Abstract:Artificial Intelligence (AI) has emerged in popularity recently, recording great progress in various industries. However, the environmental impact of AI is a growing concern, in terms of the energy consumption and carbon footprint of Machine Learning (ML) and Deep Learning (DL) models, making essential investigate Green AI, an attempt to reduce the climate impact of AI systems. This paper presents an assessment of different programming languages and Feature Selection (FS) methods to improve computation performance of AI focusing on Network Intrusion Detection (NID) and cyber-attack classification tasks. Experiments were conducted using five ML models - Random Forest, XGBoost, LightGBM, Multi-Layer Perceptron, and Long Short-Term Memory - implemented in four programming languages - Python, Java, R, and Rust - along with three FS methods - Information Gain, Recursive Feature Elimination, and Chi-Square. The obtained results demonstrated that FS plays an important role enhancing the computational efficiency of AI models without compromising detection accuracy, highlighting languages like Python and R, that benefit from a rich AI libraries environment. These conclusions can be useful to design efficient and sustainable AI systems that still provide a good generalization and a reliable detection.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is in the field of network security, how to improve the computational efficiency of intrusion detection systems and reduce environmental impacts by optimizing programming languages and feature selection (FS) methods. Specifically, the research focuses on the following points:
1. **Reducing the environmental impact of AI models**: With the rapid development of artificial intelligence (AI) in various industries, its impact on the environment has become increasingly significant, especially in terms of energy consumption and carbon footprint. The paper explores how to reduce the environmental burden of AI systems through green AI without affecting performance.
2. **Optimizing the performance of intrusion detection systems**: The research evaluates the impact of different programming languages (Python, Java, R, Rust) and feature selection methods (information gain, recursive feature elimination, chi - square test) on the computational efficiency and detection accuracy of machine learning (ML) and deep learning (DL) models in intrusion detection tasks.
3. **Balancing performance and resource consumption**: Through experimental verification, find the optimal configuration that can significantly reduce computing time and resource consumption while maintaining high detection accuracy, thereby providing a reference for designing efficient and sustainable AI systems.
### Research Background
In recent years, the wide application of AI technology has brought significant progress, but it has also been accompanied by huge energy consumption and carbon emission problems. For example, training a large - scale AI model such as OpenAI's GPT - 3 may consume about 1,287 MWh of energy, which is equivalent to the electricity consumption of 125 households in the United States in a year. Therefore, the concept of green AI emerged, aiming to reduce the energy consumption and environmental impact of AI systems by optimizing algorithms and hardware usage.
### Research Methods
To achieve the above - mentioned goals, the research adopts the following methods:
- **Selection of programming languages**: Compare the performance differences of four programming languages (Python, Java, R, Rust) when performing the same task.
- **Feature selection methods**: Apply three feature selection methods (information gain, recursive feature elimination, chi - square test) to reduce the input data dimension and improve the model efficiency.
- **Model selection**: Select five common ML and DL models (random forest, XGBoost, LightGBM, multi - layer perceptron, long - short - term memory network) for experiments.
- **Data sets**: Use two data sets (BotIoT and Hikari - 22) to conduct intrusion detection experiments to ensure the diversity and representativeness of the results.
### Experimental Results
The experimental results show that the feature selection method plays an important role in improving computational efficiency, and the choice of different programming languages also has a significant impact on the model performance. In particular, Python and R have achieved a good balance between performance and resource consumption due to their rich support for AI libraries. In addition, the XGBoost and random forest models perform excellently in most cases, but there are problems of slow training speed under certain programming languages.
### Conclusions and Prospects
Through this research, the author hopes to provide valuable information for researchers and practitioners, helping them to ensure high - efficiency performance and reduce environmental impacts when designing intrusion detection systems. Future research directions include further exploring the combinations of other programming languages and feature selection methods, and applying the principles of green AI to a wider range of network security scenarios.
In summary, this paper explores how to achieve efficient and environmentally - friendly AI solutions in intrusion detection systems through systematic experiments and analysis, providing an important reference for promoting the development of green AI.