Artificial Intelligence Investing
Valentijn van Nieuwenhuijzen, Chief Investment Officer at NN Investment Partners, the Dutch fund manager, posted the following on LinkedIn:
— NNInvestmentPartners (@NNIP) July 5, 2018
Last week I attended an excellent conference on Artificial Intelligence (AI) and Sentiment Analysis as applied to the field of finance, hosted by UNICOM in London. It was an inspiring experience, in which a broad range of AI specialists, sentiment data providers, academics, financial industry representatives and regulators shared their insights, innovations and applied technologies on these topics. Like many of us, I have been amazed by the digitalization of our societies and I am wondering more and more how to apply the benefits of digitalization in my personal and professional lives. This conference created a very insightful platform to get a better understand the way forward, especially in the field of asset management.
One of the interesting perspectives on the history of AI, which dates back to the early 1960s, showed how much of the technology has already been around for a long time. The big game-changer in recent years, however, has been the availability of exponentially growing amounts of “Big Data”, combined with increased computer power. This allows for a much broader and more effective utilization of AI techniques to facilitate problem-solving in a faster and more dynamic way than was previously possible, in many academic and business domains.
The conference provided fascinating examples of how to apply this in the field of finance. There was a special focus on what the new sentiment data can offer to drive stronger investment results and which new tools can extract the relevant information from sentiment and alternative data. For example, multiple examples of new “deep learning” techniques were presented on topics such as:
- Interpreting language
- Trading equity and bond markets with news sentiment data
- Identifying financial bubbles through usage of social media data
- Identifying influencers and superforecasters by analysing Twitter or Bloomberg news feeds
- Using innovative visualisations to interpret large data samples faster and better
- Forecasting macroeconomic variables with news data
- Identifying risk events or investment themes through deep learning or neural networks
The line-up included Marketpsych, which presented its sentiment indices. These indices constantly score the news and other media, analyse it and extract meaning, all using AI techniques. NN IP already uses these indices in its investment processes and is cooperating closely with Marketpsych on continued development of proprietary indices.
Along with the amazement and inspiration, there were a couple of important takeaways from all these new data, new AI tools and new ways to apply combinations. First, it is truly fascinating to see how much the digitalization of our societies is creating new opportunities to improve business models and problem-solving in almost every industry. Second, this certainly also holds for finance – an industry that combines a somewhat counterintuitive combination of a long history of data analysis and a still relatively modest use of Big Data and AI techniques. Thirdly, the latter is changing rapidly and FinTech companies are providing a rapidly expanding universe of useful data and tools (AI or other) that can be plugged into investment decision making processes. New money managers, purely data driven machine learning and other AI techniques have started to emerge. Finally, Big Data and AI are not a holy grail. Many experts at the conference shared insights on the power and the limitations of both. Although opinions also differ amongst the speakers, most thought that smart combinations between traditional statistics, AI and openminded human oversight are all important parts of an optimal decision making or forecasting model.
This conference enriched my insight on the way forward for AI investing and the use of sentiment and other alternative data sources in our investment approach. In the end, it is about the creativity and robustness of integrating these type of new information sources into our investment process. How it can be applied will vary across asset classes and market segments, and it will always be done in an adaptable and research-driven way. But it will be done, and I am convinced it will become an increasingly important driver of investment returns in what will be a more data-driven future, a future where smart digital augmentation of our human investment skill will prove to be a key differentiator.
More views on macroeconomics and asset allocation? Follow me on Twitter: @ValentijnvN
Deal combines Knoema’s knowledge management platform and data repository with Adaptive’s alternative data.
AI and machine learning have supplanted trading speed as a differentiator for capital markets firms.
Focus areas include listings, ESG data, and sustainable bonds.
Is it time to revisit and expand the Gramm-Leach-Bliley Act to include fintechs?
The new offering for Japanese institutional investors is set to go live in December.