AI & Machine Learning

Lead researcher - Cisco Chair of AI and IoT Professor Wei Xiang

About - Water is a critical and precious natural resource in Australia. However, water quality is declining in both natural ecosystems, as well as those managed by commercial industry.

We are working with leading Australian water quality monitoring service provider, Eco Detection, on using Artificial Intelligence of Things to create a novel embedded AI engine to advanced state-of-the-art water quality management systems with the help of machine learning and data techniques.

Our project will help Eco Detection develop a world-first 24/7 water quality monitoring solution and sensing commercial product.

Partner - Eco Detection

Australian Turntable Company (ATC) is a regional manufacturer based in Melbourne. It is seeing increased demand for its turntable products within current, new industries and export markets. The primary objective of this project is to leverage innovations in the Industrial Internet of Things (IIoT), Industry 4.0, Artificial Intelligence (AI), and machine learning, within the Turntables' manufacturing environment to capture process improvements and collect data at the source. In this project led by Cisco Chair of AI and IoT Professor Wei Xiang, an ERP system, along with an Internet of Things solution, is proposed to facilitate interconnectedness within the organization, which will capture information in real-time, as well as guide the organization through improvement processes, with the aim of unlocking the full potential of automation, robotics, and other advanced manufacturing processes for the organization. Through the implementation of the digital transformation, continuous improvements are achieved to maintain competitiveness, such as improving inventory management and reducing maintenance costs.

Water is one of the most critical and precious natural resources for Australia and its people. It is an issue of national significance to protect Australia’s precious water supply. The effect of declining water quality water quality is evident in a number of areas, e.g., human health, aquatic ecosystem (the most famous example being the Great Barrier Reef that is dying rapidly because of the poor marine water health), agriculture and aquaculture, etc. Improving and maintaining water quality is, without a doubt, one of Australia’s key priorities, given that the combined value of the agriculture and aquaculture industries to the Australian economy is $15.6 billion in 2021/2022, according to the Australian Bureau of Statistics.

In this project led by Cisco Chair of AI and IoT Professor Wei Xiang, we are working with Eco Detection on using novel Artificial Intelligence of Things (AIoT) technologies, which is a leading Australian WQM service provider and offers a promising world-first WQM solution and commercial product to realise 24/7 in-situ WQ sensing. The project is seeing the creation of a novel embedded AI engine to advance state-of-the-art water quality management systems by using novel machine learning (ML) and data fusion techniques.

This world-first Proof-of-Concept (PoC) trial aimed to determine the feasibility of using Wi-Fi to detect mobile devices at the Melbourne Central Station and to provide real-time insights on passenger behaviour. High smartphone penetration rates in Australia mean that most passengers passing through Melbourne Central Station possess mobile devices that periodically transmit Wi-Fi messages. The trial demonstrated that train arrival/departure events, passenger counts on the platforms and trains, and distribution of passengers on the platform could be estimated from the collected Wi-Fi data.

The PoC trial is a collaboration among Cisco, Victorian Department of Transport, Cohda Wireless, La Trobe University, and University of Melbourne. The La Trobe team led by Cisco Chair of AI and IoT Professor Wei Xiang developed a powerful Transformer-based machine learning model that is capable of providing real-time passenger number forecasting at the Melbourne Central Station. The experimental results based upon collected data from the MCS demonstrated that the developed Transformer-based prediction model can be used to mitigate congestion and improve passenger experience.

Lead researchers – Dr Lianhua Chi, Haishuai Wang, Zongyuan Ge and Wei Xiang

About – Our team have developed an effective Autism Spectrum Disorder (ASD) prediction model. This uses a person’s characteristics and behavioural data, along with facial images to provide an accurate prediction for diagnosis. ASD is a developmental disorder that affects communication and behaviour. It is critically important to diagnose ASD during the early stages of child development to empower parents and enable them to access available support services.

We have also created a web-based preliminary ASD screening tool. This reduces waiting times to only a few seconds and can provide an accurate prediction using uploaded facial images that helps medical researchers and ASD specialists to deliver an ASD diagnosis for an individual. Our project helped parents quickly and accurately identify children at potential risk for autism. We also provided new insights into ASD research enhancing how doctors evaluate patients with other behavioural disorders.

Partners – Harvard Medical School, Fairfield University, Monash University