Modeling Reality: Then and Now

Explore a 30-minute conference talk delivered by Iwona Białynicka-Birula at the Symposium celebrating Professor Iwo Birula-Białynicki's 90th birthday. Delve into the evolution of modeling reality, from past approaches to current methodologies. Gain insights into neural networks, their functionality, and the crucial factors driving their advancement, including computing power and data availability. Examine self-supervised learning techniques and language models, and understand various neural network architectures. Learn about lookup tables, translation processes, transformers, and decoders. Conclude with practical guidelines for working with neural networks, followed by a summary and Q&A session. This talk, presented at the University of Warsaw's Faculty of Physics in collaboration with the Center for Theoretical Physics of the Polish Academy of Sciences, offers a comprehensive overview of the progress in modeling reality and the current state of neural network technology.

Intro
Neural Networks Today
How Neural Networks Work
Computing Power
Data
Selfsupervised learning
Language models
Neural network architectures
Lookup table
Translation
Transformers
Decoder
Neural Network
Guidelines
Summary
Questions