TelecomParis_IPParis.png Telecom Paris
Dep. Informatique & Réseaux

nils.png Nils HolzenbergerHome page

July 2024

5

NeurSymAI.png Neuro-Symbolic Artificial Intelligence

with             Simon Coumes     portrait.jpg         and         Zacchary Sadeddine     Zac.jpg

            other AI courses


5


Introduction

The course is organized around weekly topics. For each topic, there is a lecture followed by a lab session. For the lab sessions, students connect to the relevant page on the web site. They will find some text to read, small open questions and small programming exercises that they can try on their own machine (or on the machines provided in the lab room). Answers are recorded and contribute to the final grade. For most questions, a possible solution becomes accessible after answering. Lab sessions can be completed during until just before the next lecture (see dates in the table below). Answers are no longer recorded beyond the deadline indicated for each topic, which means students will need to work on a weekly basis.

Exam

There will be a small quiz (on paper, no documents, no turned-on device) at the end of the course. The final quiz will consist in small short and independent exercises about Prolog, logic and other topics. Answers to lab exercises will be read and evaluated. They will contribute to the final grade (~40%). No documents, no functionning devices. See below for examples of past exams.

Labs

Lab sessions are in rooms equipped with machines, but your are welcome to use your own. We will be working with the free Prolog Interpreter SWI-Prolog. Don’t hesitate to ask questions to teachers during the labs, they are there FOR YOU.

Topics



Topics
Dates are 2025     Overview
18 feb 15:00 ➜ 4 mar 13:30     First steps in Prolog
4 mar 15:00 ➜ 11 mar 13:30     Problem solving and Knowledge representation
11 mar 15:00 ➜ 18 mar 13:30     Propositional Logic
18 mar 13:30 ➜ 25 mar 13:30     Predicate Logic
25 mar 15:00 ➜ 1 apr 13:30     Machine Learning
1 apr 15:00 ➜ 8 apr 13:30     ProbLog: Probabilistic Prolog
8 apr 15:00 ➜ 18 apr 23:59     Statistical Machine Learning in Problog
(no lab session)     Review session
15 apr 15:15 ➜ 16:45     Exam   Past exams:   202420232022

   read →
PdfIcon.png     A paper on combining external knowledge and large language models    
Gabriel Poesia, Kanishk Gandhi, Eric Zelikman and Noah D. Goodman, Certified Deductive Reasoning with Language Models, Arxiv 2023.
   read →
PdfIcon.png     Interesting paper in The New Yorker on ChatGPT seen as a lossy compression of the Web    
Chiang, T. (2023). ChatGPT is a blurry JPEG of the Web. Annals of Technology - The New Yorker, Feb.
   read →
PdfIcon.png     An interesting short paper on the discovery of concepts and rules in science    
Evans, J. & Rzhetsky, A. (2010). Machine science. Science, 329 (5990), 399-400.

Line.jpg