Resilient distributed machine learning using network reconfiguration
Machine learning (ML) is currently making great impact in our today's life. However computations may become excessively huge on big data scenarios, so recently distributed models have become an interesting field of study. This thesis presents a reconfiguration strategy over a network of distrib...
- Autores:
-
Ángel Imitola, Jesús Gabriel
- Tipo de recurso:
- Fecha de publicación:
- 2020
- Institución:
- Universidad de los Andes
- Repositorio:
- Séneca: repositorio Uniandes
- Idioma:
- eng
- OAI Identifier:
- oai:repositorio.uniandes.edu.co:1992/50992
- Acceso en línea:
- http://hdl.handle.net/1992/50992
- Palabra clave:
- Aprendizaje automático (Inteligencia artificial)
Big Data
Seguridad en computadores
Ingeniería
- Rights
- openAccess
- License
- https://repositorio.uniandes.edu.co/static/pdf/aceptacion_uso_es.pdf
| Summary: | Machine learning (ML) is currently making great impact in our today's life. However computations may become excessively huge on big data scenarios, so recently distributed models have become an interesting field of study. This thesis presents a reconfiguration strategy over a network of distributed machine learning models, in order to boost its reliability against certain types of byzantine faults or attacks. The attacks considered are some attacks that are usually used against machine learning models or concerning consensus. Simulations on a distributed support vector machine (DSVM) network are made to evaluate our reconfiguration strategy under the previously considered attacks. |
|---|