6 May 2019, Ballroom, Level 1, National Convention Centre, Canberra
|9:30||9:40||Opening and Welcome (Tomasz Bednarz, Jenny Vuong, Huyen Nguyen)|
|Session chair: Conan Bourke (UNSW)|
|9:40||10:15||Keynote: Narrative Visualisation: Telling the Story Behind the Data by Steven Wark|
Situation awareness is a key requirement for military operators, analysts, and decision makers, so that they can understand what has happened, what is likely to happen, and what needs to be done to achieve desired outcomes. Visualisation is an important tool that allows them to become immersed in the relevant data and to explore and manipulate it so that they can understand the context, relationships, what is known, what is not known, what is important and what is not important to a particular situation. When automation is introduced to avoid information overload by handling the ‘big data’ problems of volume, velocity, and variety, this pathway to understanding is largely lost. New visualisation capabilities and approaches are needed to help the user to make sense of the complexity and establish the context needed to achieve situation awareness in these circumstances. Narrative is an approach commonly used by humans to immerse an audience in the context and relationships contained within a complex situation so that they can make sense of it and extract key insights. Narrative visualisation offers potentially powerful approaches to help a user exploit automation to make sense of the complexity associated with big data. In this talk I will present some of the big data challenges faced by Defence, and how we are exploring narrative visualisation techniques to tell the story behind the data for command and control and analysis of force options.
|10:15||10:35||Visualising Data with Python by Martin Schweitzer (Australian Bureau of Meteorology)|
Python has become one of the leading languages for data science and includes a wide array of tools for visualisation. In this talk I will introduce some of these tools with a view to helping users select the best tools for the results they would like to achieve. I am also hoping to convince you that Python makes it really easy to generate sparkling results This presentation does not assume a strong knowledge of Python.
|10:35||10:55||The Art of Heritage Visualisation by Kit Devine (ANU)|
How do you create a believable past with a fragmentary data record? If you only put in what is known for certainty then the visualisation is incomplete, yet the more you try to fill in the missing pieces the more speculative the visualisation will become. In this talk I will discuss the VR artwork Artistic License: Sydney Cove circa 1800 and explore the way that it engages viewers, directly and specifically, with the conundrum of heritage visualisation.
|Session chair: Justin Baker (CSIRO)|
|11:10||11:30||Data61 Immersive Environments by Matt Adcock (CSIRO Data61)|
The combination of computer graphics and computer vision can enable real-time augmentation for front-line workers and consumers. This can often require a detailed tracking of the user’s context and tasks – things that are not necessarily supported by generic Augmented Reality display technologies. This talk will describe the application of novel software architectures, interactive techniques, and algorithms that better understand the objects and events in a person’s physical environment, thereby enabling the delivery of interactive data right where and when it is needed most.
|11:30||11:50||The role of Visual and Decision Analytics Lab in Visualising Mental Health Care Ecosystem by Nasser Bagheri (ANU), Jina Chung, Luis Salvador-Carulla|
The Visual and Decision Analysis (VIDEA) lab is a new service and research organisation at the Centre for Mental Health Research at ANU. VIDEA is Australia’s first lab combining expertise in two recent and related areas of knowledge: visual analytics and decision analytics. Our vision is to make VIDEA Lab to be at the centre of a national and international visual and decision analytics hub coordinated by CMHR. This hub will bring together decision-scientists and visual analytics experts from many disciplines across Australia and worldwide, to research specific areas of policy and society. The main activities of the VIDEA lab are:
• Visual analytics: We are using new and advanced technologies and methods (e.g. machine learning, artificial intelligence, network analysis, etc) with data science capacity to visualise pattern of mental healthcare across communities to provide better communication with policy makers and stakeholders in public health sector. Our existing capacity in machine learning enables us to apply this unique expertise to analyse and visualise complex healthcare data sets.
• Decision analytics: Decision analytics applies a unique participatory approach to developing the models, providing better transparency of models and their assumptions, and enabling us to generate evidence-informed knowledge in a way that captures the complex and dynamic nature of health and social problems. Decision analytics utilises technologies and methods (visual analytics approaches) to develop adaptable decision support tools that forecast the impact of alternative decision options before they are implemented in the real world.
• Geospatial analysis and visualisation: Another arm of VIDEA Lab is capacity of geospatial analysis and spatial visualisation of healthcare patterns using GIS (Geographical Information Systems) technology. VIDEA Lab has extensive expertise in medical geography and spatial epidemiology to support spatial analysis at Centre for Mental Health Research.
|11:50||12:10||Getting Shirty with Science by Madison Hartill-Law (Shirty Science)|
In a battle against nerdy science shirts, Shirty Science was born – bringing artists and scientists together to create conversation sparking shirts about real research. Partnered through speed networking and working to a one month deadline, this project forces scientists to collaborate in a unique way, and artists to explore new worlds. No periodic table puns allowed!
|Session chair: Huyen Nguyen (UNSW)|
|13:10||13:45||Keynote: Beyond being ordinary: In the mix of science and art by Eleanor Gates-Stuart (Charles Sturt University)|
Ideally, if asked how I would like my pictures to look, I would probably say that I would like a different view every time. To be able to break down the pieces of that image and recreated them as something else – something new. A mix of materials that could be reconfigured again and again, possibility with a few other things thrown in the mix, for complete inconsistency. Easy, we might say with today’s technology, but the picture (the artwork) needs to be created first and however contradictory this may seem, given that the aim is to deconstruct the work, the picture requires being formed of many parts, in my case layers, that complete the overall visualisation blend ready for its multipurpose. This creative process is indeed lengthy as this visualisation blend often involves a second context to that of being a work of art, as a means of communicating and experiencing science in a non-traditional way. These artworks often involve collaborating with other artists and scientists in bringing the work together, a mindful mixture of interrelated components and means to think beyond the ordinary in new methods of making art in communicating science.
|13:45||14:05||Tangible Diagrams as User Experience by Henry Gardner (ANU)|
Group discussions in workshops are documented in familiar ways using notes and diagrams. Such diagrams can be formal or informal and can make use of tangible media such as the sticky note. Sometimes these diagrams can be very structured and their composition can be the principal focus of ideation and discussion of the workshops. This is particularly the case in two workshop techniques that are presently used to capture software requirements: User Story Mapping (USM) and Big Picture Event Storming (BPES). In this talk I will describe how two design theories of visual notation, the “”Physics of Notations”” and the “”Patterns of User Experience”” can be applied to critique the sticky-note diagrams developed by USM and BPES and the cooperative processes involved in building, modifying, comprehending, and discussing these diagrams.
|14:05||14:25||Climate Visualisation for Impact and Affect by Geoff Hinchcliffe (ANU)|
This talk showcases two recently completed data visualisation projects focused on creating widely accessible and affective representations of complex climate data. The web-based visualisations feature distinctive radial graphic forms and employ narrative sequencing, animation, interaction and localisation to give users a personal stake in the climate data being represented. The works also feature novel physical outputs, such as a lasercut bamboo “climate coaster”. These physical forms see visualisation moving off the digital screen and into the hands, kitchens and living rooms of our audience; another strategy to amplify engagement and affect. The talk will include demos of the projects and a brief overview of their technical composition.
|14:25||14:45||Visualising High-Dimensional Physics Data by Ursula Laa (Monash University)|
In physics we often encounter high-dimensional data, in the form of multivariate measurements or of models with multiple free parameters. The information encoded is increasingly explored using machine learning, but is not typically explored visually. I will describe how such visual exploration is possible using the grand tour, which visualises high-dimensional distributions as animations of smoothly interpolated projections, allowing the viewer to extrapolate the shape of the parameter space in high dimensions.
|Session chair: Florence Wang (CSIRO)|
|15:00||15:20||Faster Deep Learning Training with Automatic Mixed Precision by Maggie Zhang (NVIDIA)|
Automatic Mixed Precision is an easy to use method for improving training performance up to 3x by utilizing NVIDIA Tensor Cores. We will review the theory behind mixed precision using Tensor Cores, show how to use it in model training scripts, and highlight some real world performance improvements observed by customers. Compared to single precision, mixed precision offers many benefits: 2x better use of the available DRAM bandwidth, smaller memory footprints which allow larger batch sizes or network architectures with more parameters to fit in GPU memory, and allow usage of Volta’s Tensor Cores to boost raw math throughput by up to 8x.
|15:20||15:40||[Un]natural Histories: The Ramblings of a 21st Century Naturalist and Artist by Dan Power|
Art and the Natural Sciences have been intrinsically linked throughout history, they have co-evolved for thousands of years. Now, at the start of the Anthropocene, the role of Art in Science and Science in Art is evolving to stay relevant and make sense in the light of the 21st century.
|15:40||16:00||General Adversarial Networks in Contemporary Art by Baden Pailthorpe (ANU)|
This paper will give an overview of current uses of General Adversarial Networks (GANs) by contemporary artists internationally. It will also preview my own experimental application of a Deep Convolutional General Adversarial Network for an upcoming exhibition The Invisible Hand, 4A Centre for Contemporary Asian Art, Sydney (June, 2019).
|16:30||18:00||Computer Animation Festival by Tomasz Bednarz (UNSW) in Torrens Room, Level 1|