This partnership will drive forward the development of cutting-edge machine learning and AI technology in sleep health. SleepCogni is thrilled to welcome Professor George Panoutsos, an esteemed leader in Computational Intelligence, to the team as their AI Science Advisor.
Professor Panoutsos, who was recently announced as the Head of the School of Electrical and Electronic Engineering at Sheffield, brings a wealth of knowledge and expertise to SleepCogni. Over the past eight years, SleepCogni has cultivated a strong relationship with Professor Panoutsos, working closely with him and his research group. This partnership includes significant contributions from Joseph Hawkins, SleepCogni’s Head of Operations, who enhanced SleepCogni’s machine learning capabilities as an undergraduate student.
Professor Panoutsos expressed his enthusiasm for the collaboration: "I have long admired the innovative work of SleepCogni and am thrilled to deepen our collaboration. The physiological, behavioural, and environmental data collected by the SleepCogni device is significant and provides a great opportunity to advance sleep health technology. Not only can SleepCogni generate immediate impact in sleep health, they can also use the same framework to set the future research agenda in digital health for sleep, such as personalised advisory systems, and autonomous sleep health optimisation. With SleepCogni we are exploring state of the art machine learning models to apply and evaluate their potential towards full feedback systems that would improve sleep health. Our initial analyses have uncovered potential correlations in the data, that could be used to forecast sleep efficiency, a critical indicator of sleep health. The alignment of SleepCogni's feedback algorithm with a class of models known as discrete event systems is excellent, as it opens opportunities to use a plethora of system analysis tools to demonstrate system robustness.
“We are now developing a new simulation model – a Digital Twin of the actual device - that would interact iteratively with the feedback algorithm in real time, to test millions of algorithm variations, ensuring safe and effective biofeedback loops. This will allow us to personalise therapy sessions based on the severity of insomnia. The opportunity to work on such a data-rich, ground-breaking technology is incredibly exciting, and I’m looking forward to further technical developments within this collaboration, as well as new research."
Richard Mills, CEO of SleepCogni, added: “We are thrilled to welcome Professor Panoutsos to our team. His leadership in AI and machine learning will help us develop a clear and strategic path forward, leveraging the best knowledge and expertise available globally. The University of Sheffield's leading computational resources, combined with the AI expertise of Professor Panoutsos, his colleagues, and students, make this partnership a significant asset for our business, which we expect to grow. Originally being based at the University of Sheffield's Bioincubator, and now at the Kroto Innovation Centre has been an added advantage, supporting our long-standing collaboration.
“Professor Panoutsos is the ultimate professional, with a keen understanding of our business's speed and ambition. It has been a pleasure working with him over the years, and we look forward to growing our relationship with him and the School of Electrical and Electronic Engineering at The University of Sheffield.
“We are also excited about the developments in large language models (LLMs), such as the GPT-4o’s voice capabilities, which has the potential to be a game changer for SleepCogni and its users.
“ Our aim is to develop a personalised sleep coaching assistant using underpinning Machine Learning research for sleep optimisation and decision making, integrated with LLMs based on CBT-I standards, our own proprietary device data and sleep medical advisory team input. This will be delivered through our bedside device, it will offer real-time, 24-hour support and advice in 50 languages, being able to perceive emotional cues from users and personalise empathetic audio responses.”