Moderator Highlights: Quant Strats May 2022

Quant Strats 5 May 2022, New York, was a huge success. We had over 200+ attendees at the event, finally networking face-to-face after 2 years! There were incredible sessions covering Quantitative Data Strategies, AI, Machine Learning, and Alternative Assets.

At Quant Strats, attendees had the opportunity to learn from 40+ top experts in the industry, with our star moderators guiding the sessions. Hear from the moderators, as we talk with them about their key takeaways and more:

Moderator: Lilian Quah, Managing Director, Epoch Investment Partners

Session: Panel – Training ML models after Covid – reducing risk and increasing yield

Panelists: Oliver Faltin-Trager, Portfolio Manager, Emso Asset Management; Peng Cheng, Head of Machine Learning Strategies, JP Morgan

Hi Lilian, thank you for joining us today and for moderating a great panel session. We were wondering what your key takeaways were from your session?

Thanks, I am happy to help out.

I think my key takeaways would be:

1. Using ML models is a craft that requires judgment and care. Judgment is required for model selection, training data selection, and interpretation.

2. There is a range of potentially suitable ML techniques for different use cases. Knowing what to use when depends on context. In some cases using an ensemble of models can be the right answer.

3. It is very difficult to know when to turn off or retire a model. Having a strong prior about when a model “should work” can help, as can monitoring the cost of implementing a trading strategy based on the model’s predictions.

What do you think our community will be looking forward to for next year?

I think for next year the “hot” topics will be crypto assets, with perhaps a focus on new product and portfolio management.

Moderator: Edward Tong, Executive Director, Applied AI & Machine Learning, JP Morgan

Session: Panel – Understanding ML application and modeling and building capability in your organization

Panelists: Michael Beal, CEO, Data Capital Management; Mayank Saxena, VP – QIS Trading, Société Générale; Dr Aitor Muguraza, Head of Quantitative Modelling and Data Analytics, Kaiju Capital Management; Anthony Maylath, Quant Research – Equities, JP Morgan 

Thank you, Edward, for moderating our panel at Quant Strats 2022.

Thanks for having me.

What were your key points from your session?

1. Full automation of machine learning (ML) trading is feasible and there have been examples of successful implementation.

2. ML can be used as a tool to boost the performance of existing performant trading strategies, e.g. meta-labeling by Ernie Chan, and Marcos Lopez de Prado.

3. ML may not be best for highly non-stationary data. Consider Markov chain, transition state, or econometrics methodologies, e.g. use econometrics ideas to transform data to stationary conditions before deploying ML, ensemble models with econometrics methods

What were your highlights of the event?

I enjoyed the KEYNOTE FIRESIDE CHAT: The Computer-Assisted Human: how partnership between engineers and investment professionals can deliver exceptional results with the speakers Andrew Janian, Head of Equities Engineering, Citadel and Perry Vais, Head of Equity Quant Research, Citadel.

It was an illuminating discussion from Citadel senior leadership on how to foster high-performing teams of investment professionals, quants, and engineers to harvest alpha. Learning about their ideas for gradually sharing alpha discoveries among their teams, their opinions on optimizing fundamental, quantitative strategies, and apprehensions for quantamental approaches.

What are your predictions for what will be the ‘hot topic’ at Quant Strats 2023?

Deep hedging methods for option and exotic hedge optimization and the associated auto-encoder simulator. Then, ML methods for refining ESG data.

Thanks for your input, we hope to see you at Quant Strats again soon.

Look forward to next time!

Moderator: Vasant Dhar, Founder, SCT Capital

Session: Keynote Fireside Chat with David Ruksin, CTO, WorldQuant

Hi Vasant, thanks for moderating our first session of the day, it was a great introduction to the day! Could you please give us your three takeaways:

1. Platforms that manage money need a sophisticated engineering setup to rapidly evaluate new strategies/managers to assess how additive they are; this requires an integration of alpha generation and execution to recognize liquidity constraints.

2. Platforms need an engineering set up to evaluate the potential alpha from new datasets accurately and efficiently. WorldQuant evaluates on average two new datasets per week.

3. Investment platforms are increasingly global, with teams across multiple time zones and markets. Platforms must address the management and technology challenges associated with such teams.

Thankyou Vasant, these are very insightful.

Great, hope to see you next year.

Yes, hope to see you at Quant Strats US 2023 (March 14, 2023)!

Are you interested in hearing more on these topics? We are not done yet…Quant Strats will be back in London live in October 2022. Make sure to keep up to date with our page to find out more soon!