Zhu, Zhao Wei (A.A. 2024/2025) Cognitive artificial intelligence as a cultural interpreter: understanding and reproducing deep cultural semantics across domains. Tesi di Laurea in Machine learning, Luiss Guido Carli, relatore Giuseppe Francesco Italiano, pp. 67. [Master's Degree Thesis]

PDF (Full text) Restricted to Registered users only Download (796kB) | Request a copy

Abstract/Index

Cultures don’t just “translate,” they think differently. Hypothesis: cognitive AI can embody, not just translate, meaning across cultural contexts. Understanding culture beyond language. What is culture? Why translation fails. Embodied culture: language + behavior + emotion. Curiosity: historical roots of cultural difference: Rome vs. ancient China. Implications for AI. Technical pathways to human-like AI. Overview of current AI architectures and their limitations. Neuro-symbolic systems: where logic meets learning. Graph neural networks: structured relational thinking. Multimodal foundation models (e.g., GPT-4V, Gemini). Affective computing: recognizing and responding to emotions. Memory-augmented neural networks and RAG frameworks. Cultural knowledge in action: real-world applications and translation tasks. Adaptive cultural translation. Food, dress, body language, and movement. Music, religion, holidays, and philosophy: interpreting traditions without “flattening”. Case study: a poem and a recipe-GPT vs. cognitive AI. Experimental study: human vs AI reasoning patterns. Evaluation criteria. Example: Chinese idiom–对牛弹琴 ("playing the lute to a cow"). Applications and commercial potential. AI-enhanced tourism: guiding experiences with cultural depth. AI in international marketing: cultural adaptation of brands and ads. Cultural personalization in global education platforms. Diplomatic and cross-border communication facilitation. AI-driven cultural onboarding for immigrants and expats. Risks and ethical considerations. Cultural misrepresentation or stereotyping by AI. Bias reinforcement in training data. Who owns cultural knowledge? Consent, identity and emotional manipulation. Guidelines for culturally respectful AI systems. Discussion and future directions. Can AI truly understand culture, or is it just high-resolution mimicry? Future models: embodied AI, continual learning and dynamic context memory. The potential for multilingual, multisensory cultural agents. A personal vision: toward humble, helpful, and human-compatible AI.

References

Bibliografia: pp. 65-66.

Thesis Type:	Master's Degree Thesis
Institution:	Luiss Guido Carli
Degree Program:	Master's Degree Programs > Master's Degree Program in Data Science e Management (LM-91)
Chair:	Machine learning
Thesis Supervisor:	Italiano, Giuseppe Francesco
Thesis Co-Supervisor:	Simeone, Antonio
Academic Year:	2024/2025
Session:	Summer
Deposited by:	Alessandro Perfetti
Date Deposited:	11 Sep 2025 14:15
Last Modified:	11 Sep 2025 14:15
URI:	https://tesi.luiss.it/id/eprint/43153

Downloads

Repository Staff Only

View Item