Università degli Studi di Pavia

Facoltà di Ingegneria


Artificial Intelligence

A.A. 2024-2025

First Semester

Fri: 11:00 a.m. - 1:00 p.m., Aula 5

Fri: 2:00 p.m. - 4:00 p.m., Aula 5

Lectures & Suggested Readings:

  • Reports of errors in the resources below are always welcome
    1. 2024.10.04 (theory)

      Introduction [pdf]

      Alan Turing (Wikipedia)
      Computer chess (Wikipedia)

      Shannon, C., "Programming a Computer for Playing Chess", Philosophical Magazine, 41 (314), 1950 [pdf]

      BBC2 Horizon, "Out of Control", 2012 [video, on Dailymotion]

      D. Silver, et al., "Mastering the game of Go with deep neural networks and tree search", Nature, 529, 2016 [link]

      "AlphaGo - The Movie | Full Documentary", YouTube, 2020 [video]

    2. 2024.10.04 (theory)

      Symbolic reasoning [pdf]
      Language, schemas and reasoning

      Syllogism (ancient logic) (Wikipedia)

    3. 2024.10.11 (theory)

      Propositional logic [pdf]
      Boolean algebras, formal propositional language and its semantics, satisfiability, entailment

      Rules of inference, justified by entailment (Wikipedia)

    4. 2024.10.18 (theory)

      Entailment and algorithms [pdf]
      Turing machine, decision problems, computational complexity, entailment as a satisfiability problem (refutation)

      Automated Symbolic Calculus [pdf]
      Resolution by refutation, soundess and completeness, computational complexity

    5. 2024.10.25 (theory)

      First-order logic [pdf]
      First-order semantic structures, formal language, variables and quantifiers, satisfaction, entailment

    6. 2024.11.08 (theory)

      Semi-decidability of First-Order logic [pdf]
      Prenex normal form, skolemization, Herbrand's theorem

      First-Order resolution [pdf]
      Clausal form, unification, resolution method for first-order logic

    7. 2024.11.15 (theory)

      SLD resolution [pdf]
      Horn clauses, SLD resolution, logic programming

      Depth-first search (Wikipedia)
      Breadth-first search (Wikipedia)

      Two Prolog examples that can be executed SWI-Prolog online [link]:
      - Re-definition of Prolog append/3 using the function cons/2 [pl]
      A note [pdf] describing the trace of an example of SLD resolution with append/3
      - Regression to infinity and fairness [pl]

      Minimal Models [pdf]
      Horn clauses, Herbrand model, implicit database
      NOTE: This presentation is an additional resource, provided for completeness.

      Plausible reasoning [pdf]
      Negation as failure, closed world assumption, deductive, inductive and abductive reasoning

    8. 2024.11.22 (theory)

      Probabilitistic reasoning: representation and inference [pdf]
      Foundations of probability, probability space, random variables, Bayes' theorem, probabilistic inference

      Countable set (Wikipedia)
      Sigma algebra (Wikipedia - see simple set-based examples)
      Probability space (Wikipedia)
      Random variable (Wikipedia)

      "The Banach-Tarski Paradox" (semi-serious but also very informative), YouTube, 2015 [video]

    9. 2024.11.29 (theory)

      Graphical models [pdf]
      Graphical models and factorization, d-separation, inference, computational methods

      GeNIe Modeler (free academic license) available on BayesFusion LLC [download]

      Examples for GeNIe Modeler:
      - Alarm.xdsl [model]
      - Alarm_dataset.csv [csv]

    10. 2024.12.06 (theory)

      Supervised learning [pdf]
      Learning proababilities from observations: maximum likelihood estimator, maximum a posteriori probability, conjugate prior distribution

      Binomial distribution (Wikipedia)
      Beta distribution (Wikipedia)
      Conjugate prior (Wikipedia)

      Beta distribution interactive plot [link]

      Numerical supervised learning [pdf]
      Learning with models that cannot be optimized analytically, logistic regression, gradient descent, stochastic gradient descent, mini-batch gradient descent

      Logistic regression (Wikipedia)

    11. 2024.12.13 (theory)

      Unsupervised learning [pdf]
      Learning from incomplete observations: k-means, hidden random variables, EM algorithm, observability model, missing data

      - k-means (a.k.a. LBG) algorithm - online demo [link]

      - Voronoi tesselation with Lloyd relaxation - online demo [link]

      - Expectation Maximization in Python (Colab Notebook) [link]

      - Andrew Ng, "The EM algorithm", CS229 Lecture notes, Stanford [pdf]

      EM algorithm for missing data, examples for GeNIe Modeler:
      - Alarm.xdsl [model]
      - Alarm_missing_values_dataset.csv [csv]

    12. 2024.12.20 (theory)

      Causal Inference [pdf]
      Simpson's paradox, structural causal model, intervention, do-calculus, counterfactuals, path-specific counterfactual inference

      Examples with "GeNIe" (free academic version) available at BayesFusion llc [link]:
      - Berkeley Admission test [model]
      - Berkeley Admission test (modified) [model]
      - Berkeley dataset [csv]

    Instructor

    1. Marco Piastra

    2. Contact: marco.piastra@unipv.it


    Kiro

    1. Course info


    Exams

    1. See Faculty website


    Further resources:

    (There are no required textbooks for this course. The following books are recommended as optional readings)

    1. Mordechai Ben-Ari, Mathematical Logic for Computer Science (3rd Edition). Springer, 2012

    2. Kevin P. Murphy, Probabilistic Machine Learning: Advanced Topics, MIT Press, 2023. [Pre-print]


    Links

    1. Artificial Intelligence Reading Group


    1. Artificial Intelligence, A.A. 2023-2024 and before