000.000 / 000.000
LOG // CI
Prof. Dr.
Oliver Kramer
Professor of Computational Intelligence · University of Oldenburg
> We develop computational intelligence methods that connect evolutionary optimization, learning, cognition, and AI in application areas like computational biology.
01_evolutionary_projects
EvoLabEVOLUTION STRATEGIES STUDIO [EXPLORE]Interactive studio for evolution strategies on ten classic benchmarks — tune population, selection, recombination, and step size while the search unfolds.
MolEvolver LabEVOLVING MOLECULES [REPO]Evolves a molecule to match a pharmacophore you place in 3D, searching a library of 10,150 synthesis-constrained compounds right in the browser.
Machine Learning with Evolution StrategiesEVOLUTIONARY MACHINE LEARNING [COLAB]Hands-on Colab course: SVM, MLP, clustering, and CNNs written from scratch in NumPy and trained with one derivative-free optimizer — the (1+1)-ES.02_ai_projects
LangDashONE PHRASE · EIGHT LANGUAGES [APP]One phrase rendered side by side across the eight most spoken languages — with notes on idioms and grammar.
WIKIPEDIA DISTILLEDWIKIPEDIA, DISTILLED [APP]Explore nine connected Wikipedia articles at a time, distilled by AI into concise three-sentence summaries.
LMAPMANIFOLD LEARNING WITH LOCAL MODELS [MAP]Unfolds messy, high-dimensional data into a visible map of its structure by combining local PCA models with global MDS — privately, in the browser.03_teaching
Introduction to Artificial IntelligenceLECTURE SCRIPT [PDF]Scripted lecture from k-nearest neighbors and k-means through convolutional networks, attention, and large language models — each chapter with worked examples, exercises, and Colab notebooks.
Tiny TransformerINTERACTIVE EXPLAINER [INTERACTIVE]A tiny transformer opened up live in the browser — tokenization, attention and sampling explained step by step.
Introduction to Evolution StrategiesMEDIUM ESSAY [BLOG · 2026]A tour of evolution strategies on Medium — from the (1+1)-ES and the 1/5th success rule to self-adaptive step sizes on hard landscapes.04_selected_publications
- Linear Evaluation Complexity of Surrogate-Assisted (1+1)-EA on OneMaxESANN 2026 [2026]
- Adaptive Search in Collatz Exponent-Code Space via 2-adic and 3-adic ConstraintsCoRR abs/2607.10041 [2026]
- An LLM-Based Multi-Agent Framework for Evolutionary Blackbox OptimizationGECCO Companion 2025: 671–674 [2025]
- Enhancing Evolutionary Algorithms Through Meta-Evolution StrategiesIEEE CAI 2025: 1292–1297 [2025]
- LLaMA Tunes CMA-ESESANN 2024 [2024]
05_books
2026Artificial Intelligence Essentials[SPRINGER]
2017Genetic Algorithm Essentials[SPRINGER]
2016Machine Learning for Evolution Strategies[SPRINGER]
2014A Brief Introduction to Continuous Evolutionary Optimization[SPRINGER]
2013Dimensionality Reduction with Unsupervised Nearest Neighbors[SPRINGER]
2009Computational Intelligence[SPRINGER]
2008Self-Adaptive Heuristics for Evolutionary Computation[SPRINGER]
06_blog_articles
Cognitive BASICREASONING LANGUAGE [2025 · Medium]Cognitive operations as program lines make prompting explicit, inspectable, and systematic.
Beyond KohlbergMORAL PSYCHOLOGY FOR ASI [2025 · Medium]Human stage theories of moral development were never written for machines — a moral psychology for superintelligence that evolves instead of being imposed.
Unsupervised Regression for Dimensionality ReductionMANIFOLD LEARNING [2023 · Towards Data Science]Optimize latent coordinates instead of fitting a regression model — a flexible approach to manifold learning.07_misc
END OF TRANSMISSION