CEU Electronic Theses and Dissertations, 2022
| Author | Endes-Nagy, Peter |
|---|---|
| Title | Siamese Networks: Ranking Objects by Poorly Defined Similarity |
| Summary | Capstone project summary. A machine learning tool was developed for a Hungarian SME, specialised on manufacturing custom-made machine tools and metal products. The unit cost of creating operation plans is relatively high, developing a model that retrieves Top-k similar objects from the historical dataset has high added value for the business. Challenge: similar in terms of "how to manufacture", but similarity is poorly defined. Solution: generate pairwise labels with heuristics and train a siamese network |
| Supervisor | Divényi, János |
| Department | Economics MSc |
| Full text | https://www.etd.ceu.edu/2022/endes-nagy_peter.pdf |
Visit the CEU Library.
© 2007-2021, Central European University