Chapter 14 - Analyzing Adversarial Performance - The Deep Learning Architect's Handbook
Chapter 14 - Analyzing Adversarial Performance - The Deep Learning Architect's Handbook
In this part of the book, you will dive into the exciting realm of deploying, monitoring, and governing deep learning models in production, drawing
parallels with MLOps and DevOps. This part will provide you with a comprehensive understanding of the essential components required to ensure the
success and impact of your deep learning models in production with real-world utilization.
Throughout the chapters in this part, we’ll explore the various aspects of deploying deep learning models in production, touching upon important
considerations such as hardware infrastructure, model packaging, and user interfaces. We’ll also delve into the three fundamental pillars of model
governance, which are model utilization, model monitoring, and model maintenance. You’ll learn about the concept of drift and its impact on the
performance of deployed deep learning models over time, as well as strategies to handle drift e ectively. We’ll also discuss the benefits of AI
platforms such as DataRobot, which streamline the complex stages of the machine learning life cycle and accelerate the creation, training,
deployment, and governance of intricate deep learning models.
As a bonus, building upon the foundational transformer method from Chapter 6, Understanding Neural Network Transformers, we’ll delve into Large
Language Model (LLM ) solutions, which are revolutionizing various domains. You’ll learn about architecting LLM solutions and building autonomous
agents, equipping you with the knowledge to harness their potential.
End of Chapter 17
Previous Chapter Next Chapter