Jess Adams and Adriana Toutoudaki were the lucky winners to our Re:Work competition and 
they got to attend the Re:Work Deep Learning Healthcare Summit. Check out their event 
recap below. 

On the 21st-22nd September, at a swanky venue in the centre of London’s financial district, AI and deep learning experts gathered to network and share their research and developments in the field. My colleague, Adriana, and I were lucky enough to win tickets from One HealthTech to attend this event.

Working in bioinformatics and keeping up with tech developments in healthcare means that ‘deep learning’ and ‘artificial intelligence’ are buzzwords we have been seeing cropping up more and more frequently. So naturally, we were very keen to jump on the opportunity to attend the Re-Work Deep Learning in Healthcare Summit and see what all the fuss was about from the people doing the research.

The re-work team did a really great job getting some fascinating speakers and I have to applaud the visible effort to get some really successful female speakers too. Here's a brief run-down of our 2-day crash course on deep learning in health care.

Unfortunately, I couldn’t attend the morning of day 1, but Adriana did. This is what she had to say about it:

“What a fantastic opportunity it was to attend this Re-Work event. The talk I was most looking forward to was the first one of the Summit. Sarah Culkin, Strategic Data Lead of NHS England gave an overview of where the NHS stands at the moment in terms of including AI techniques in routine clinical practice and what there is to come. It was interesting to see NHS England really thinking about this, developing guidance for clinicians and scientists to ensure successful incorporation of such technologies in a data-driven NHS. As Sarah Culkin mentioned, "we are at the foothills of getting out house in order".

Additional talks that I think were noteworthy included ones from Mark Gooding and Spiros Denaxas. Mark Gooding, Chief Scientist from Mirada Medical spoke about the importance of packaging when developing AI tools for clinicians/clinical scientists. How it is important to create tools that make routine practice easier for them rather than having to get lost in convoluted menus and outdated graphical interfaces. Lastly, Spiros Denaxas, Associate Professor in Biomedical Informatics  at UCL, spoke about aggregating and manipulating Electronic Health Record data for research and how this "free data" can lead to some important discoveries.”

In the afternoon my favourite talk was from Nikolas Pontikos about Eye2Gene- a tool to aid in the genetic diagnosis of progressive sight-loss caused by an inherited retinal dystrophy. It was good to hear about deep learning in genetic diagnostics as this is our field, but working in diagnostics means we don’t get to see early-stage concepts in our day to day jobs, as all the tools we use need to be rigorously tested and validated before we can implement them into service.

From the first day, I also want to acknowledge what a fab job Maxine Mackintosh, PhD student and co-founder of One HealthTech, did of comparing all day -- keeping it lively and entertaining throughout.

On day 2 in the morning I really enjoyed the talk given by Maithra Raghu from Google Brain, who spoke about using machine learning to predict patient cases where a second opinion could have a significant effect on diagnosis and treatment of the patient, ultimately saving time and, potentially, lives. From this talk, and many of the talks I attended, there was an underlying theme of using deep learning tools as aids rather than relying on AI entirely- which I think is both sensible and – I'm sure – reassuring for healthcare professionals!

Steven Finkbeiner from the Gladstone Institute also gave an interesting talk in the morning about using machine learning to identify early stages of neurodegenerative disorders and designing better clinical trials. This was particularly appropriate as it was actually World Alzheimer’s Day, so it was great to hear about research in this really worthwhile area.

After a delicious lunch, the focus was on ethics and data handling in the panel sessions, and it was really encouraging to see that this is something being considered in the early stages of AI development. Among the discussion were comments on the Asilomar AI principles and the Topol review for the NHS, which were interesting and particularly relevant to us. It was also nice to have a representative from the PHG foundation on the panel, as some of their policy work is particularly tied to the NHS.

We are really grateful to One HealthTech for the opportunity to attend. The day was informative and well organised and we really enjoyed ourselves- thank you!