El Hif, Chaima (2023) NLP based approaches to bring a high level of innovation to address IT issues PFE - Project Graduation, ENSTA.
![]() | PDF Restricted to Registered users only 699Kb |
Abstract
With the remarkable advancement of AI, businesses are striving to integrate generative AI into all of their processes. This raises a question about the confidentiality of company data when utilizing these technologies. With the advent of ChatGPT, companies are becoming increasingly concerned. Should they prohibit the use of these AIs, or should they adopt them with a significant risk of data leakage? We will propose a solution for detecting confidential data and personal identifiable information in the texts shared with these AIs. In a second phase, we will try to optimize IT processes within the company. In fact, despite the numerous advantages that IT can offer to a company, one cannot overlook the incidents that follow. These problems can lead to a loss of time and money for the company. This is why all stakeholders are eager to adopt predictive maintenance solutions. By using the historical incident data within the company, we will employ NLP to cluster incidents and attempt to find patterns to identify potential new incidents. Analyzing these incidents will also help us automate their resolutions, thereby reducing workload and consequently, maintenance costs.
Item Type: | Thesis (PFE - Project Graduation) |
---|---|
Uncontrolled Keywords: | T, incidents, confidentiality, PII, NLP, transformers, Large Language Models (LLM). |
Subjects: | Mathematics and Applications |
ID Code: | 9845 |
Deposited By: | Chaima Elhif |
Deposited On: | 30 oct. 2023 14:53 |
Dernière modification: | 30 oct. 2023 14:54 |
Repository Staff Only: item control page