Deep Learning Crash Course placeholder cover

Deep Learning Crash Course

by Giovanni Volpe, Benjamin Midtvedt, Jesús Pineda, Henrik Klein Moberg, Harshith Bachimanchi, Joana B. Pereira, and Carlo Manzo
Spring 2025, 472 pp.
ISBN-13: 
9781718503922
Use coupon code PREORDER to get 25% off!

Deep Learning Crash Course starts from the basics to reach the most modern techniques and applications that are of great interest right now, and whose popularity will only grow in the future. It covers advanced topics such as generative models (the technology behind deep fakes), self-supervised learning, attention mechanisms (the technology behind ChatGPT), diffusion models (the technology behind text2image models such as DALL-E), graph neural networks (the technology behind AlphaFold), and deep reinforcement learning (the technology behind AlphaGo). These cutting-edge concepts and techniques address the current demands and trends in deep learning, giving you practical skills to tackle complex real-world problems.

Author Bio 

Giovanni Volpe is a Professor at the Physics Department of the University of Gothenburg in Sweden. His research interests include deep learning, brain connectivity, statistical mechanics, and soft matter. He has authored more than 200 articles and reviews on these topics. Moreover, he has co-authored two books, Optical Tweezers: Principles and Applications (Cambridge University Press, 2015) and Simulation of Complex Systems (IOP, 2021), and is currently co-editing the book Active Matter, which will appear in early 2024 (Springer Verlag). He has also developed several software packages for microscopy, deep learning, and brain connectivity.

Benjamin Midtvedt is a doctoral researcher that combines a solid grounding in physics with a keen interest in the potential of deep learning in life sciences. His background includes a Bachelor’s in Physics and a Master’s degree in Engineering Mathematics and Computer Science. Benjamin has made significant strides in the field of microscopy through deep learning. The unifying focus of his research has been the development of accessible and practical AI optimized to the needs of the user. He is also been the lead developer of several Python-based open-source deep learning frameworks.

Jesús Pineda is a doctoral researcher in physics interested in the intersection between
deep learning and computer vision. Jesús holds a Bachelor's degree in Mechatronics and a Master's in Electrical and Electronic engineering. He co-authored several articles in high-impact journals, focusing on the application of deep learning to unveil meaningful insights derived from microscopy data. Jesús is also a core developer of the deep learning software packages DeepTrack and Deeplay.

Henrik Klein Moberg is a Ph.D. candidate at Chalmers University of Technology, specializing in the integration of Artificial Intelligence with physical sciences. His academic background includes a Bachelor's degree in Physics and a Master’s degree in Complex Adaptive Systems. His research focuses on applying deep learning techniques to nanofluidic microscopy and nanophotonics, aiming to enhance the precision and efficiency of these technologies. He has also organized and spoken at numerous conferences related to AI and scientific data analysis.

Harshith Bachimanchi is a PhD student whose research combines holographic microscopy and deep learning to better understand marine microorganisms. His academic journey began with an integrated Bachelor's-Master's program in physics, focusing initially on experimental nonlinear optics. Since beginning his PhD in 2020, Harshith has applied his skills in experimental optics alongside deep learning techniques to track both biological and synthetic particles, enhancing our understanding of these complex systems. He has also developed simulations and tutorials demonstrating the practical applications of deep learning in microscopy. Moving forward, Harshith aims to continue blending experimental and computational approaches to solve complex challenges in biophysics.

Joana B. Pereira is an Associate Professor at Karolinska Institute in Sweden, where she focuses on investigating new biomarkers for neurodegenerative disorders, in particular Alzheimer’s disease. She has published over 90 articles in highly ranked journals including "Nature Aging" and "Nature Communications”, which have been featured several times by the press. Since 2020 she has been organizing an interdisciplinary conference called “Emerging Topics in Artificial Intelligence” held annually in San Diego, CA. She is also the scientific coordinator at Karolinska Institute of an innovative, trans-European Network of Excellence for brain research and technologies called NeurotechEU. In 2021, she won the De Leon prize for best neuroimaging article in Alzheimer’s disease.

Carlo Manzo is an Associate Professor at the University of Vic, Spain, where he leads the Quantitative Bioimaging Lab. His research is dedicated to the quantitative analysis of biophysical processes, merging advanced deep-learning techniques with state-of-the-art imaging technologies to achieve single-molecule sensitivity. His work primarily investigates the spatiotemporal organization and dynamics of cellular membrane components, focusing on their implications in health and disease. He has contributed to over 50 peer-reviewed articles and reviews in top-tier journals, such as “Nature Methods” and “Nature Machine Intelligence”. He is also a developer of several software packages and the founder of the Anomalous Diffusion (AnDi) challenge, an initiative that galvanizes the scientific community to refine methods for analyzing single-molecule trajectories. His contributions to the field of biophysics were recognized in 2017 when he was awarded the “E. Pérez Payá” prize by the Sociedad de Biofísica de España.

Table of contents 

Introduction
Chapter 1: Dense Neural Networks for Classification 
Chapter 2: Dense Neural Networks for Regression
Chapter 3: Convolutional Neural Networks for Image Analysis
Chapter 4: Encoders–Decoders for Latent Space Manipulation
Chapter 5: U-Nets for Image Transformation
Chapter 6: Self-Supervised Learning to Exploit Symmetries 

Chapter 7: Recurrent Neural Networks for Timeseries Analysis
Chapter 8: Attention and Transformers for Sequence Processing
Chapter 9: Generative Adversarial Networks for Image Synthesis
Chapter 10: Diffusion Models for Data Representation and Exploration
Chapter 11: Graph Neural Networks for Relational Data Analysis 
Chapter 12: Active Learning for Continuous Learning
Chapter 13: Reinforcement Learning for Strategy Optimization 
Chapter 14: Reservoir Computing for Predicting Chaos 
Conclusion and Outlook

The chapters in red are included in this Early Access PDF.