| New PDF pages (1-based) | # slides | What was added (titles / representative)                                                                                          |
| ----------------------- | -------- | --------------------------------------------------------------------------------------------------------------------------------- |
| 32                      | 1        | Hype Cycle for Artificial intelligence, 2025                                                                                      |
| 42                      | 1        | Chromosome, Genes and Genomes                                                                                                     |
| 60–61                   | 2        | Proportional and rank based selection; Roullete wheels for the proportional                                                       |
| 154                     | 1        | Example: two moons                                                                                                                |
| 157                     | 1        | Word bias have several meanings                                                                                                   |
| 168–175                 | 8        | Good question — measuring bias and variance…; Revision question:; Practical steps 2/3; Implementation; Interpretation             |
| 307                     | 1        | Another option:                                                                                                                   |
| 356                     | 1        | A few deep learning sucesses                                                                                                      |
| 504–505                 | 2        | Prompt engineering (PE); Preference optimization of LLMs                                                                          |
| 510–530                 | 21       | Direct preference optimization (DPO); (LLM training/usage topics incl. knowledge injection, where LLMs go, compute hunger, etc.)  |
| 532–536                 | 5        | Ethical dilemmas of using AI; Responsible use of AI in research; EU guidelines…; EU AI Act; Slovenska tovarna umetne intelligence |
| 599–606                 | 8        | XGBoost overview; XGBoost Workflow; XGBoost Step by Step…; Key properties…; Small Example: First Tree                             |
| 641–668                 | 28       | Transformers for tabular data / foundation models (contents → masked cell modeling → benefits)                                    |
| 670–684                 | 15       | Transformers for time series / TS foundation models (tokens idea → step-by-step → summary)                                        |
| 686–697                 | 12       | AutoML (why → pipeline → NAS → Bayesian optimization → strengths/limits)                                                          |
| 764                     | 1        | Example video: Bipedals                                                                                                           |
| 862–866                 | 5        | RL with human feedback (RLHF) → PPO → DPO → Training Procedure for DPO                                                            |