Proceedings of Third International Conference on Sustainable Expert Systems

Subarna Shakya, Valentina Emilia Bălaş, Haoxiang WangLecture Notes in Networks and Systems 2023 · DOI · OpenAlex

Overview

The Third International Conference on Sustainable Expert Systems (ICSES 2022) serves as a pivotal forum for researchers to present high-impact work at the intersection of Artificial Intelligence (AI) and global sustainability. This volume compiles 76 peer-reviewed papers that demonstrate a significant shift from general-purpose AI toward highly specialized, resource-efficient "Expert Systems" tailored for medicine, engineering, and social sciences.

Editors and affiliations for the ICSES 2022 proceedings.

Addressing the Implementation Gap

A primary focus of these proceedings is bridging the "implementation gap" in modern AI. While theoretical advancements are rapid, deploying these models to solve specific, real-world sustainability challenges remains a hurdle. The contributors tackle two core dimensions:

  1. Domain Specificity: Moving beyond generic models to address hyper-local problems, such as detecting malaria in thin blood smears or forecasting volcanic eruptions at Mount Etna.
  2. Sustainability: Ensuring these systems can be deployed within existing infrastructures—including IoT, 5G, and low-power edge devices—without requiring excessive computational resources.

Methodological Innovations

The methodologies presented represent the current state-of-the-art toolkit for engineers and data scientists:

  • Deep Learning & Transfer Learning: Extensive use of Convolutional Neural Networks (CNNs) for diagnostics and Generative Adversarial Networks (GANs) for data augmentation.
  • Hybrid Systems: Integration of Fuzzy Logic with deep learning to manage uncertainty in sensor data, particularly in renewable energy systems.
  • Decentralized Frameworks: Utilization of Blockchain for secure identity management and remote infrastructure monitoring via UAVs.
  • Optimization: Application of hybrid evolutionary algorithms and meta-embeddings to enhance web services and natural language processing.

Diverse Data Benchmarks

The research leverages a wide array of specialized datasets, including:

  • Medical: Luna16 for lung cancer and Giemsa-stained smears for malaria.
  • Linguistic: The LinCE dataset for multilingual code-switching.
  • Environmental: Historical geological data from Mount Etna and remote sensing for land-use analysis.
  • Industrial: Digital simulations for predictive maintenance and real-time traffic feeds for urban monitoring.

Key Insights: The Rise of Edge Intelligence

A recurring theme throughout the volume is the transition toward "Edge" Expert Systems. By moving intelligence closer to the data source—using hardware like Raspberry Pi for patient monitoring or FPGA-based modeling—researchers are paving the way for AI that is both sustainable and responsive. Furthermore, the proceedings highlight the multidisciplinary synergy required for success, where deep engineering domain knowledge is integrated with advanced machine learning.

Results and Selectivity

The conference maintained a high standard of academic rigor, accepting only 76 papers out of 327 submissions, representing a competitive ~23% acceptance rate. Key results include:

  • High Diagnostic Accuracy: Significant precision in detecting neurodegenerative disorders and chronic kidney disease.
  • Efficiency Gains: Reductions in computational overhead for vehicular network security and web service optimization.

Limitations and Future Directions

Despite these successes, challenges remain. Data scarcity in niche fields and the "black box" nature of deep learning continue to be barriers to adoption. Future work is expected to focus on "Network Sustainability," aiming for resilient, self-healing communication infrastructures powered by intelligent expert systems.

Summary updated

BibTeX
@inproceedings{shakya2023proceedings,
  author = {Subarna Shakya and Valentina Emilia Bălaş and Haoxiang Wang},
  title = {{Proceedings of Third International Conference on Sustainable Expert Systems}},
  booktitle = {Lecture Notes in Networks and Systems},
  year = {2023},
  doi = {10.1007/978-981-19-7874-6},
}
Figure