ERC consolidator grant awarded for optoacoustic neural network project
Professor Birgit Stiller and her team are investigating alternative computing architectures for AI and machine learning based on light and hypersonic sound interactions in glass fibres and photonic chips (Image credit: Sonja Smalian/PhoenixD, Leibniz University Hannover)
Birgit Stiller, a professor at Leibniz University Hannover’s PhoenixD Cluster of Excellence and the Max Planck Institute for the Science of Light in Erlangen, has been awarded an ERC consolidator grant. Over the next five years, Stiller will receive up to €3 million in funding to advance her research project, titled Computing with Light and Sound. The funding line from the European Research Council (ERC) is intended for researchers with seven to twelve years of experience since completing their doctoral degree whose independent research group is currently in the consolidation phase.
“My project focuses on alternative computing architectures for neuromorphic computing, which mimics the way the human brain functions. I want to investigate new approaches to improving machine learning and artificial intelligence,” says Stiller, who was appointed to a W3 professorship at Leibniz University Hannover this year and is a member of the executive board of the PhoenixD Cluster of Excellence. Since 2019 she has also led an independent research group at the Max Planck Institute for the Science of Light in Erlangen. “We will use light-sound interactions in glass fibres and photonic chips to produce versatile neural network structures – so-called optoacoustic neural networks.”
Stiller wants to use her research to find ways to reduce the heavy energy consumption of today’s data processing centres. To achieve this goal, she is using photonics, the science and technology of light. This requires her to work with individual particles of light, known as photons. “Photonics is a promising solution because it enables parallel computing and therefore high bandwidth,” says Stiller. Numerous research groups, including the various teams of the PhoenixD Cluster of Excellence in Hannover and Braunschweig, are working on the optical implementation of neural networks.
“Acoustics is a real novelty in this area and will add flexible, reconfigurable building blocks that will ultimately give the optical neural networks new functions,” explains Stiller. Among other things, the desired functions could include self-learning and recurrent neural networks. Stiller has decided on optical fibres or photonic chips for her research because both platforms are highly developed. “Glass fibres are everywhere – today they are in almost every street. We will also push forward into the quantum domain and implement several ideas for quantum-based neural networks,” says Stiller of her research plans.
“The generous funding provided through the ERC Consolidator Grant makes it possible for me and my team to realise this new research approach,” she says. “I am extremely pleased about this chance.” Stiller now has the opportunity to carry out classic fibre-optic and quantum-optic experiments with new high-end equipment and to expand her research team in Hannover and Erlangen.
