Logo Lanfrica

AI-driven Data Management of Traditional Tunisian Nutritional Dishes: A Cultural Heritage Conservation

Domaine:

natural language processing

Type de record:

paper
Créateur:
OlfMarHibMal
Éditeur:
Ins
Hôte:
The potential loss of traditional Tunisian dishes threatens the sustainability of valuable cultural and nutritional traditions. To help preserve this rich heritage, a conversational AI system has been developed that employs advanced language processing and machine learning techniques to bring Tunisia’s culinary traditions to life in a digital space. Multilingual transformer models have been adapted to understand Tunisian dialects and combined with a detailed Food Heritage Knowledge Graph, allowing personalized, interactive access to authentic recipes and the stories behind them. A hybrid dialogue system operated by a chatbot has been implemented to ensure smooth, meaningful conversations that respect cultural sensitivities and build user trust and engagement.Despite challenges such as dialect diversity and limited data, it is demonstrated that modern AI can effectively capture and share complex cultural knowledge. Plans are underway to expand dialect support through federated learning and to improve contextual understanding with smarter memory models. Overall, this project highlights how technology and tradition can be connected through AI, supporting cultural preservation, promoting gastronomic tourism, and encouraging healthier eating habits in Tunisia.

Similaires