KEYWORD |
Vehicular data analysis based on Deep Learning
Thesis in external company
keywords CONTRASTIVE LEARNING, DEEP LEARNING, GRAPH NEURAL NETWORKS, MACHINE LEARNING
Reference persons LUCA CAGLIERO, FRANCESCO VACCARINO, LUCA VASSIO
Research Groups DAUIN - GR-04 - DATABASE AND DATA MINING GROUP - DBDM
Thesis type APPLIED RESEARCH
Description Research collaboration between PoliTo and Tierra Spa -- multinational telematics service provider (https://www.tierratelematics.com/)
Objectives:
- Profiling of fleets of industrial vechicles based the analysis on geo-spatial and CAN Bus Data (i.e., IOT signals monitoring engine and vehicle status)
- Definition, identification, and characterization of Points-Of-Interests (e.g., job sites, deposits, filling stations)
- Design, development and testing of machine learning models for smart vehicle management (based upon domain experts' indications)
Techniques:
- Clustering
- Contrastive learning
- Graph Neural Networks
Required skills Fundamentals of Python, Databases, Data Mining and Machine Learning
Deadline 31/08/2024
PROPONI LA TUA CANDIDATURA