MIT je ponudio vrlo ozbiljni besplatni tečaj oi strojnom učenju: Machine Learning with Python: from Linear Models to Deep Learning
koji traje 15 tjedana, a počinje 1.02.2023.g. pa se prijavite što prije na poveznici:

https://www.edx.org/course/machine-learning-with-python-from-linear-models-to?index=product&queryID=c5ed75f297498e8695711e4cb4a9a985&position=1

O tečaju kažu:

“Machine Learning with Python: From Linear Models to Deep Learning” is a course provided by MIT on the edX platform which you can audit for free. You can also upgrade for certification. There is just one little problem, its starts FEB 2023 but you can enroll now. Once the course start, you will get a notification and you can start learning.
The course level is advanced.
Prerequisites:
   • Proficiency in Python programming
   • Probability theory course
   • A college-level single and multi-variable calculus course
   • Vectors and matrices course
Curriculum:
   • Linear classifier, perception algorithms, etc.
   • Linear regression
   • Stochastic gradient descent, over-fitting, generalization
   • Recommender problems, collaborative filtering
   • Non-linear classification, kernels
   • Deep learning, backpropagation
   • Unsupervised learning: clustering
much more
This is a very rigorous course and definitely worth the time.
You will also have the complete three projects.
   • Automatic Review Analyzer
   • Digit Recognition with Neural Networks
   • Reinforcement Learning
This course is taught by Regina Barzilay (Delta Electronics Professor in the Department of Electrical Engineering and Computer Science and a member of the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology), Tommi Jaakkola (Thomas Siebel Professor in EECS and IDSS at MIT), and Karene Chu (Digital Learning Scientist and Research Scientist at Massachusetts Institute of Technology).
If you are looking to advance your knowledge of machine learning models, this is a golden opportunity.

Tečaj je besplatan ako ga samo slušate. Plaća se jedino ukoliko želite potvrdu da ste ga prošli.

Zanimljiv animirani film  - POVEZNICA - nastao na isječcima intervjua znanstvenika Stuart Russell na Radio Davosu na temu "Obećanja i opasnosti umjetne inteligencije" (The promises and perils of AI)" u kojemu se istražuju trenutna ograničenja umjetne inteligencije i mogućnosti stvaranja tehnologije kompatibilne s ljudima.

U nadolazećim godinama, umjetna inteligencija će vjerojatno promijeniti naš život - a vjerojatno i cijeli svijet. Ali ljudi se teško slažu oko toga kako će AI utjecati na naše društvo. Možemo li izgraditi AI sustave koji nam pomažu popraviti svijet? Ili smo osuđeni na robotsko preuzimanje? Istražite ograničenja umjetne inteligencije i mogućnosti stvaranja tehnologije kompatibilne s ljudima.

Ovaj se video temelji na isječcima intervjua iz Radio Davos Podcasta. Epizoda se zove "Obećanja i opasnosti AI - Stuart Russell na Radio Davosu".
Cijelu epizodu poslušajte ovdje: POVEZNICA. Stuard Russell je poznati britanski znanstvenik u području umjetne inteligencije i koautor jednog od najboljih udžbenika umjetne inteligencije:

Stuart J. Russell, Peter Norvig Artificial Intelligence: A Modern Approach 2nd Edition, Prentice Hall; 2nd edition (January 1, 2003) - POVEZNICA

Dubokom učenju se ne može vjerovati kaže Stephen Grossberg pionir modeliranja aktivnosti mozga ...

https://spectrum.ieee.org/deep-learning-cant-be-trusted

Izvrstan pregled različitih algoritama strojnog učenja:

https://towardsdatascience.com/an-overview-of-classic-reinforcement-learning-algorithms-part-1-f79c8b87e5af