Leaders in Artificial Intelligence from industry and academia gathered in New York today to attend the first day of RE•WORK’s AI in Finance Summit. Over 250 attendees participated in a plethora of workshops, heard groundbreaking presentations from over 50 speakers and joined together to discuss the latest AI tools and techniques for finance during multiple networking breaks.
The AI in Finance Summit marks the fourth in the series of RE•WORK’s global AI and Deep Learning for Finance Summits. Being the first U.S edition however, we were once again delighted that speakers joined us in New York from all over the world, representing a wide range of leading global corporations such as Deutsche Bank, Citi and ING Bank. Furthermore, each RE•WORK summit welcomes a diverse audience, and today’s summit was no different, with attendees ranging from founders CEOs, and data scientists, to professors, PhD students, ethics officers and many more. The companies in which they represented were no different, with traditional financial institutions such as HSBC, KPMG and Intuit in attendance alongside Amazon, Uber and Walmart.
Upon arrival, attendees discussed what they were most looking forward to about the day ahead. Ying Sun of Vanguard believed that;
The summit represents a good opportunity to discover the real applications at an industry and company level, giving an overview of the most popular applications of AI research. There are many benefits to attending a RE•WORK conference!
On social media, Dave Jones, @InstinctiveDave mentioned that he was also “Looking forward to some new insights at #reworkFIN”
Courtney McGill of Bank of New York Mellon, began the day introducing Uday Singh, the Head of Process Automation and Robotics at Credit Suisse. He shared that automating IT infrastructure by deploying AI and ML across Credit Suisse has helped the company enhance customer and counterpart experience as well as improving the risk profile, fortifying controls and streamlining considerations.
Continuing the thread of how AI can currently assist financial services, Ambika Sukla of Morgan Stanley concluded his presentation stating that although AI provides companies with a lot of opportunities, there is “still is a lot of work that needs to be done for AI advancement in Finance, and we need to empower everyone to be able to contribute to this. We need to fully understand how models work and it is vital that we continue to involve humans in the loop”.
Finishing this introductory session, Yuanyuan Liu of AIG gave us a detailed insight into the relationship between AI and Insurtech, sharing the ‘Nine Killer Applications of Digital Technology in General Insurance”. Yuanyuan referred to AIG’s own journey, echoing Uday’s work at Credit Suisse, noting that AI has been extremely important in improving customer experience, something the company realizes as they get quicker and quicker at underwriting.
Breaking for coffee and cake, attendees eagerly explored the exhibition area. HyperScience, a machine learning company focused on automating office work, discussed how solutions allowed organizations to automate large amounts of manual data entry, even for documents with handwriting, poor image quality, or variable structure. FinBrain Technologies brought their Deep Learning algorithms, which analyze and predict the future movements of Financial Assets, as well as a delicious surprise - homemade Turkish Delight!
Alongside this, attendees relished the opportunity to “Ask the Experts Your Questions on AI in Finance”, where Vincent Tang (FINRA), Yibei McDermott (Deutsche Bank) and Valentino Zocca (Citi) led a roundtable discussion with attendees on everything data science!
Moving to delve deeper into the latest advancements in AI tools and techniques, Roxana Geambasu of Columbia University spoke about the importance of assessing robustness against malicious behaviour in machine learning models.
“As Deep Neural Networks deliver exceptional performance on tasks including credit fraud detection and malware classification, we must not forget that it is incredibly easy to fool DNN’s”
Finishing the morning session, Igor Halperin (NYU) shared his latest research on “Reinforcement Learning for Portfolio Optimization and Market Modeling” and shared his motivation behind his passion:
“Challenges that I have faced in my teaching has made me focus on what we do modelling wise when we don’t have enough data”
He drew attention to an interesting concern in the advancement of AI research and model testing; how do we advance research further if we do not have access to data? The difference between industry and academia was left in the air; do global companies who are able to more easily access testing data in comparison to academic institutions able to advance research faster? And quicker?
Moving into the lunch break, attendees reflected on the morning via social media:
1) Valentino Zocca, @ItalyHighTech - Many interesting talks at #rework summit on #AI in Finance in New York #reworkFIN
2) Dror Katzav, @drorkatzav - Exciting day today at @teamrework AI in Finance summit! #reworkFIN
The afternoon sessions began with the much anticipated session on “NLP and Fraud Detection”, both extremely hot topics at the moment.
Eric Charton (National Bank of Canada) explained that “chatbots vendors are notoriously difficult to deal with. Additionally, most chatbots today are not robust, and fail in most cases as they cannot deal with the infinite capacity of language. National Bank of Canada are using Rasa Stack, an open source model to build the chatbot to overcome this”.
Following this, as a Wall Street regulator, FINRA receives a lot of regulatory forms with text fields, and Yvonne Li mentioned that mining these fields using NLP is incredibly useful. Discussing the model selection that FINRA consider when tackling this consideration, Yvonne said “just because you have labels doesn’t mean you automatically choose supervised learning. To overcome data sparsity, we choose Global Vectors for Word Representation (GLoVe) algorithm at FINRA to try to automatically categorize termination reasons on a U5 Form”.
Another addition to the AI in Finance Summit was the workshop, “Investor Panel & Networking Session” which saw a panel of VC’s and startup mentors share their most protected industry insights and tips. Joe Chittenden-Veal of Greycroft, John Frankel of ff Venture Capital, Scarlett Sieber of Village Capital and Jillian Canning of Techstars NYC came together with moderator, Steven Kuyan of NYU Tandon Future Labs to address what all companies, and startups in particular, want to hear:
- How does the intersection of AI and finance differ from other industries? Are there more opportunities or is the opportunity space smaller with large enterprises investing in AI internally?
- What should AI companies be looking for from an investor when they are doing due diligence?
- How do you consider new investments in AI startups, especially those that employ some level of machine learning? Is there deep technical due diligence?
On a similar note, at the summit, RE•WORK assisted companies look for new talent in AI. We heard from innovative companies looking to hire including: National Bank of Canada, Skopos Labs and inpher who posted on our dedicated Job Board.
Other presentation highlights include;
- Wells Fargo's Bernhard Hientzsch gave a deeply technical presentation entitled "Deep Learning and Computational Graph Techniques for Derivatives Pricing and Analytics", talking us through a variety of approaches for FBSDE including; Han-Jentzen-E, Wang-Chen-Sudjianto-Liu-Shen and Beck-E-Jentzen.
- "The next three years’ demand for AI adoption in Financial Services will be higher than High Tech. The competition for patents and IP is accelerating." - Yuanyuan Liu, Director, Statistical Machine Learning, AIG Insurance
- Yaron Golgher & Lipa Roitman of I Know First, were quoted saying “The historical correlation between past algorithm predictions and accrual market movement for each asset is the key to identify and focus on the most predictable markets and securities, enhancing the overall performance” during their presentation on “Investment Selection By Combining Chaos Theory with AI”
- "$50 billion per year is spent on wages for data entry. HyperScience makes human-readable documents machine readable." — Peter Brodsky, CEO, HyperScience
After much knowledge was shared during the various presentations throughout the day, attendees reflected on what they had learnt and what this means for the future of AI in Finance:
It is clear that the resources available within a company affects their AI capabilities. To progress, it is important to consider the data science environment as a whole. In an ideal world, data engineers first transform the data in preparation for consumption by data scientists. What tools can data scientists use to their advantage to improve results? - Summit Attendee
To wrap up the day, a panel discussion on “Assessing the Regulatory Hurdle: AI in Finance” saw Peggy Tsai (Morgan Stanley Wealth Management), Jordan Brandt (Inpher), Gabrielle Haddad (Sigma Ratings) & Adrien Delle-Case (International Institute of Finance) agree that in the battle of AI vs Regulation, AI should in fact be thought of as a friend, rather than a foe.
Adrien, of the International Institute of Finance wants to demystify the idea that regulation is seen to be killing innovation and the application/use of machine learning and AI.
Gabrielle also does not view regulation as a barrier to technology, stating the Sigma Ratings have been engaging with regulators since the beginning of their journey to become the world’s first non-credit rating agency, viewing regulators as their partners. However, Gabrielle noted that there are still two very different perspectives on AI and regulation due to the long standing concern that “There is no AI rulebook, there is no law on AI”.
Coming from an entrepreneurial perspective, Jordan Brandt said that “Regulation is actually a catalyst to innovation, it is an opportunity (rather than a challenge).”
As well as moderating the panel discussion, Peggy Tsai also led a workshop entitled “Data Governance Strategy to Support Machine Learning” to the delight of attendees, who noted the session as a particular highlight of their day.
In celebration of the great wealth of knowledge shared throughout the day, attendees enjoyed a selection of drinks, solidifying new friendships and exploring their common interest of furthering the opportunities and reach of AI in Finance.
With HyperScience leading the way in application based AI for finance today, join us for Day 2 of the AI in Finance Summit tomorrow to hear ING Bank, Capital One and Bank of America Merrill Lynch & many more discuss this further!