With the development of machine learning techniques, the attention of research has been moved from single-modal learning to multi-modal learning, as real-world data exist in the form of different modalities. However, multi-modal models often carry more information than single-modal models and they are usually applied in sensitive scenarios, such as medical report generation or disease identification. Compared with the existing membership inference against machine learning classifiers, we focus on the problem that the input and output of the multi-modal models are in different modalities, such as image captioning. This work studies the privacy leakage of multi-modal models through the lens of membership inference attack, a process of determining whether a data record involves in the model training process or not.

Most of the successful deep neural network architectures are structured, often consisting of elements like convolutional neural networks and gated recurrent neural networks. Recently, graph neural networks (GNNs) have been successfully applied to graph-structured data such as point cloud and molecular data. These networks often only consider pairwise dependencies, as they operate on a graph structure. We generalize the GNN into a factor graph neural network (FGNN) providing a simple way to incorporate dependencies among multiple variables. We show that FGNN is able to represent Max-Product belief propagation, an approximate inference method on probabilistic graphical models, providing a theoretical understanding on the capabilities of FGNN and related GNNs. Experiments on synthetic and real datasets demonstrate the potential of the proposed architecture.

With the development of machine learning techniques, the attention of research has been moved from single-modal learning to multi-modal learning, as real-world data exist in the form of different modalities. However, multi-modal models often carry more information than single-modal models and they are usually applied in sensitive scenarios, such as medical report generation or disease identification. Compared with the existing membership inference against machine learning classifiers, we focus on the problem that the input and output of the multi-modal models are in different modalities, such as image captioning. This work studies the privacy leakage of multi-modal models through the lens of membership inference attack, a process of determining whether a data record involves in the model training process or not.

A common task in forensic biology is to interpret and evaluate short tandem repeat DNA profiles. The first step in these interpretations is to assign a number of contributors to the profiles, a task that is most often performed manually by a scientist using their knowledge of DNA profile behaviour. Studies using constructed DNA profiles have shown that as DNA profiles become more complex, and the number of DNA-donating individuals increases, the ability for scientists to assign the target number. There have been a number of machine learning algorithms developed that seek to assign the number of contributors to a DNA profile, however due to practical limitations in being able to generate DNA profiles in a laboratory, the algorithms have been based on summaries of the available information. In this work we develop an analysis pipeline that simulates the electrophoretic signal of an STR profile, allowing virtually unlimited, pre-labelled training material to be generated. We show that by simulating 100 000 profiles and training a number of contributors estimation tool using a deep neural network architecture (in an algorithm named deepNoC) that a high level of performance is achieved (89% for 1 to 10 contributors). The trained network can then have fine-tuning training performed with only a few hundred profiles in order to achieve the same accuracy within a specific laboratory. We also build into deepNoC secondary outputs that provide a level of explainability to a user of algorithm, and show how they can be displayed in an intuitive manner.

This paper discusses the effectiveness of leveraging Chatbot: Generative Pre-trained Transformer (ChatGPT) versions 3.5 and 4 for analyzing research papers for effective writing of scientific literature surveys. The study selected the \textit{Application of Artificial Intelligence in Breast Cancer Treatment} as the research topic. Research papers related to this topic were collected from three major publication databases Google Scholar, Pubmed, and Scopus. ChatGPT models were used to identify the category, scope, and relevant information from the research papers for automatic identification of relevant papers related to Breast Cancer Treatment (BCT), organization of papers according to scope, and identification of key information for survey paper writing. Evaluations performed using ground truth data annotated using subject experts reveal, that GPT-4 achieves 77.3\% accuracy in identifying the research paper categories and 50\% of the papers were correctly identified by GPT-4 for their scopes. Further, the results demonstrate that GPT-4 can generate reasons for its decisions with an average of 27\% new words, and 67\% of the reasons given by the model were completely agreeable to the subject experts.

For this experiment, the researchers relied on 61 hours of video from a helmet camera worn by a child who lives near Adelaide, Australia. That child, Sam, wore the camera off and on for one and a half years, from the time he was six months old until a little after his second birthday. The camera captured the things Sam looked at and paid attention to during about 1% of his waking hours. It recorded Sam's two cats, his parents, his crib and toys, his house, his meals, and much more. "This data set was totally unique," Lake says.

Automated multi-label chest X-rays (CXR) image classification has achieved substantial progress in clinical diagnosis via utilizing sophisticated deep learning approaches. However, most deep models have high computational demands, which makes them less feasible for compact devices with low computational requirements. To overcome this problem, we propose a knowledge distillation (KD) strategy to create the compact deep learning model for the real-time multi-label CXR image classification. We study different alternatives of CNNs and Transforms as the teacher to distill the knowledge to a smaller student. Then, we employed explainable artificial intelligence (XAI) to provide the visual explanation for the model decision improved by the KD. Our results on three benchmark CXR datasets show that our KD strategy provides the improved performance on the compact student model, thus being the feasible choice for many limited hardware platforms. For instance, when using DenseNet161 as the teacher network, EEEA-Net-C2 achieved an AUC of 83.7%, 87.1%, and 88.7% on the ChestX-ray14, CheXpert, and PadChest datasets, respectively, with fewer parameters of 4.7 million and computational cost of 0.3 billion FLOPS.

Become an expert and make a difference to society. The University of South Australia (UniSA) is Australia's University of Enterprise. We are South Australia's largest university and one of the very best young universities in the world. At UniSA, we are authentic, resilient, and influential - and we deliver results. We pride ourselves on our dynamic and agile culture, which embraces challenges and thrives on breaking new ground.

ADELAIDE, AUSTRALIA--Cosmos Magazine reports that Daryl Wesley of Flinders University and Mimal and Marrku Traditional Owners of the Wilton River area used machine learning to analyze changes in rock art styles in northern Australia's Arnhem Land. The computer was supplied with information of more than 1,000 types of objects and a mathematical model to determine how similar two images are to one another. The model was then applied to images of the rock art. "One amazing outcome is that the machine learning approach ordered the styles in the same chronology that archaeologists have ordered them in by inspecting which appear on top of which," said team member Jarrad Kowlessar of Flinders University. Styles of artwork that are closer to each other in age are also closer to each other in appearance, he explained.

Adelaide (Australia): Researchers at the University of South Australia have developed a cost-effective new technique to monitor soil moisture using a standard digital camera and machine learning technology. The United Nations predicts that by 2050 many areas of the planet may not have enough fresh water to meet the demands of agriculture if we continue our current patterns of use. One solution to this global dilemma is the development of more efficient irrigation, central to which is precision monitoring of soil moisture, allowing sensors to guide'smart' irrigation systems to ensure water is applied at the optimum time and rate. Current methods for sensing soil moisture are problematic -- buried sensors are susceptible to salts in the substrate and require specialised hardware for connections, while thermal imaging cameras are expensive and can be compromised by climatic conditions such as sunlight intensity, fog, and clouds. Researchers from The University of South Australia and Baghdad's Middle Technical University have developed a cost-effective alternative that may make precision soil monitoring simple and affordable in almost any circumstance.