In the previous article, we introduced the DDPG method, which allows training models in a continuous action space. However, like other Q-learning methods, DDPG is prone to overestimating Q-function values. This problem often results in training an agent with a suboptimal strategy. In this article, we will look at some approaches to overcome the mentioned issue.
In this article, we expand the range of tasks of our agent. The training process will include some aspects of money and risk management, which are an integral part of any trading strategy.
Data classification for purposes of analysis and forecasting is a very diverse arena within machine learning and it features a large number of approaches and methods. This piece looks at one such approach, namely Agglomerative Hierarchical Classification.
In this article, we will have a look at yet another reinforcement learning approach. It is called goal-conditioned reinforcement learning (GCRL). In this approach, an agent is trained to achieve different goals in specific scenarios.
Training useful skills without an explicit reward function is one of the main challenges in hierarchical reinforcement learning. Previously, we already got acquainted with two algorithms for solving this problem. But the question of the completeness of environmental research remains open. This article demonstrates a different approach to skill training, the use of which directly depends on the current state of the system.
In the previous article, we introduced the DIAYN method, which offers the algorithm for learning a variety of skills. The acquired skills can be used for various tasks. But such skills can be quite unpredictable, which can make them difficult to use. In this article, we will look at an algorithm for learning predictable skills.
The problem of reinforcement learning lies in the need to define a reward function. It can be complex or difficult to formalize. To address this problem, activity-based and environment-based approaches are being explored to learn skills without an explicit reward function.
In the context of reinforcement learning, model procrastination can be caused by several reasons. The article considers some of the possible causes of model procrastination and methods for overcoming them.
With the rapid development of artificial intelligence today, language models (LLMs) are an important part of artificial intelligence, so we should think about how to integrate powerful LLMs into our algorithmic trading. For most people, it is difficult to fine-tune these powerful models according to their needs, deploy them locally, and then apply them to algorithmic trading. This series of articles will take a step-by-step approach to achieve this goal.
The article describes hierarchical training models that offer an effective approach to solving complex machine learning problems. Hierarchical models consist of several levels, each of which is responsible for different aspects of the task.
This article discusses the use of the Go-Explore algorithm over a long training period, since the random action selection strategy may not lead to a profitable pass as training time increases.
With the rapid development of artificial intelligence today, language models (LLMs) are an important part of artificial intelligence, so we should think about how to integrate powerful LLMs into our algorithmic trading. For most people, it is difficult to fine-tune these powerful models according to their needs, deploy them locally, and then apply them to algorithmic trading. This series of articles will take a step-by-step approach to achieve this goal.
Dive into the world of ONNX, the powerful open-standard format for exchanging machine learning models. Discover how leveraging ONNX can revolutionize algorithmic trading in MQL5, allowing traders to seamlessly integrate cutting-edge AI models and elevate their strategies to new heights. Uncover the secrets to cross-platform compatibility and learn how to unlock the full potential of ONNX in your MQL5 trading endeavors. Elevate your trading game with this comprehensive guide to Mastering ONNX
This series of articles introduces several time series labeling methods, which can create data that meets most artificial intelligence models, and targeted data labeling according to needs can make the trained artificial intelligence model more in line with the expected design, improve the accuracy of our model, and even help the model make a qualitative leap!
In this article we continue with our theme in the last of tackling everyday trading indicators viewed in a ‘new’ light. We are handling horizontal composition of natural transformations for this piece and the best indicator for this, that expands on what we just covered, is the double exponential moving average (DEMA).
In this article, we will explore the application of all classification models available in the Scikit-Learn library to solve the classification task of Fisher's Iris dataset. We will attempt to convert these models into ONNX format and utilize the resulting models in MQL5 programs. Additionally, we will compare the accuracy of the original models with their ONNX versions on the full Iris dataset.
In this article we attempt to simplify our illustration of concepts covered in these series by dwelling on just one indicator, the most common and probably the easiest to understand. The moving average. In doing so we consider significance and possible applications of vertical natural transformations.
We continue studying the environment in reinforcement learning models. And in this article we will look at another algorithm – Go-Explore, which allows you to effectively explore the environment at the model training stage.
One of the key problems within reinforcement learning is environmental exploration. Previously, we have already seen the research method based on Intrinsic Curiosity. Today I propose to look at another algorithm: Exploration via Disagreement.
This article, the 21st in our series, continues with a look at Natural Transformations and how they can be implemented using linear discriminant analysis. We present applications of this in a signal class format, like in the previous article.
Regression is a task of predicting a real value from an unlabeled example. The so-called regression metrics are used to assess the accuracy of regression model predictions.
We digress in our series by pondering at part of the algorithm to chatGPT. Are there any similarities or concepts borrowed from natural transformations? We attempt to answer these and other questions in a fun piece, with our code in a signal class format.
This series of articles introduces several time series labeling methods, which can create data that meets most artificial intelligence models, and targeted data labeling according to needs can make the trained artificial intelligence model more in line with the expected design, improve the accuracy of our model, and even help the model make a qualitative leap!
In the previous article, we discussed relational models which use attention mechanisms in their architecture. One of the specific features of these models is the intensive utilization of computing resources. In this article, we will consider one of the mechanisms for reducing the number of computational operations inside the Self-Attention block. This will increase the general performance of the model.
We continue our look at natural transformations by considering naturality square induction. Slight restraints on multicurrency implementation for experts assembled with the MQL5 wizard mean we are showcasing our data classification abilities with a script. Principle applications considered are price change classification and thus its forecasting.
This series of articles introduces several time series labeling methods, which can create data that meets most artificial intelligence models, and targeted data labeling according to needs can make the trained artificial intelligence model more in line with the expected design, improve the accuracy of our model, and even help the model make a qualitative leap!
This article continues our series into category theory by introducing natural transformations, a key pillar within the subject. We look at the seemingly complex definition, then delve into examples and applications with this series’ ‘bread and butter’; volatility forecasting.
Object-oriented programming enables creation of a more compact code that is easy to read and modify. Here we will have a look at the example for three ONNX models.
This article, the final in our series to tackle functors as a subject, revisits monoids as a category. Monoids which we have already introduced in these series are used here to aid in position sizing, together with multi-layer perceptrons.
This article, the 16th in our series, continues with a look at Functors and how they can be implemented using artificial neural networks. We depart from our approach so far in the series, that has involved forecasting volatility and try to implement a custom signal class for setting position entry and exit signals.
This article on Category Theory implementation in MQL5, continues the series by looking at Functors but this time as a bridge between Graphs and a set. We revisit calendar data, and despite its limitations in Strategy Tester use, make the case using functors in forecasting volatility with the help of correlation.
This article which is part of a broader series on Category Theory implementation in MQL5, delves into Functors. We examine how a Linear Order can be mapped to a set, thanks to Functors; by considering two sets of data that one would typically dismiss as having any connection.
This article, that follows Category Theory implementation of Orders in MQL5, considers how database schemas can be incorporated for classification in MQL5. We take an introductory look at how database schema concepts could be married with category theory when identifying trade relevant text(string) information. Calendar events are the focus.
This material provides a complete guide to creating a class in MQL5 for efficient management of CSV files. We will see the implementation of methods for opening, writing, reading, and transforming data. We will also consider how to use them to store and access information. In addition, we will discuss the limitations and the most important aspects of using such a class. This article ca be a valuable resource for those who want to learn how to process CSV files in MQL5.
This article which is part of a series that follows Category Theory implementation of Graphs in MQL5, delves in Orders. We examine how concepts of Order-Theory can support monoid sets in informing trade decisions by considering two major ordering types.
Here we will describe only one of the aspects of machine learning - activation functions. In artificial neural networks, a neuron activation function calculates an output signal value based on the values of an input signal or a set of input signals. We will delve into the inner workings of the process.
This article continues the series on category theory implementation in MQL5. Here we continue monoid-actions as a means of transforming monoids, covered in the previous article, leading to increased applications.
The article provides an example of using a perceptron as a self-sufficient price prediction tool by showcasing general concepts and the simplest ready-made Expert Advisor followed by the results of its optimization.
In this article we discuss methods related to the analysis of timeseries in the frequency domain. Emphasizing the utility of examining the power spectra of time series when building predictive models. In this article we will discuss some of the useful perspectives to be gained by analyzing time series in the frequency domain using the discrete fourier transform (dft).
Neural networks are an ultimate tool in traders' toolkit. Let's check if this assumption is true. MetaTrader 5 is approached as a self-sufficient medium for using neural networks in trading. A simple explanation is provided.