Top 7 livros para aprender sobre machine learning e inteligência artificial

My Favorite Machine Learning and AI Books

Key Concepts: Data manipulation, machine learning models, deep learning, neural networks, rational agents, dynamic systems, control theory, TensorFlow, Scikit-learn, Keras, Pandas, NumPy, activation functions, loss functions, Transformers, diffusion models, reinforcement learning.

1. Python for Data Analysis by Wes McKinney

• Main Topic: Data manipulation and analysis using Python.
• Key Points:
  • Focuses on using Python for data manipulation, analysis, and preparation for machine learning modeling.
  • The author, Wes McKinney, is one of the creators of the Pandas library.
  • Covers the essential knowledge required for data analysis in machine learning.
• Technical Terms: Pandas (a Python data analysis library).

2. Machine Learning with Scikit-Learn, Keras & TensorFlow

• Main Topic: Practical machine learning with Scikit-learn, Keras, and TensorFlow.
• Key Points:
  • Focuses on finding patterns in data using machine learning models.
  • Includes practical projects for solving problems using TensorFlow.
  • The speaker replicated and adapted the examples to create portfolio projects and YouTube videos.
  • The TensorFlow examples covered approximately 80% of the material needed for the TensorFlow certification.
• Technical Terms: Scikit-learn (a Python machine learning library), Keras (a high-level neural networks API), TensorFlow (an open-source machine learning framework).

3. Neural Networks from Scratch with Python

• Main Topic: Understanding the inner workings of neural networks using Python and NumPy.
• Key Points:
  • Focuses on understanding how neural networks work, including parameters, activation functions, and loss functions.
  • Uses only Python and NumPy for implementation.
  • Includes colorful illustrations and animations (accessible via QR codes).
• Technical Terms: Neural network, activation functions, loss functions, NumPy (a Python library for numerical computing).

4. Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville

• Main Topic: Foundational concepts and techniques in deep learning.
• Key Points:
  • Provides a solid foundation in deep learning techniques and vocabulary.
  • Includes a review of linear algebra (Chapter 2 addresses gaps in matrix mathematics).
  • A free online version is available.
• Technical Terms: Deep learning, linear algebra, matrix mathematics.

5. Understand Deep Learning by Simon Prince

• Main Topic: Modern deep learning concepts and techniques.
• Key Points:
  • A more modern version of the Goodfellow book, covering similar topics but also including recent advancements.
  • Covers Transformers, diffusion models, and reinforcement learning in detail.
  • A free online version with examples is available.
• Technical Terms: Transformers, diffusion models, reinforcement learning.

6. Artificial Intelligence: A Modern Approach by Russell and Norvig

• Main Topic: General principles of rational agents and AI development.
• Key Points:
  • Focuses on the principles of rational agents and the components needed to develop them.
  • Defines an agent as an intelligent entity capable of making rational decisions.
  • Covers machine learning, deep learning, and reinforcement learning within the context of rational agents.
  • The speaker uses it primarily as a reference book.
• Technical Terms: Rational agent, machine learning, deep learning, reinforcement learning.

7. Data-Driven Science and Engineering: Machine Learning, Dynamical Systems, and Control

• Main Topic: Machine learning and data-based methods for engineering and physics.
• Key Points:
  • Focuses on machine learning for dynamic systems and control theory.
  • Includes chapters on deep learning and reinforcement learning in physics and engineering.
  • Provides examples in MATLAB and Python.
  • Each chapter has an accompanying video.
• Technical Terms: Dynamic systems, control theory, MATLAB.

Synthesis/Conclusion:

The speaker presents a curated list of books that have been instrumental in their machine learning and AI journey. The selection emphasizes a practical, code-first approach, supplemented by theoretical understanding as needed. The books cover a range of topics from data manipulation and analysis to deep learning, rational agents, and applications in science and engineering. The speaker highlights the importance of hands-on experience and adapting existing solutions to new problems. The list progresses from introductory texts to more advanced and specialized topics, reflecting the speaker's own learning path.