The current and next generation of all-sky surveys such as Gaia, SKA and Euclid will revolutionize astronomy. With the rise of big data era, the astronomy community have adopted data-driven approaches such as machine learning algorithms and visualisation techniques into their research. These techniques are being used in a broad range of astronomical problems, from solar system objects to galaxy evolution to cosmology.

There are several opportunities provided by this mysterious black box set of algorithms that can outperform traditional approaches. However, one might be skeptical toward the end results, e.g. the reliability and robustness of the output. Addressing these concerns and understanding the limits and challenges of applying ML to both the real and simulated astronomical datasets is the main goal of this session. Therefore, we would like to invite researchers to discuss the capabilities of ML algorithms and to realize their challenges and limitations. In addition, we invite scientists to rethink the science questions based on what we have learned from analysing current datasets so far and push the community further towards next possible steps.

Programme

The symposium consists of 6 blocks of 1.5 hour and includes invited talks, contributed talks, and posters (including poster flash talks). The following represent an overview of the expected talks and discussions in this session.

Invited speakers

Scientific organisers

Contact