Filippo Castellarin
DeepBlocks: visually constructing, debugging and training Deep Neural Networks.
Rel. Luigi De Russis, Tommaso Calo'. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2024
|
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Share Alike. Download (3MB) | Preview |
|
Archive (ZIP) (Documenti_allegati)
- Altro
Licenza: Creative Commons Attribution Share Alike. Download (647kB) |
Abstract: |
Programming Deep Neural Networks (DNNs) can be a daunting task for individuals with limited programming expertise. Moreover, the potential of Deep Learning (DL) models could be vastly expanded if more people had access to this technology. Conversely, for those already coding Neural Networks (NNs), comprehending runtime errors, pinpointing issues, and resolving them remains a significant hurdle. DeepBlocks is introduced as a Visual Programming (VP) tool that enables the construction, debugging, training, and evaluation of DNNs through the addition and connection of blocks, along with the configuration of training parameters. The application was tested by actual users, whose feedback was instrumental in further development and also conveyed gratitude for democratizing the power of Deep Learning. This was achieved by providing an intuitive tool that allows a wider audience to experiment with Deep Neural Networks. |
---|---|
Relatori: | Luigi De Russis, Tommaso Calo' |
Anno accademico: | 2024/25 |
Tipo di pubblicazione: | Elettronica |
Numero di pagine: | 94 |
Soggetti: | |
Corso di laurea: | Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering) |
Classe di laurea: | Nuovo ordinamento > Laurea magistrale > LM-32 - INGEGNERIA INFORMATICA |
Aziende collaboratrici: | NON SPECIFICATO |
URI: | http://webthesis.biblio.polito.it/id/eprint/33246 |
Modifica (riservato agli operatori) |