Applied Lessons in AI with UK.AI & the Hippocratic Oath for Artificial Intelligence
London, 7th of September
Why launch a community of AI practitioners in London
The City of London, one of the world’s top financial and technology hubs, gained momentum in the race towards artificial intelligence when the likes of DeepMind, Benevolent, and — most recently FIVE AI — started to reveal impressive achievements in computer vision, deep learning and natural language processing.
London’s ingenuity and entrepreneurial spirit is invariably attributed to its diversity and supportive ecosystem for industry incumbents challenging businesses that have yet to benefit from the power of cognitive services and automation. At present, London is a “serious playground” for AI startups, generously topped up with talent, tech, and capital.
Building a community to nurture, connect and share this knowledge around applied AI was the next logical step for us, Catalina and Jamie, the two City.AI Ambassadors starting UK.AI as a local chapter.
“As a VC fund looking to invest in the most exciting Applied AI businesses, I believe that London is the richest territory for Forward Partners to operate. The city is located at the crossroads of business, science and academia so it’s naturally placed to provide both the deep thinking and commercial application needed to fuel groundbreaking AI. The world class quality of businesses being created here has already been underlined by the high profile acquisitions of London based AI businesses by US tech titans, namely Microsoft’s acquisition of Swifkey, which started out at UCL and Google’s acquisition of DeepMind.
Couple the above factors with the fact that London attracts the most talented entrepreneurs from across Europe and its clear that our ecosystem here is especially conducive to fuelling the next generation of AI startups.” —
Nic Brisbourne, Managing Partner at Forward Partners
UK.AI — Applied AI Lessons
The launch of UK.AI on 7th of September was kindly supported by Central Working and Imperial College, now heavily focused on growing London’s first hub for applied sciences, from researchers working in STEM to companies applying AI.
The first to share a practical lesson on applying AI to solve the problem of message analysis, classification and automation in customer experience was Yoram Bachrach, Head of Research at DigitalGenius.
DigitalGenius deployed the Humans+AI Product in contact centres for KLM, TravelBirds, L’Oreal and Magoosh. Using neural networks to pre-classify messages, the system was also capable of pre-filling customer support responses depending on the message received via Facebook, WhatsApp and other channels.
Whilst repetitive questions were handled by the DigitalGenius product, reducing workload, complex tickets were left to human agents, who could provide adequate levels of care and support to customers with unique cases.
View the entire presentation here or watch the video below.
—
The second set of speakers, Phil Westcott and Sean Anderson of Filament, shared their insights from applying machine learning to various use cases, from training a chatbot to recognise asthma symptoms better for Think.Act.Breathe on behalf of Boehringer Ingelheim, to helping T-Mobile’s chatbot recognise user moods to strengthen its connection with users and provide more appropriate responses depending on these emotional indicators.
You can check out their presentation here as well.
—
The third speaker, Giovanni Vacanti from Seldon, emphasised the importance of using machine learning only when it is truly necessary. His lessons was focused on explaining how machine learning algorithms work using types of training: matrix factorisation, which always needs data to train on (“the cold start problem), and the hybrid model which combines matrix factorisation with deep text learning (movie recommendations, reviews) to create suggestions for users who have just started using a service. The latter works very well for services such as Netflix that provide custom recommendations to users. Instead of waiting for the user to watch enough movies and rate them, the engine uses a hybrid model to suggest movies based on text, reviews and content written by other people outside Netflix.
To learn about applying a hybrid model for content recommendation engines, check his presentation here.
To spice things up, we replaced the traditional panel discussion with a blind date between the three speakers and two industry leaders: Alex Harvey, Head of Robotics and Autonomous Systems at Ocado, and Anton Fishman — Chair of AI and Robotics Conference taking place on the 14th of September in London. If you are a startup applying AI or working with robotics, get in touch with Anton who will kindly be able to offer free space in the exhibition area. More details about the programme here.
As all blind dates go, each “bachelor’ — in our case, industry leader, asked two questions and each speaker had to give their best answer. At the end, the two guests decided who the winner is and whether they would go on a business date with them or not.
SO MUCH FUN!
Towards the end of the evening we heard from Benedict Dellot, Associate Director at The RSA, who walked us through the risks and benefits of using AI in decision making processes that may have a significant impact on people’s lives.
The problem of ethics and bias in cognitive systems has received a lot of attention from luminaries in the UK and worldwide, and organisations including The RSA, IEEE, The Future of Humanity institute and Leverhulme Centre for Artificial Intelligence are working on various solutions of addressing this problem and raising awareness.
Benedict’s proposal was inspired by the legendary Hippocratic Oath, used in medicine to guide practitioners in using science and knowledge for the benefit of humans and society.
The launch of UK.AI was a success on all fronts: a full house, a great community, practical lessons and case studies, and a truly supportive ecosystem. We are looking forward to multiple events in the next few months, and will be announcing our first set of interviews shortly.
Share your AI challenge
City.ai would like to get your feedback on the kinds of challenges you face when applying or using AI in your company. Are you unsure which technology to use? Are you concerned about data bias? Are you looking into the pros and cons of building an in-house solution versus hiring experts to handle specific projects, such as chatbot training?
Do let us know what you think by filling this feedback form.
About the UK.AI Ambassadors
Catalina Butnaru is currently working on the first training methodology for helping people make the best use of AI, such that is augments jobs instead of replacing them. Like the agile methodology for managers, but for humans to work with AI. She is also committee member of IEEE, contributing towards a framework for advancing Wellbeing with the use of AI. In the day life she is VP Marketing at Reincubate.
Jamie’s background is in emerging technology innovation, start-up incubation and venture building, having worked with large corporations and nimble start-up businesses. Most recently, Jamie founded and led the start-up accelerator of EY and co-founded an EdTech start-up. He also advises a number of early-stage ventures.