User ticketing system with automatic resolution suggestions
| Autor: | Hameed, Aqib |
|---|---|
| Přispěvatelé: | Padró, Lluís, Mirats Tur, Josep Maria, Universitat Politècnica de Catalunya. Departament de Ciències de la Computació |
| Jazyk: | angličtina |
| Rok vydání: | 2022 |
| Předmět: |
Neural networks (Computer science)
sistema de venda d'entrades Natural language processing (Computer science) Deep learning (Machine learning) xarxes neuronals artificials Xarxes neuronals (Informàtica) Deep learning Informàtica::Intel·ligència artificial [Àrees temàtiques de la UPC] Tractament del llenguatge natural (Informàtica) NLP artificial neural networks ticketing-system Aprenentatge profund |
| Zdroj: | UPCommons. Portal del coneixement obert de la UPC Universitat Politècnica de Catalunya (UPC) |
| Popis: | In the recent years, neural networks are very popular in the field of the artificial intelligence (AI). We decided to design the project on the basis of it. Neural network is the branch of the machine learning that uses the different layers to represent the data. Data are transformed to different or multi layers and generate the output. Our project is User ticketing system with automatic resolution suggestions, in which data is text based and with the help of Natural language processing (NLP) we can solve the problem. NLP is a sub-field of artificial intelligence which creates interactions between computers and human language. The goal is a computer capable of "understanding" the contents of documents, including the contextual nuances of the language within them. In this project the main goal is measuring the similarity between the texts. We design different models and use different layers and hyper-parameters later we will discuss in details. Our models are based on the Siamese neural network (SNN). SNN (sometimes called a twin neural network) is an artificial neural network that uses the same weights while working in tandem on two different input vectors to compute comparable output vectors. In this project, three main approaches have been used. Firstly, on the basis of the pre-trained model we get the output. Secondly, we follow the transfer learning approach. The transfer learning consists of taking features learned on one problem, and leveraging them on a new, similar problem. Thirdly, the model is fine tunned and designed on the basis of mutilingual (Spanish and English). |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|