensional modal space allows great flexibility in processing the output information, as well as controlling the input information for various purposes. In this talk, I will discuss a coherent suite of topics including (i) processing the output information from a multimode fibre in the context of sensing using machine learning, (ii) controlling the input information into multimode fibre as a mean of tailoring the output for laser applications, and (iii) the introduction of physical neural network, an emerging field in complex optics toward realisation of energy-efficient and low-latency machine learning.

AIML Special Presentation: Meaning and Intelligence in Language Models — From Philosophy to Appropriate LLM Responses

Date: Mon, 5 May 2025, 2:00 pm - 3:00 pm
Location: AIML

Language Models have been around for decades but have suddenly taken the world by storm. In a surprising third act for anyone doing NLP in the 70s, 80s, 90s, or 2000s, in much of the popular media, artificial intelligence is now synonymous with language models. In this talk, Prof Chris Manning will take a look backward at where language models came from and why they were so slow to emerge, a look inward to give some thoughts on meaning, intelligence, and what language models understand and know, and a look at some recent work on steering language models to respond well to people’s questions and commands.

AI on the Ground Seminar: Translating an AI algorithm for antenatal hydronephrosis

Mandy, Lauren, Melissa - AI on the Ground

Date: Fri, 2 May 2025, 10:30 am - 11:30 am
Location: AIML

Antenatal HN is the most common congenital anomaly, affecting up to 5% of pregnancies and identified as a dilatation of the urinary tract on prenatal imaging. While the majority of cases are physiologic and self-resolving, a subset necessitate surgical intervention. Currently, infants are monitored with serial ultrasounds and may undergo invasive and burdensome investigations, including catheterisation and radionuclide imaging, which contribute to patient distress and healthcare costs.

AIML Special Presentation: How tough is your data? Applying structural graph theory to robust fitting and clustering

Date: Thu, 10 Apr 2025, 11:00 am - 12:00 pm
Location: AIML

Structural graph theory is making great advances in understanding what characteristics of graphs act to make those graphs difficult or easy for algorithms for a range of common problem types (maximum clique, graph cuts, etc.). Modern machine learning is largely driven by the success (or otherwise) of a proposed model/algorithm on standard test datasets. This involves a plethora of different measures and sometimes one might suspect that a dataset has “easy examples” and “hard examples” and that a statistic might be affected by the proportion of hard or easy samples. So a structural graph theory based understanding of AI problems should help both understand benchmark results and also to design algorithms according to the range of complexities of data a problem setting can “throw up”.

AIML Special Presentation: Introducing the Responsible AI Research (RAIR) Centre

Date: Tue, 8 Apr 2025, 10:30 am - 10:45 am
Location: AIML

We're pleased to welcome Kate Klimeš, Program Manager for the newly established Responsible AI Research (RAIR) Centre, who will be presenting on the vision and goals of the new centre. The Centre is based at AIML in the Lot Fourteen business and technology precinct in ��. RAIR represents a powerful collaboration between AIML’s world-class researchers in artificial intelligence and machine learning, scientific experts from CSIRO, and is supported by the South Australian Government.

AIML Special Presentation: Accelerating Science with NVIDIA

Date: Thu, 3 Apr 2025, 2:45 pm - 3:45 pm
Location: AIML

NVIDIA's advanced GPU technologies and AI platforms are revolutionising scientific computing across multiple domains. This talk will explore how NVIDIA's solutions are enabling unprecedented acceleration of research and discovery in fields like generative AI, digital twins, data science, climate modelling, robotics and drug discovery. Dr Johan Barthélemy will introduce the latest AI-powered tools such as RAPIDS, Modulus, (Bio)NeMo, Omniverse, Cosmos, and Groot, pushing the boundaries of what's possible. Real-world examples will demonstrate how researchers are leveraging those technologies to tackle complex problems orders of magnitude faster and more efficiently than traditional methods.

AIML Special Presentation: Continuum Robots for Minimally Invasive Surgery: Smaller, Softer, and Smarter

Date: Thu, 27 Mar 2025, 11:30 am - 12:30 pm
Location: AIML

Surgical robots have achieved commercial success and are experiencing rapid global growth. The robotics community continues to push the boundaries of safety, ease of use, and minimally invasive techniques in surgical robots. Continuum robots, with their intrinsic flexibility and miniature sizes, are emerging as ideal candidates for the next generation of minimally invasive surgical systems. Their development, however, presents three major challenges: design and manufacturing, human-robot interaction, and intelligent perception. In this talk, Dr Liao Wu will share his efforts to overcome these challenges, focusing on advancements that make surgical continuum robots smaller, softer, and smarter.

AI on the Ground: The Great Chatbot Debate Watch Party

Date: Wed, 26 Mar 2025, 12:30 pm - 2:30 pm
Location: AIML

Chatbots based on large language models (LLMs), like ChatGPT, answer sophisticated questions, pass professional exams, analyze texts, generate everything from poems to computer programs, and more. But is there genuine understanding behind what LLMs can do? Do they really understand our world? Or, are they a triumph of mathematics and masses of data and calculations simulating true understanding?

AIML Research Seminar: The Search for LLM-Based Disinformation

Date: Tue, 25 Mar 2025, 11:00 am - 11:30 am
Location: AIML

Disinformation is one of the biggest threats to social progress and stability. Lies and post-truths radicalize decision-making and pervert knowledge. Accordingly, there has been an explosion of research seeking to detect disinformation in online content. Here, the frontier of disinformation detection research leverages a variety of ML techniques. Yet, a recent meta-analysis discovered existing techniques are only 79% accurate. More problematic is existing techniques only work against generated artifacts. Meaning, articles, tweets, and Reddit posts are in the wild before detection occurs. We need a technique to detect potential disinformation in the ML models before content gets loose. Fortunately, work is underway attempting to construct such a technique. This presentation will bring you up to speed on misinformation concepts. We will cover existing machine learning detection methods. The feature will be a progress snapshot of working at the model level. Keep in mind this is not only about fake news. No. Detecting misinformation is about preserving all human knowledge.

AIML Special Presentation: ARC Discovery Project on Hyperspectral Video Tracking

Date: Fri, 21 Mar 2025, 11:00 am - 12:00 pm
Location: AIML

Video tracking is a fundamental AI technology with more than 100 billion dollars market values. Traditional tracking models based on colour or grayscale videos have inherent limitations in detecting and tracking objects. Focused on challenging scenarios faced by conventional camera systems, this research in the past five years harnesses the capability of hyperspectral video cameras in material identification within and beyond the visible spectrum to capture and model the spectral, spatial, and temporal information for object tracking. This talk gives an overview of hyperspectral imaging technology and how it is used to develop hyperspectral tracking methods, which form the foundation of a recently funded ARC Discovery Project.

��