Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/108206
Title: | Detección de eventos anómalos en un entorno industrial mediante el uso de técnicas de Federated Learning |
Author: | García Carretero, Darío Martín |
Tutor: | Parada Medina, Raúl |
Others: | Casas-Roma, Jordi |
Abstract: | An anomalous event is one that occurs suddenly and without foresight. In an industrial environment, these events, generally machine failures, can cause great economic and personal damage, so their detection can help prevent irreversible situations. The objective of this project is to show how to detect anomalies in industrial equipment using a machine learning model. To teach a model to distinguish between normal and abnormal behaviour, data is needed, the more data, the better. Today, most components within an industrial environment are monitored by specialized measuring devices. We can always have the data provided by our measuring devices to train the model, but what if we could have more data? Many industrial equipment is of generic use and can be used for many tasks and in many types of installation. If we could have access to the data of all those devices, we could create a much robust model. For various reasons companies refuse to share their data. For this reason, this paper proposes the use of Federated Learning (FL). Thanks to FL, models can be built taking advantage of all available information while maintaining data privacy. |
Keywords: | federated learning anomalous events industrial environment |
Document type: | info:eu-repo/semantics/masterThesis |
Issue Date: | 8-Jan-2020 |
Publication license: | http://creativecommons.org/licenses/by/3.0/es/ |
Appears in Collections: | Trabajos finales de carrera, trabajos de investigación, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
TFM - Audio.mp4 | Presentación | 117,55 MB | MP4 | View/Open |
code.zip | Código fuente | 28,61 kB | Unknown | View/Open |
darioTFM0120memoria.pdf | Memoria del TFM | 1,68 MB | Adobe PDF | View/Open |
Share:
This item is licensed under a Creative Commons License