We're working to enhance the delivery of services throughout the community, using machine learning and constraints-based optimisation.
Real world problems often require linking population-sized datasets. The aim of this project was to automate and improve linkage quality, fairness and speed to ensure optimal performance of linkages in a real-world setting.
Developing an API to enable the fast and free discovery of historical aerial photography from Geoscience Australia's archives.
Data Standardisation: Standardata
The goal of the project was to investigate the effect of the type of logic on logical rule authoring and to understand how to manage complexity at scale in logical rule authoring. In addition, the project aims to identify the benefit in terms of effort and ease-of-use of using DDL instead of classical logic is investigated, as well as to identify additional logical features to ensure alignment between legal texts and their corresponding logical representation.
The Longitudinal Linked Employer-Employee Data System (L-LEEDS) Project explores approaches for building software components to support a privacy-preserving, federated data sharing model, as a foundation for a future whole-of-government distributed privacy-preserving, data analytics environment.
Federated Magda aims to form a comprehensive solution for the management of data within government agencies. Aligning with wider federal government initiatives around data use, Federated Magda enables more of the data held by government to be released and shared.
CSIRO’s Data61, in collaboration with government and other CSIRO business units, developed a proof-of-concept applying satellite observation and machine learning for more effective crop management in Australia.
An artificial intelligence engine for traffic congestion management developed for Transport for NSW.
The project aim was to make a synthetic version of the Australian Taxation Office (ATO) Information Longitudinal Files (ALife), that provides known levels of privacy and targeted utility. A privacy-preserved synthetic dataset gives an individual “plausible deniability” that their data contributed to the synthetic data. A potential use case is in assisting researchers’ decision-making process on whether to formally apply to access ALife and incur the costs of doing so.
The efficacy of aggregation in ensuring privacy and confidentiality can at times be lacking, with a residual risk of learning information about contributors to aggregated data. Rules are often used to determine suppression of the publication of risky data cells, however it may still be possible to reconstruct suppressed data. Secondary suppression can be used to control this risk. This project has looked at incorporating risk models into a prototype system for use by APRA to assess risk in suppressed cells in an APRA publication.
The Protari (Protection Against Re-Identification) Data Explorer enables a wide range of users to gain insights from sensitive datasets while maintaining the privacy of individuals in the data. The project was a collaborative effort between analysts, designers and engineers at Data61, the Department of Prime Minister and Cabinet (PM&C), the Department of Social Services (DSS) and the Australian Bureau of Statistics (ABS).
The CSIRO Data61 “Knowledge Extraction in Context” system is a proof-of-concept prototype that was developed in collaboration with the Department of Agriculture, Water and Environment. The prototype demonstrates the feasibility of converting content from conservation statutory documents into knowledge that can be searched, browsed, visualised and used to provide insights about threats, habitats and interactions concerning Australian threatened species listed under the Environmental Protection and Biodiversity Conservation (EPBC) Act 1999. The resulting prototype helps to unlock the knowledge stored only in PDFs so that it can potentially feed into downstream conservation and environment-related data analytics.
YDYR enables the transformation of spatial data between different spatial geographies, allowing users to re-project data onto a different geographic region. It improves on existing methods by modelling the relationship between the data of interest and ancillary data available on a common set of regions. As well as improving the accuracy of the transformation, it delivers estimates of uncertainty that can be propagated through to downstream analysis and inform the user of the quality of the transformation
We’re working with Rio Tinto in Western Australia’s Pilbara region to maximise the throughput of their rail system by using simulations to balance train scheduling with maintenance and other system constraints.
We're re-imagining regulation as an open platform based on digital logic to help support a growing ecosystem of digital regulation tools and services.
We're helping government agencies to identify and segment their social media followers, allowing them to develop targeted messaging for bigger impact and reach.
Our social media monitoring tool Vizie is transforming the way government agencies listen to, understand and respond to customer feedback from social media.
Enter a valid email address, for example jane.doe@csiro.au.
First name must be filled in
Surname must be filled in
We'll need to know what you want to contact us about so we can give you an answer.
For security reasons attachments are not accepted.
Sorry, an error has occured, please try again.
We have received your enquiry and will reply soon.
