Machine learning has emerged as a promising paradigm for enabling connected, automated vehicles to autonomously cruise the streets and react to unexpected situations. Reacting to such situations requires accurate classification for uncommon events, which in turn depends on the selection of large, diverse, and high-quality training data. In fact, the data available at a vehicle (e.g., photos of road signs) may be affected by errors or have different levels of resolution and freshness. To tackle this challenge, we propose an active learning framework that, leveraging the information collected through onboard sensors as well as received from other vehicles, effectively deals with scarce and noisy data. Given the information received from neighboring vehicles, our solution: (i) selects which vehicles can reliably generate high-quality training data, and (ii) obtains a reliable subset of data to add to the training set by trading off between two essential features, i.e., quality and diversity. The results, obtained with different real-world datasets, demonstrate that our framework significantly outperforms state-of-the-art solutions, providing high classification accuracy with a limited bandwidth requirement for the data exchange between vehicles.

Active Learning with Noisy Labelers for Improving Classification Accuracy of Connected Vehicles / Abdellatif, ALAA AWAD ABDELHADY; Chiasserini, Carla Fabiana; Malandrino, Francesco; Mohamed, Amr; Erbad, Aiman. - In: IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY. - ISSN 0018-9545. - STAMPA. - 70:4(2021), pp. 3059-3070. [10.1109/TVT.2021.3066210]

Active Learning with Noisy Labelers for Improving Classification Accuracy of Connected Vehicles

Alaa Awad Abdellatif;Carla Fabiana Chiasserini;
2021

Abstract

Machine learning has emerged as a promising paradigm for enabling connected, automated vehicles to autonomously cruise the streets and react to unexpected situations. Reacting to such situations requires accurate classification for uncommon events, which in turn depends on the selection of large, diverse, and high-quality training data. In fact, the data available at a vehicle (e.g., photos of road signs) may be affected by errors or have different levels of resolution and freshness. To tackle this challenge, we propose an active learning framework that, leveraging the information collected through onboard sensors as well as received from other vehicles, effectively deals with scarce and noisy data. Given the information received from neighboring vehicles, our solution: (i) selects which vehicles can reliably generate high-quality training data, and (ii) obtains a reliable subset of data to add to the training set by trading off between two essential features, i.e., quality and diversity. The results, obtained with different real-world datasets, demonstrate that our framework significantly outperforms state-of-the-art solutions, providing high classification accuracy with a limited bandwidth requirement for the data exchange between vehicles.
File in questo prodotto:
File Dimensione Formato  
TVT_RAINBOW_POLITO.pdf

accesso aperto

Descrizione: Articolo principale
Tipologia: 2. Post-print / Author's Accepted Manuscript
Licenza: PUBBLICO - Tutti i diritti riservati
Dimensione 1.36 MB
Formato Adobe PDF
1.36 MB Adobe PDF Visualizza/Apri
TVT_final.pdf

non disponibili

Descrizione: Articolo principale
Tipologia: 2a Post-print versione editoriale / Version of Record
Licenza: Non Pubblico - Accesso privato/ristretto
Dimensione 1.61 MB
Formato Adobe PDF
1.61 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2873984