Deep Learning and AI Assistant Summits London, Day 1 Highlights
“In 10 years, AI will be as exciting as electricity is to us today. You walk into a room right now and you don't even have to switch the lights on. AI will be like that. AI will be an obvious part of our business processes. You wouldn't want to run a business without it”
Michael Natusch, Prudential
Today saw RE•WORK hosting the Deep Learning in Finance Summit, Deep Learning in Retail & Advertising Summit and the AI Assistant Summit, bringing together some of the leading global minds applying these technologies in multiple industries.
On the finance track, the mornings sessions kicked off with Claire Calmejane from Lloyds Banking Group who is one of the only women in the European Top 40 Fintech list, welcoming everyone to the summit: ‘It’s amazing that attendees have travelled from as far as Toronto and São Paulo to come to London to learn about Deep Learning - we’re going to learn lots today!’
Soledad Galli, Lead Data Scientist at LV= began by discussing ways in which we can interpret machine learning models for insurance claims. ‘Data and machine learning is becoming the core of many businesses in both the public and private sectors. The amount of available data makes it much easier to apply machine learning to improve the decisions we make - this is especially relevant in Finance because of the vast amount of data we have.’
At RE•WORK we really value showcasing diversity in AI, so we were very glad to kick off the day with three female speakers across the tracks.
On the AI Assistant track, having already heard discussions on NLP and voice recognition implementation, Yariv Adan from Google Assistant explained 'the next tech revolution is happening now and it’s powered by AI. We’re shifting to a world where machines are adapted to humans.' Google's mission is ‘to organise the world's information and make it universally accessible and useful’, and Yariv explained how it’s incredibly important for AI and conversational UI to be combined as this has the power to scale expert capabilities that are currently accessible to only a few, and bring them to everyone breaking barriers across geography, age, literacy, language, culture, and affordability. ’Google Assistant understands context in a conversation, context is one of the main things my team is working on.’
@Yssybyl: You can’t think of a worse question to ask a machine than “When?” Context and understanding the meaning of the point of the question is critical- great demo of Google Assistant by Yariv Adan #ai #reworkRETAIL
@timanderson : How many engineers are working on Google Assistant? Answer "A lot" #reworkai
AI is also hugely transforming retail. Your first reaction may be to think of online shopping, which is of course being disrupted by AI, and we heard various presentations from the likes of Walmart Labs, who explained how they use AI:
‘We heavily use ML in product recommendation. Say you’re looking at a pair of jeans and we want to recommend you something similar so we do this based on user history, and each product can be abstracted using an embedding in word2vec"
daco @daco_io: "Model accuracy is not always the best metric, improving classification accuracy with image is hard" - Alessandro Magnani, Data Scientist @WalmartLabs at @teamrework #reworkAI conference in London #LiveTweet
We will also hear from ASOS tomorrow on this topic.
Taking a different angle on retail, Kat James from Royal Mail explained how they are leveraging AI in recommender systems. Royal Mail ‘are legally obliged to deliver approximately 50 million letters to 24k addresses 6 days a week within 1-3 working days’ which is a huge amount, so they are working to improve efficiency with AI. They ‘have implemented a recommender system to help the B2B sales team, which is probably our most traditional use but we are also working on using recommenders as information filtering algorithms, hoping to simplify making data driven decisions.’ Interestingly, Kat explained, ‘Our main challenge, which is probably relatively unique to Royal Mail is that we have too much data. There isn’t an area of the business that isn’t crying out for help from our team and combining advanced data science techniques and legacy systems is a constant challenge for us.’
After a morning full of exciting presentations, it was time for a networking coffee break, where attendees discussed their morning so far:
‘My goal is to meet people in finance and also our team are looking forward to participating in the panel later today'.
Adil Yousif, ElementAI
‘I’ve enjoyed the talks around conversation design. I’m really looking forward to more discussions on this topic since it puts the end user at the very heart of the reflection’ - Aurelien Francois, Axa
‘The most exciting thing about working in AI is that it’s advancing so quickly so you are always learning new things. The tech that I used at the start of my Phd is completely different to what I'm using now, which forces me to keep up and challenges me.’
Daniel Trigueros, Biometrics
'Most of the conferences you go to are about pushing products - this one is about sharing knowledge. One of the best ones I've been to. I definitely want to recommend the next event to a colleague!'
David Swann, Nationwide
There is no ceiling for the limit of AI progressions! This is my second RE•WORK event after Deep Learning Summit London last year, and my company is more than happy to pay for me to attend!
Quentin Vajda, ING Belgium
As well as presentations, we were also recording podcasts, filming interviews, and attendees were welcomed into interactive workshops. Vasco Pedro from Unbabel led us through a session exploring how AI is redefining omni-channel customer service for the better. The workshop, ‘Transforming Customer Service with Multilingual Conversations’ explored how AI plus humans can deliver exceptional customer experience, in 28 languages, at scale, with multilingual live chats. Vasco Pedro explained how the ‘Unbabel API works with multiple business platforms such as Salesforce, Slack, MailChimp and others. It adapts to any workflow in customer service.’ One of the current challenges in chatbots is that the management of dialogue is still fundamentally based on scripts which lacks the necessary detail to handle more complex customer requests.
Whilst attendees were learning from Unbabel, over in the AI Assistant track, Marc Paulina from Google explained how we can create empathic assistants to enhance the user experience. Users have high expectations - they want to be able to speak to their assistant in their natural language. When your Google home for instance starts speaking to you, your expectations get even higher. It next asks you a question and you realise it can learn from you, so people get even higher expectations. ‘These devices are everywhere - they don’t have an off switch, they’re ready to listen when we need them. This emotional part of the experience isn’t something we need to wait for face detection or other emerging technologies, we can already do a lot in the design process of the existing assistants. These devices are with you throughout the emotional rollercoaster of every day, so it’s important that they really understand you.’
At the lunch break, we managed to speak to lots more attendees who shared their highlights so far:
“I believe the next big change in AI will be in education. My research is based on how AI is changing the way we think and educate.I think it’s important to educate women right now so there is less of a gender gap in 5 years. From a financial perspective, AI is definitely starting to infiltrate all areas, but especially important is its effects on sales prediction, chat bots and operational AI (especially day to day).”
Mona Kaminski, Tel Aviv University
“I already have some great ideas from the retail track. I’ve been working on technology in shopping carts and developing apps to replace coin access. This is actually my third RE•WORK event and I’m booked for a fourth. The presentations and information from each of these events has helped patent applications.”
This afternoon, in a panel discussion ‘What are the Key Challenges & Risks of Implementing AI Into the Financial Sector?’ Karthik Ramakrishnan from ElementAI explained that 'executives across finance need to adopt AI as a strategy to be an AI First bank'. Joining Karthik on the panel were AJ Hashim from Barclays, Elena Chatzimichali from HSBC and John Spindler from AI Seed who was moderating. Elena touched on how both healthcare and finance are both behind the AI curve. 'There are a few challenges: firstly the lack of trust, education and information from the point of doctors in healthcare, and also the fear of taking on a challenge - should we trust these machines? Similarly, banks have to push the boundary as there is a lot of fintech companies who are pushing forward and they have to keep up.'
Bringing the day to a close on the Deep Learning in Retail and Advertising track was another panel discussion,'How Can You Best Enrich the Customer Experience Using AI?, which raised some interesting points on the current use of AI in online sales. In the discussion, Filipe Barbosa from Unbabel explained that language has always been important in making sales offline and that needs to be replicated online to connect with customers. Adam McMurchie from RBS left us with an interesting thought, saying that 'every time a chatbot fails and passes the customer on to a human, your product is perceived to have failed. Take time developing chat bots to reduce likelihood of failure.'
We’re excited to be back tomorrow to hear from more leading minds from the likes of Bloomsbury AI, Deutsche Telekom Group, Tractica, ASOS and many more. In the afternoon the three tracks will be coming together in the plenary session, a panel surrounding privacy, security and ethics. Running alongside on the workshop track, RE•WORK are also offering the opportunity for startups to meet with leading AI investors, VCs and industry experts to pitch ideas, and gain advice on their business ideas and strategies.
Follow us on Twitter @teamrework and join the conversation using the #reworkRETAIL, #reworkFIN and #reworkAI tags.