Communication and Generative Ai
A.Y. 2025/2026
Learning objectives
Within the area of computer science, students will acquire knowledge and understanding of:
- what artificial intelligence is and how it works — language models and automatic generation of images, audio and video;
- basic principles of generative models and how to interact with them (prompting);
- AI applications in digital communication;
- criticalities: reliability, bias and transparency of AI-generated contents.
- what artificial intelligence is and how it works — language models and automatic generation of images, audio and video;
- basic principles of generative models and how to interact with them (prompting);
- AI applications in digital communication;
- criticalities: reliability, bias and transparency of AI-generated contents.
Expected learning outcomes
- Using generative AI for digital communication projects.
- Writing effective prompts to get relevant and coherent results from generative models.
- Critically analysing AI-generated contents, evaluating their communicative quality, context adherence and potential bias.
- Incorporating generative AI in the writing and editing process, and in the production of multimedia contents.
- Writing effective prompts to get relevant and coherent results from generative models.
- Critically analysing AI-generated contents, evaluating their communicative quality, context adherence and potential bias.
- Incorporating generative AI in the writing and editing process, and in the production of multimedia contents.
Lesson period: First semester
Assessment methods: Esame
Assessment result: voto verbalizzato in trentesimi
Single course
This course can be attended as a single course.
Course syllabus and organization
Single session
Responsible
Lesson period
First semester
Course syllabus
Module 1 - Foundations of Artificial Intelligence and Basic Terminology
Definition of Artificial Intelligence, Machine Learning, Deep Learning, generative models, and information retrieval.
Basic concepts of machine learning: types of learning, data, tasks, models, and loss functions.
Module 2 - Natural Language Processing: Concepts and Tools
Classic NLP techniques: Bag of Words, TF-IDF.
Tokenization, stemming, lemmatization.
Semantic similarity and text representation models.
Introduction to representation learning models for textual data.
Module 3 - Generative Models: Development and Functioning
Model architecture and training (Transformer, LLMs).
Training phases, inference, fine-tuning.
Reinforcement Learning from Human Feedback (RLHF).
Retrieval-Augmented Generation (RAG), Deep Search, and hybrid models.
Module 4 - Evaluation of Generative AI Models
Reference datasets and benchmarks.
Evaluation tasks and metrics.
Module 5 - Prompt Engineering
Core techniques: zero-shot, one-shot, and few-shot prompting.
Advanced prompting: Chain of Thought (CoT), ReAct, system/role prompting.
Best practices in prompting and contextualization in communication.
Module 6 - Multimodal Generative AI
Overview of representation and models for the generation of images, audio, and video.
Limitations of multimodal generative AI.
Module 7 - Critical and Ethical Issues
Bias in data and models: sources, impacts, and mitigation strategies.
Reliability and transparency: explainability, watermarking, traceability.
Definition of Artificial Intelligence, Machine Learning, Deep Learning, generative models, and information retrieval.
Basic concepts of machine learning: types of learning, data, tasks, models, and loss functions.
Module 2 - Natural Language Processing: Concepts and Tools
Classic NLP techniques: Bag of Words, TF-IDF.
Tokenization, stemming, lemmatization.
Semantic similarity and text representation models.
Introduction to representation learning models for textual data.
Module 3 - Generative Models: Development and Functioning
Model architecture and training (Transformer, LLMs).
Training phases, inference, fine-tuning.
Reinforcement Learning from Human Feedback (RLHF).
Retrieval-Augmented Generation (RAG), Deep Search, and hybrid models.
Module 4 - Evaluation of Generative AI Models
Reference datasets and benchmarks.
Evaluation tasks and metrics.
Module 5 - Prompt Engineering
Core techniques: zero-shot, one-shot, and few-shot prompting.
Advanced prompting: Chain of Thought (CoT), ReAct, system/role prompting.
Best practices in prompting and contextualization in communication.
Module 6 - Multimodal Generative AI
Overview of representation and models for the generation of images, audio, and video.
Limitations of multimodal generative AI.
Module 7 - Critical and Ethical Issues
Bias in data and models: sources, impacts, and mitigation strategies.
Reliability and transparency: explainability, watermarking, traceability.
Prerequisites for admission
No
Teaching methods
Lectures, group discussions, and computer-based exercises
Teaching Resources
Slides and lecture notes.
Prompt engineering book by Lee Boonstra
Prompt engineering book by Lee Boonstra
Assessment methods and Criteria
Written test
Educational website(s)
Professor(s)