2. Professors
  3. LA ROCCA Michele
  4. Teaching

NEURAL NETWORKS AND ARTIFICIAL INTELLIGENCE

Michele LA ROCCA NEURAL NETWORKS AND ARTIFICIAL INTELLIGENCE

0222400045
DEPARTMENT OF ECONOMICS AND STATISTICS
EQF7
STATISTICAL SCIENCES FOR FINANCE
2024/2025

OBBLIGATORIO
YEAR OF COURSE 1
YEAR OF DIDACTIC SYSTEM 2014
SPRING SEMESTER
CFUHOURSACTIVITY
530LESSONS
MICHELE LA ROCCA T Curriculum
ExamDate
LA ROCCA09/06/2025 - 14:30
LA ROCCA24/06/2025 - 14:30
LA ROCCA08/07/2025 - 14:30
LA ROCCA02/09/2025 - 14:30
Objectives
KNOWLEDGE AND UNDERSTANDING

THE COURSE AIMS TO PROVIDE ADVANCED STATISTICAL LEARNING TOOLS BASED ON NEURAL NETWORKS. STUDENTS WILL LEARN BOTH THE BASIC THEORETICAL CONCEPTS AND THE COMPUTATIONAL SKILLS NECESSARY FOR THE SPECIFICATION AND ESTIMATION OF BOTH SHALLOW AND DEEP NEURAL NETWORKS. PARTICULAR EMPHASIS WILL BE PLACED ON THE USE OF THIS CLASS OF MODELS FOR THE STUDY OF THE DEPENDENCE BETWEEN STATISTICAL VARIABLES AND THE STUDY AND FORECAST OF DYNAMIC PHENOMENA, WITH A SPECIFIC FOCUS ON BIG DATA AND FINANCIAL AND INSURANCE APPLICATIONS.

APPLYING KNOWLEDGE AND UNDERSTANDING

BASED ON THE KNOWLEDGE LEARNED, THE STUDENT WILL DEVELOP THE ABILITY TO:

– APPLYING NEURON MODELS TO DIFFERENT TYPES OF DATA (CONTINUOUS, DISCRETE, ETC.), TO DIFFERENT DEPENDENCY SCHEMES (REGRESSION AND TIME SERIES), AND TO DIFFERENT APPLICATION AREAS WITH A PARTICULAR FOCUS ON FINANCIAL AND INSURANCE PROBLEMS.

– USE THE STATISTICAL LANGUAGE R FOR THE IMPLEMENTATION OF THE COURSE OBJECT MODELS

– AUTONOMOUSLY AND CRITICALLY ANALYZE AND EVALUATE DOCUMENTS AND REPORTS THAT INCLUDE INFORMATION GENERATED WITH NEURONAL NETWORKS, FORMULATING CRITICAL JUDGMENTS ON THE PARTICULAR ARCHITECTURE USED, ON THE INFERENCE TECHNIQUES AND ON THE PREDICTIVE MODELS CONSTRUCTED AS WELL AS ON THE VALIDITY, INTERNAL AND EXTERNAL, OF THE CONCLUSIONS REACHED.
Prerequisites
KNOWLEDGE OF MATRIX CALCULATION NOTIONS, BASIC PROGRAMMING, THE R STATISTICAL LANGUAGE, AND REGRESSION MODELS IS REQUIRED (AT LEAST AT AN INTRODUCTORY LEVEL).
Contents
MODELS INSPIRED BY BIOLOGICAL NEURAL NETWORKS AND SOME HISTORICAL NOTES ON THEIR EVOLUTION. SHALLOW AND DEEP NEURAL NETWORKS. RECURRENT NEURAL NETWORKS. APPLICATIONS OF ARTIFICIAL INTELLIGENCE TO NOTABLE PROBLEMS (4H). UNIVERSAL APPROXIMATION THEOREMS. ACTIVATION FUNCTIONS AND LOSS FUNCTIONS (4H). NEURAL NETWORKS AND REGRESSION. NUMERICAL OPTIMIZATION METHODS FOR NEURAL NETWORKS: GRADIENT DESCENT AND ITS VARIANTS, BFGS AND LBFGS, BACKPROPAGATION. MODEL VERIFICATION AND VALIDATION TECHNIQUES. K-FOLD CROSS-VALIDATION AND DIAGNOSTIC TESTS (8H). APPLICATIONS AND CASE STUDIES ON THE REGRESSION PROBLEM IN R USING KERAS/TENSORFLOW, TORCH, AND THEIR VARIANTS (4H). NEURAL NETWORKS AND CLASSIFICATION. APPLICATIONS AND CASE STUDIES ON THE CLASSIFICATION PROBLEM IN R USING KERAS/TENSORFLOW, TORCH, AND THEIR VARIANTS (4H). NEURAL NETWORKS FOR THE ANALYSIS AND FORECASTING OF NON-LINEAR TIME SERIES. CROSS-VALIDATION IN A TIME SERIES CONTEXT. APPLICATIONS AND CASE STUDIES IN R FOR NON-LINEAR TIME SERIES FORECASTING (6H).
Teaching Methods
THE COURSE INCLUDES 30 HOURS OF CLASSROOM TEACHING. ALTHOUGH ATTENDANCE IS NOT MANDATORY, GIVEN THE NATURE OF THE COURSE, IT IS STRONGLY RECOMMENDED.
DURING THE LESSONS, THEORETICAL ISSUES WILL BE DISCUSSED, CONSTANTLY SUPPORTED BY THE IMPLEMENTATION OF THE METHODOLOGIES PROPOSED IN THE STATISTICAL LANGUAGE R, BY THE PRESENTATION OF CASE STUDIES THROUGH WHICH THE OPERATIONAL METHODS OF IMPLEMENTATION OF THE TECHNIQUES WILL BE ILLUSTRATED, THE CONTEXTS OF USE OF THE DIFFERENT TOOLS AND WILL BE CLARIFIED THE POSSIBLE INTERPRETATIONS OF THE RESULTS OBTAINED. THE EXERCISES WILL, THEREFORE, CONSTITUTE AN INTEGRAL PART OF THE PLANNED LESSONS. THE COURSE WILL BE DELIVERED IN ENHANCED WEB MODE. THEREFORE, THE LECTURES WILL BE ACCOMPANIED BY A WEB SPACE FOR DISTRIBUTING HANDOUTS AND SUPPLEMENTARY READINGS, USEFUL DATASETS FOR DEVELOPING CASE STUDIES AND ANY CLARIFICATIONS ON THE TOPICS COVERED IN THE COURSE. THE AVAILABILITY OF THE WEB SPACE IS NOT TO BE UNDERSTOOD AS AN ALTERNATIVE OR REPLACEMENT FOR IN-PERSON LESSONS.
Verification of learning
THE STUDENT WILL BE EVALUATED DURING THE FINAL EXAM, WHICH WILL BE HELD ON THE DEPARTMENT'S SCHEDULED EXAM DATES.
THE STUDENT MUST DISCUSS PROJECT WORK DURING THE FINAL EXAM AND TAKE AN ORAL TEST ON THE SCHEDULED TOPICS. THE PROJECT WORK MUST BE AGREED UPON WITH THE TEACHER DURING THE COURSE, CARRIED OUT INDIVIDUALLY, AND AIMS TO EVALUATE THE STUDENT'S ABILITY TO SPECIFY AND VALIDATE NEURAL MODELS TO SOLVE A SPECIFIC PROBLEM AND COMMUNICATE THE RESULTS THROUGH A STATISTICAL REPORT. THE EVALUATION OF THE PROJECTS WILL BE CARRIED OUT TAKING INTO ACCOUNT THE FOLLOWING ASPECTS:
- EFFECTIVE FORMULATION AND FRAMING OF THE PROBLEM WITH CLEAR RESEARCH QUESTIONS;
- CORRECTNESS AND EFFECTIVENESS OF THE SPECIFICATION AND VALIDATION OF THE MODELS PROPOSED FOR THE SOLUTION OF THE PROBLEMS FORMULATED IN THE PREVIOUS POINT;
- CORRECTNESS AND EFFICIENCY OF THE COMPUTATIONAL SOLUTIONS ADOPTED;
- CORRECTNESS AND EFFECTIVENESS OF THE COMMENTS ON THE RESULTS OBTAINED;
- CONTENT, STRUCTURE AND COMMUNICATIVE EFFECTIVENESS OF THE REPORT.
THE FINAL GRADE, AWARDED OUT OF THIRTY, WILL CONSIDER THE QUALITY OF THE PROJECT WORK DEVELOPED, THE LEVEL OF THEORETICAL KNOWLEDGE ACQUIRED ON THE TOPICS IN THE PROGRAM, THE AUTONOMY OF ANALYSIS AND JUDGMENT, AND THE STUDENT'S PRESENTATION SKILLS IN THE ORAL TEST.
Texts
LECTURE NOTES, WEB RESOURCES, AND ARTICLES SUGGESTED BY THE TEACHER DURING THE COURSE WILL BE MADE AVAILABLE TO ALL ATTENDING STUDENTS THROUGH THE WEB PLATFORM, COMMUNICATED AT THE BEGINNING OF THE COURSE.

EFFECTIVE STATISTICAL LEARNING FOR ACTUARIES: NEURAL NETWORKS AND EXTENSIONS. MICHEL DENOUIT, DONATIEN HAINAUT, JULIEN TRUFIN, SPRINGER

TO FLEXIBLY RESPOND TO THE SPECIFIC NEEDS OF EACH STUDENT, THE TEACHER CAN RECOMMEND ALTERNATIVE OR ADDITIONAL READINGS TO STUDENTS WHO REQUEST THEM.
More Information
THE TEACHER PROVIDES FURTHER EXPLANATIONS AND METHODOLOGICAL SUPPORT TO STUDENTS DURING RECEPTION HOURS.
DAYS, TIMES AND PLACE OF RECEPTION, AS WELL AS ANY CHANGES, ARE COMMUNICATED ON THE TEACHER'S WEB PAGE.
Lessons Timetable

  BETA VERSION Data source ESSE3 [Ultima Sincronizzazione: 2025-04-14]