Problem Solving with Artificial Intelligence: A Hands-On Guide to AI Learning

This article introduces the online open-source book Problem Solving with Artificial Intelligence from the Hands-On Computer Science series. Adopting a hands-on teaching philosophy, it aims to address the disconnect between theory and practice in AI education, helping learners master core AI concepts and technologies through solving real-world problems, thus offering unique learning value.

Hands-On Computer Science Series

This series is centered on the core philosophy of 'learning by doing', with features including open-source and free access, online-first, code-driven, and project-oriented.

Publication Advantages of Problem Solving with Artificial Intelligence

As an online open-source book, it has advantages such as real-time updates, community contributions, multimedia support, and global accessibility, with a unique focus on solving real-world problems using AI.

Problem-Oriented Learning Path

Starting from real-world problems (e.g., predicting house prices, recognizing handwritten digits), it guides learners to explore solutions, implement code hands-on, and reflect on and expand principles.

Core Topic Coverage

Covers machine learning fundamentals (supervised/unsupervised learning, model evaluation, etc.), deep learning basics (neural networks, CNN/RNN, etc.), practical application areas (computer vision, NLP, etc.), and engineering practice skills (data preprocessing, model deployment, etc.).

Code Practice Design

Progressive complexity, runnability, modifiability, and completeness, using modern toolkits like PyTorch/TensorFlow.

Constructivist Learning Theory

Constructs knowledge through solving real-world problems. Code experiments provide interactive opportunities, and error debugging becomes part of the learning process.

Cognitive Load Management

Chunked learning, scaffolding support, and repeated reinforcement of core concepts.

Fromom Concrete to Abstract

Path: Concrete code → Run results → Pattern observation → Concept abstraction → Principle understanding.

Instant Feedback Loop

Modify code to see results instantly; error messages and visualization help with understanding.

Target Audience

• Beginner learners: Need basic programming (Python) and high school math; it is recommended to read in order and run code hands-on.
• Professionals: With software engineering experience, can quickly skim familiar content and focus on AI-specific methods.
• Educators/self-learners: Use as a textbook or self-study guide guide, extract examples to improve courses.

Tech Stack

• Language: Python
• Core libraries: Scikit-learn, PyTorch/TensorFlow, Pandas/NumPy, Matplotlib/Seaborn
• Environment: Jupyter Notebook, Google Colab

Advantages

Free access, community-driven, rapid iteration, transparent and open, customizability.

Challenges

Quality control, sustainability, consistency, discoverability.

Learning Recommendations

Build a solid foundation, practice code hands-on, build a project portfolio, participate in the community, and keep learning.

Future Directions

Enhance interactive content, multimedia support, enrich the practical project library, certification assessment, multilingual support.

Problem Solving with Artificial Intelligence represents a practice-oriented, open-source, and shared AI education paradigm. It lowers the learning threshold, bridges the gap between theory and practice, promotes self-directed learning, and provides shared resources for the community. It is an excellent starting point for AI learning, helping to cultivate AI talents and drive the democratization and responsible development of technology.