We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Combining Reinforcement Learning and Inverse Reinforcement Learning for Asset Allocation Recommendations

Abstract: We suggest a simple practical method to combine the human and artificial intelligence to both learn best investment practices of fund managers, and provide recommendations to improve them. Our approach is based on a combination of Inverse Reinforcement Learning (IRL) and RL. First, the IRL component learns the intent of fund managers as suggested by their trading history, and recovers their implied reward function. At the second step, this reward function is used by a direct RL algorithm to optimize asset allocation decisions. We show that our method is able to improve over the performance of individual fund managers.
Comments: 9 pages, 12 figures
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computational Finance (q-fin.CP); Portfolio Management (q-fin.PM)
Cite as: arXiv:2201.01874 [cs.LG]
  (or arXiv:2201.01874v1 [cs.LG] for this version)

Submission history

From: Igor Halperin [view email]
[v1] Thu, 6 Jan 2022 00:08:17 GMT (1826kb,D)

Link back to: arXiv, form interface, contact.