2
$\begingroup$

I've been struggling to find engaging papers on the application of deep reinforcement learning in quantitative risk analysis, portfolio management, algorithmic trading and/or options pricing. What are some papers with interesting findings on this topic that you know of?

$\endgroup$
1
  • 2
    $\begingroup$ Your title says 'reinforcement learning', but your text demands 'DEEP reinforcement learning'. RL may be sufficient for many problems in the areas you specified. You may want to clarify which one you need. $\endgroup$ – ir7 Jan 10 at 17:05
2
$\begingroup$

You can find a lot of good papers by just typing keywords like "deep reinforcement learning finance" in the arXiv or Google Scholar or looking at top researchers websites which provides an overflow of applications and research directions to engage with. Anyway, here are a few off the top of my head:

If you are looking for a more introductory level paper, this master's thesis applied distributional reinforcement learning to the problem of optimal execution and provides code which is quite rare & nice: https://www.imperial.ac.uk/media/imperial-college/faculty-of-natural-sciences/department-of-mathematics/math-finance/TobyWestonSubmission.pdf

If you want more comprehensive, recent work, I enjoyed "Optimal Execution of Foreign Securities" by Cartea & Arribas which was a unique application of machine learning to optimal trading in the FX market: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3562251. Rama Cont and Justin Sirgnano wrote a short paper back in 2018 with respect to the problem of detecting price formation using deep learning: https://arxiv.org/pdf/1803.06917.pdf. Charles Lehalle also has some interesting work on not so much deep learning, but reinforcement learning you could find on his Google Scholar page: https://scholar.google.com/citations?hl=en&user=j44WqxsAAAAJ&view_op=list_works&sortby=pubdate. I really liked this paper by Zhang et al (2019) on deep learning to also predict price formation - it was covered in a lot of detail and they include a lot of citations if you want to look around some more: https://ieeexplore.ieee.org/abstract/document/8673598?casa_token=eealOPwLJEEAAAAA:B6eolFGeLCcidoJOLG_VUXPcBnHqKVXeNapUeHzME7A5VlG6g8PEKWjxAtWuQakSlxiMoOnf-g. While you asked for deep reinforcement learning, here's an application of q-learning to options pricing (there are many more): https://arxiv.org/pdf/1712.04609.pdf.

You can also find older, but still interesting papers like https://ieeexplore.ieee.org/abstract/document/8324237?casa_token=FXpE__5nGKwAAAAA:lgEhAx_r2-2B3My1HgzP8dJshgF2XFM88PyI6dRs23GMwg-sOl-P3NcJSuKV01_oaqKM9pPSYg which applies deep reinforcement learning to trading cryptos - warning: littered with typos the last time I read it.

$\endgroup$
1
$\begingroup$

I've worked in algorithmic trading for years. RL (or deep RL for that matter) is not used in this industry.

$\endgroup$
1
  • 1
    $\begingroup$ That is a too wide a generalization. RL is frequently applied to market making kind of problems. If this is the de facto production model for firms, is another question. What do you refer to when you say AlgoTrading, optimal execution? $\endgroup$ – Pontus Hultkrantz Jan 13 at 18:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.