A comprehensive approach to conceptual modelling and visual representation of log files
SAC '22
Log file modelling, UML
2022
Code N/A
Graphical representation of the network logs.
ABSTRACT
This paper proposes a comprehensive approach to conceptual log file modelling. It is based on the UML (Unified Modelling Language) class diagrams, which notation was extended to represent the main functionality of the nested structure of JSON (JavaScript Object Notation) log files. An implementation of the approach is illustrated in two domains: banking applications and networking. This study is a first step towards the sound development of a centralized log files management system. We use the ECS (Elastic Common Schema) initiative as a starting point. We also present the developed plugin to the StarUML software, which allows modelling logs in both UML and JSON formats.
In summary, our contributions are:
• We propose ComCa, an approach specifically designed for training-free open-vocabulary attribute detection. ComCa exploits the compositional nature of objects and attributes to build an auxiliary cache that is used as an anchor to refine the model's predictions.
• We make caching scalable by extracting priors from databases and LLMs on which objects are associated with each attribute, using them to populate the cache. Soft-labeling further accounts for the multiplicity of attributes within an image.
• We empirically show that ComCa outperforms all the cache-based baselines and achieves competitive performance against costly training-based methods, demonstrating the practical feasibility of training-free open vocabulary attribute detection.
• We propose ComCa, an approach specifically designed for training-free open-vocabulary attribute detection. ComCa exploits the compositional nature of objects and attributes to build an auxiliary cache that is used as an anchor to refine the model's predictions.
• We make caching scalable by extracting priors from databases and LLMs on which objects are associated with each attribute, using them to populate the cache. Soft-labeling further accounts for the multiplicity of attributes within an image.
• We empirically show that ComCa outperforms all the cache-based baselines and achieves competitive performance against costly training-based methods, demonstrating the practical feasibility of training-free open vocabulary attribute detection.
• Evelina Rakhmetova (University of Verona)
• Marco Garosi (University of Verona)
• Carlo Combi (University of Verona)
• Marco Garosi (University of Verona)
• Carlo Combi (University of Verona)
You can cite this work as:
@inproceedings{10.1145/3477314.3507693,
author = {Rakhmetova, Evelina and Garosi, Marco and Combi, Carlo},
title = {A comprehensive approach to conceptual modelling and visual representation of log files},
year = {2022},
publisher = {Association for Computing Machinery},
url = {https://doi.org/10.1145/3477314.3507693},
doi = {10.1145/3477314.3507693},
booktitle = {Proceedings of the 37th ACM/SIGAPP Symposium on Applied Computing},
series = {SAC '22}
}