Stephanie Lo is the Head of Quantitative Driven Research (QDRSF) for State Street's Securities Finance business. She is part of State Street Associates, State Street's academic arm. In this capacity, Stephanie oversees multiple research projects related to Securities Finance. Jian Wu is a Trading Strategist for State Street's Securities Finance business. In this capacity, he designs and implements automated algorithmic strategies, creates complex trading platforms and works on quantitative research. He has expertise in mining and modeling large datasets and prototyping machine learning algorithms to enhance trading operations.
At the AI in Finance Summit in New York this September 5-6, Stephanie and Jian will be presenting their latest work in the space, so in advance of this presentation, we sat down with both Stephanie and Jian for this interview.
How did you start your work in AI and Finance - what came first?
Stephanie: For me, the finance side came first. I spent a couple years at the beginning of my career at a proprietary trading shop. As I started to formulate my own questions on markets, I realized that I had more to learn on theory and methods, which motivated me to pursue the economics PhD. Now, several years later, the content – the financial questions and market insights – still come first, and motivate the methods I apply, whether plain vanilla regression techniques or cutting-edge AI methods.
Jian: From the beginning, I have been focused on quantitative techniques within finance. More recently, I have transitioned from being research focused to functionally trading a book of business. Algorithmic trading is heavily dependent on quantitative methods and more generally technology and AI is no different. I spend a lot of time on thinking about AI’s application and where it can add value to the business initiatives.
What's one piece of advice you'd give to someone starting out in the field?
Stephanie: Always try to understand the fundamental business needs upfront. There can be a temptation to apply techniques simply because they seem differentiated or interesting, but practitioners need to understand first whether the methods will address the end user needs.
Jian: Oftentimes practitioners want to jump into model building to arrive at conclusions faster. Taking a deep dive into the data set prior to modeling is essential as it reinforces understanding of the dynamics of the data and can save time in the long run.
From your perspective, how is AI transforming research among financial firms?
Stephanie: Over the past several years, AI has increasingly penetrated research in the financial industry. However, we are still in a growth period, and there are still different views on whether, and how fully, to embrace AI. While some firms have made AI a cornerstone of their research strategy, other firms prefer to stick to more standard quantitative techniques or qualitative analysis. Whether these two camps stay separate, or eventually meet somewhere in the middle, is yet to play out.
Jian: AI techniques are being explored in many different fields among financial firms and certain fields are seeing greater success levels than others. Being closest to trading, AI techniques have certainly introduced a new dimension of a non-linear method that is additive to the traditional econometric methods but there is still a lot of work to be done. I think it is extremely important to be aware of certain pitfalls of using AI techniques as more and more emphasis gets put onto using them for a wide range of applications, such as predictive analytics, automated trading or understanding complex alternative data.
Information Classification: Limited Access
What are the goals of your teams and what problems are you trying to solve?
Stephanie: My team at State Street Associates strives to leverage academic partnerships and research expertise while collaborating with State Street’s Securities Finance business to provide differentiated research offerings. In our view, there is often a gap between industry and academia, and by bridging that gap, we can identify more impactful questions, study them more carefully, and drive to applied products that are motivated by business needs.
Jian: My team’s mandate is to constantly evolve trading. Algorithmic trading, in particular, revolves around using cutting edge-quantitative techniques along with a sophisticated technology infrastructure. Pairing people with the right expertise and focusing on core business challenges allow us to come up with innovative solutions. How we can design new trading strategies, optimize our book, create new products, interact faster and mine the valuable details from the data are constant questions we’re asking.
What are examples of business case studies where you have leveraged quant/AI techniques?
Stephanie: We have several projects focused on the events of stocks “going special” in the borrow market. There is no standard industry definition for what a special event is, and both defining the events and then finding robust independent predictors for those events has proven to be challenging. With our differentiated platform – both the access we have to our academic collaborators and the business’s position in the market – we are developing a novel, applied understanding of these events.
Jian: My team has been collaborating with Stephanie’s team on an internal algorithm that identifies securities of interest for the desk by incorporating alternative data sources. Ultimately, the goal would be to use machine learning techniques to integrate these sources in a more dynamic way. Another example is our recent collaboration with an internal data science team to look to leverage deep learning networks to predict security borrowing rates.
What are you most looking forward to about the summit?
Stephanie: With the diversity of perspectives and approaches around AI, there is a lot to learn from industry participants in this space. Events like these can be really impactful by gathering practitioners from all angles to share and discuss these views.
Jian: By attending this conference, I expect to learn the latest progress in AI techniques from my peers. Obviously, I am excited to hear about successful use cases , but I am also interested, sometimes even more so, to discuss solutions that did not ultimately work.
Understanding the different frameworks in both end states allow us to learn from others work so we can continue to innovate in our rapidly changing market.
About the speakers:
Stephanie Lo is the Head of Securities Finance Research. She is part of State Street Associates, State Street's academic arm. In this capacity, Ms. Lo oversees multiple research projects related to Securities Finance. These projects include the use of alternative data, integration of academic insights and methods, and quantitative techniques such as machine learning and artificial intelligence. Prior to joining State Street, Ms. Lo worked as a natural gas trader at a quantitative trading firm and as a management consultant at the Boston Consulting Group. She has expertise in quantitative methods and economics. She has produced multiple academic papers in economics. Ms. Lo holds a PhD in Economics, a Masters in Economics, and a Bachelor of Arts in Economics from Harvard University.
Jian Wu is an Algorithmic Trader for State Street's Securities Finance business. In this capacity, he designs and implements automated algorithmic strategies, creates complex trading platforms and works on quantitative research. He has expertise in mining and modeling large datasets and prototyping machine learning algorithms to enhance trading operations. Prior to joining State Street Global Markets, Mr. Wu worked as a quantitative researcher at State Street Associates. In this role, he was responsible for the research and development of proprietary investor behavior indicators as alpha signals. Mr. Wu holds a Ph.D. in Electrical and Computer Engineering and an M.S. in Mathematical Finance from Rutgers University. He also earned an M.S. and a B.S. in Optical Engineering from Zhejiang University. Mr. Wu is a certified Financial Risk Manager (FRM) holder.