GENERATIVE AI FOR SCIENCE: A Hands-On Guide for Students and Researchers

Design molecules. Predict protein structures. Accelerate climate models. All with AI.AI-designed drugs achieve 80-90% Phase I success rates. AlphaFold earned the 2024 Nobel Prize. Neural weather models outperform supercomputers 1000x faster. This book teaches you to build these systems yourself.Generative AI for Science is a hands-on guide for researchers and practitioners applying cutting-edge AI to scientific discovery. 540+ pages, 13 chapters, with runnable notebooks—zero setup required.WHAT YOU'LL LEARN:Transformers and Large Language ModelsDiffusion Models for molecules and proteinsGraph Neural Networks for chemistryPhysics-Informed Neural Networks (PINNs)Fine-tuning and domain adaptationMultimodal AI and MLOps deploymentAPPLICATIONS:Drug Discovery: Molecular generation and property predictionProtein Engineering: Structure prediction with ESMFoldClimate Science: Weather AI and neural surrogatesPhysics: PDE solving and simulation accelerationWHO THIS IS FOR:Scientists wanting AI skills for researchML engineers entering scientific applicationsGraduate students seeking hands-on curriculumEvery concept includes working code. Every technique solves real scientific problems. Developed from graduate courses and refined through dozens of AI workshops.Prerequisites: Basic Python. No deep learning experience needed.FREE RESOURCES INCLUDED:All source code, PowerPoint slides, and datasets are freely available on GitHub: github.com/jpliu168/Generative_AI_For_Science50+ Google Colab notebooks—run with one click, no installationComplete PowerPoint slides for every chapterSample datasets for hands-on practiceRegular updates based on reader feedbackCOMPLETE CHAPTER COVERAGE:Foundations: AI fundamentals, scientific data workflows, text and knowledge generation\Core Techniques: Data-to-data models, autoencoders, GANs, VAEs, physics-informed neural networks, neural surrogatesDomain Applications:Chemistry and Materials ScienceBiology and BiomedicinePhysics and EngineeringGeoscience and ClimateProduction Skills: Fine-tuning, multimodal AI, evaluation and benchmarking, ethics and responsible AI, deployment and MLOpsPROVEN MATERIAL:This book originated from the graduate course "Generative AI for Science" at NC State's Data Science Academy. The content has been tested and refined through:Multiple graduate course offerings12+ hands-on AI/LLM training workshopsHundreds of students and researchersReal research collaborations in biology, chemistry, climate science, and materials"Generative AI enhances the scientific method—expanding hypotheses, sharpening experiments, revealing hidden patterns."Get the book. Clone the repo. Start discovering. Read more