Deep Reinforcement Learning for Risk-Aware Crypto Portfolio Management

Summary

The paper presents a deep reinforcement learning agent for portfolio management that balances return seeking with risk restraint. Its target policy adjusts how strongly the agent favors the optimal action, using a tunable greediness parameter to encourage lower-risk choices. The authors evaluate the approach on cryptocurrency market data, selected for its abundant minute-level observations and high volatility.

In the reported test period, the agent returned 1800% and had the lowest risk among the compared methods. Further experiments indicate robust performance under high market volatility and short training periods. The excerpt does not specify the risk measure, comparison methods, assets, transaction costs, or evaluation design, so the performance claims cannot be assessed in detail or assumed to generalize beyond the stated experiments.

Key ideas

• The proposed agent optimizes portfolio management with attention to both profit and risk restraint.
• A tunable target policy controls preference for the optimal action and is intended to favor lower-risk actions.
• The approach is evaluated using cryptocurrency market data with minute-level observations.
• The authors report a 1800% test-period return and the lowest risk among compared methods.
• Additional experiments suggest robustness to high volatility and short training periods, though the excerpt omits evaluation details.

Tags

# Automatic Financial Trading Agent for Low-risk Portfolio Management using Deep Reinforcement Learning


# Automatic Financial Trading Agent for Low-risk Portfolio Management using Deep Reinforcement Learning









The autonomous trading agent is one of the most actively studied areas of artificial intelligence to solve the capital market portfolio management problem. The two primary goals of the portfolio management problem are maximizing profit and restrainting risk. However, most approaches to this problem solely take account of maximizing returns. Therefore, this paper proposes a deep reinforcement learning based trading agent that can manage the portfolio considering not only profit maximization but also risk restraint. We also propose a new target policy to allow the trading agent to learn to prefer low-risk actions. The new target policy can be reflected in the update by adjusting the greediness for the optimal action through the hyper parameter. The proposed trading agent verifies the performance through the data of the cryptocurrency market. The Cryptocurrency market is the best test-ground for testing our trading agents because of the huge amount of data accumulated every minute and the market volatility is extremely large. As a experimental result, during the test period, our agents achieved a return of 1800% and provided the least risky investment strategy among the existing methods. And, another experiment shows that the agent can maintain robust generalized performance even if market volatility is large or training period is short.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.