Dmitriy Gizlyk / プロファイル

We invite you to explore the SEW-ResNet framework, which allows you to build deep spiking models without degradation issues and with effective gradient management. In this article, we demonstrate how to implement a basic spiking neuron and its algorithms using MQL5.

Neural networks are already changing the way we analyze markets, and new architectures are opening up even more possibilities. In this article, we wrap up our work with the SpikingBrain framework, which opens up new possibilities for us.

We present to the reader an implementation of the SpikingBrain framework’s approaches based on recurrent linear attention with gates, discussed in detail in this article. Forward pass, gradient propagation, and weight update algorithms ensure efficient processing of financial time series and enable the practical implementation of the framework’s key concepts.

The SpikingBrain framework demonstrates a unique approach to data processing: neurons respond only to significant events, effectively filtering out noise. Its event-driven architecture reduces computational costs while preserving key information about price movements. Adaptive thresholds and the ability to use pre-trained modules ensure the model's flexibility and scalability.

The article provides a detailed examination of the integration of the ST-Expert framework's approaches into the Extralonger architecture, which enables the simultaneous analysis of temporal and spatial representations of data. The results of testing on real historical data are presented, demonstrating the model's effectiveness and its robustness to market anomalies. The article describes the framework's modular structure, which ensures reproducibility, flexibility for research, and the ability to optimize components in stages.

In this article, we continue implementing the ST-Expert framework approaches, focusing on the practical aspects of applying them using MQL5. Earlier, we examined the theoretical foundations and key components of the model; now we move on to working directly with graph attention algorithms and local and global attention distribution. The main goal of this work is to demonstrate how ST-Expert's conceptual ideas are transformed into workable solutions for analyzing and forecasting financial time series.

In this article, we will explore the ST-Expert framework, which ensures the robustness of forecasts under market uncertainty by taking local and global dependencies in time series into account. Its flexible architecture promotes model adaptability and improves the accuracy of predictions.

The Extralonger framework demonstrates a unique ability to integrate spatial and temporal factors into a single model, ensuring high forecast accuracy. Its architecture allows it to adapt to different planning horizons and financial instruments while maintaining the system's transparency and manageability.

We are continuing our work on implementing the approaches proposed by the authors of the Extralonger framework. This time, we will focus on building a Global-Local Spatial Attention module using MQL5, examining both its structure and its practical integration into the overall computational process.

The Extralonger framework demonstrates an approach to integrating spatial and temporal factors into a single model, which makes it possible to account for both local patterns and long-term cycles simultaneously. This architecture makes time series forecasting more resilient to market noise and enables data analysis across different time horizons. The article takes a detailed look at how these ideas are put into practice using OpenCL and MQL5.

In this article, we conclude our work on building the SAGDFN framework using MQL5, summarizing the development process and presenting the results of its practical testing. Let's combine the modules we've already implemented into a single system, highlight the strengths of this approach, point out its weaknesses, and discuss possible ways to improve it.

In this article, we will take a detailed look at the practical implementation of the key components of the SAGDFN framework. We will show how sparse attention and the selection of significant neighbors are organized for time series forecasting. The approaches presented strike a balance between forecast accuracy and computational efficiency.

In this article, we explore the architecture of SAGDFN — a modern framework capable of transforming the approach to processing spatiotemporal data. It preserves key information even in complex graphs while reducing computational costs.

The article describes a practical implementation of the HimNet framework based on MQL5, ready for integration into automated trading. We demonstrate how heterogeneity-adapted meta-parameters transform the model into a universal tool capable of handling fluctuating volatility.

In this article, we take a detailed look at the algorithms used to implement the key components of the HimNet framework. We demonstrate how, with a minimal number of trainable components, a high degree of consistency and controllability can be achieved throughout the entire system. The presented implementation is compact and transparent, which makes it easier to adapt to real-world market tasks.

We invite you to explore the HimNet framework, which combines the flexibility of spatio-temporal adaptation with high computational efficiency, enabling accurate and stable forecasts for financial time series. The article explains in detail how its key components interact with one another, transforming complex algorithms into a manageable architecture.

TQNetフレームワークの開発とテストも、いよいよ最終段階を迎えました。今回は、理論を実際の取引環境に結び付け、過去のデータによる学習から、直近の市場データを用いたストレステストへと進みます。 モデルの頑健性と予測精度を評価し、最終的な結果を通じて、提案手法が実際の取引においてどのような価値を持つのかを明らかにします。 ここで得られるのは、単なる統計上の数値ではなく、実用性を具体的に示す結果です。

TQNetフレームワークは、モジュール性、柔軟性、高い処理性能を兼ね備え、金融時系列データのモデリングと予測に新たな可能性を切り開きます。 高度なパラメータ初期化手法も含め、大域的な相関関係を捉えるための複雑なメカニズムを実装する可能性について考察します。

時系列データを意味のある層に分解し、それらを活用して簡潔なモデルを構築するアルゴリズムをご紹介します。 そのアーキテクチャからMQL5/OpenCLによる具体的な実装、過去の市場データを用いた実証テストまで、体系的に解説します。

In this article, we continue our hands-on exploration of SSCNN — a next-generation architectural solution capable of processing fragmented time series. Instead of blind scaling — smart modularity, attention to detail, and targeted normalization. Step by step, we are creating computational blocks in the MQL5 environment and laying the foundation for reliable predictive analysis.