Deep Reinforcement Learning vs Reinforcement Learning

Deep Reinforcement Learning vs Reinforcement Learning

Deep Reinforcement Learning vs Reinforcement Learning Deep reinforcement learning and reinforcement learning are two of the most popular methods of machine learning currently in use in the financial markets.

Which is more useful?
Which is the right approach for you?

This article aims to compare these two methods.

Reinforcement Learning

Reinforcement learning is the training of machine learning models to make a sequence of decisions.

The model will usually start from random trails and then the model will train itself into a complicated model.

Reinforcement learning employs a system of rewards and penalties. With the idea to compel the computer to solve problems by itself. Human involvement is limited to changing the environment (Budek, 2018).

Think of the idea of determining when to trade a rook for a pawn in chess. That is the theory of reinforcement learning simplified.

Deep Reinforcement Learning

On the other hand, deep reinforcement learning is quite different from reinforcement learning.

As the construction of deep learning is based on the human brain. The overall model will be implanted in a much deeper data set. So deep learning models will consist of neural network layers used to learn more abstract features about particular data (Budek, 2018).

Deep learning is very useful in price forecasting in finance. Firstly, the recurrent neural network can be used for a time series database. Secondly, long short term memory models are a variation of RNN with additional parameters to support longer memory. Also, the deep learning method can be used in fraud detection in finance (Montantes, 2020).

Use Case

I also want to introduce one great example of an application of reinforcement learning in the financial market. Moreover, price setting strategies with reinforcement learning are very useful and interesting.

Dynamic stock price changes are a big challenge in understanding stock prices. For example, gated recurrent unit networks with reinforcement learning can have advantages. Also, extracting financial features that represent the intrinsic character of a stock can help stop loss and maximize profit in trading. And, I think maximizing profit with minimum capital investments is another great application.

In conclusion, both deep reinforcement learning and reinforcement learning can have great functions in the financial market. But, we should try to practice more to maximize these methods in our trading systems on a case by case basis, not in a broad application.

Rather, realizing that these approaches are as different as apples and oranges and must be used in very different applications.

Deep Reinforcement Learning vs Reinforcement Learning Written by Frank Chen & Edited by Vivian Fang


Budek, K. (2018, Jul 5). What is reinforcement learning? The complete guide. 

Chatterjee, S. (2021, Jul 16th). 7 Applications of Reinforcement Learning in Finance and Trading. Neptuneblog. 

Montantes, J. (2020, Jul 2). Deep Learning in Finance: Is This The Future of the Financial

Industry? Towards data science.

Nasdaq Head of Ai, JPM Head of Ai, Point72 Head of Data Solutions Discuss Ai with Alexander Fleiss.
Yann LeCun Deep Learning Interview