Skip to collection list
Skip to video grid
Deep Learning and the Transformation of Science Exploration
Gadi Singer, Vice President, AI Products Group, Intel
Artificial Intelligence and Big Data are contributing to elevate the exploration of scientific questions to the next level. Across all scientific fields, the ability to collect data has exploded. Artificial Intelligence – and more specifically Deep Learning – are being unleashed at this torrent of data and are delivering deeper analysis and insights, resulting in a transformation in the making and reach of science as we know it.
The session will discuss the unique characteristics of DL and how they constitute a ‘game changer’ for science exploration. In particular, it will show the high potential of pattern classification, clustering with a varying degree of supervision, feature learning, and anomaly detection.
The presentation will explore some successful examples where scientific discovery has been enhanced by DL. All those examples are based on Intel AI solutions/Hardware.
• In Meteorology, DL usage enabled early identification of forming storm conditions.
• In Particle Physics, DL-enhanced simulation of subatomic particles is deepening our understanding of particle detection and helping correct inefficiencies and inaccuracies.
• In Astronomy, the technologies help address some of cosmology's hardest challenges.
• In medicine, they are being used for genomic sequence analysis (variant detection, mRNA differential splicing and expression prediction), to identify abnormal tissue and potential tumor sites in scans and tomography, and to map the mind in real time.
In its last section, the session will address directly the AI practitioners in the audience. It will provide useful guidance on how to best approach scientific and other similarly complex domains, applying the strength and capabilities of DL in concert with more traditional field techniques to get transformational results.