History, Milestones, and AI Waves
From Symbolic AI to Deep Learning
Early AI
Early AI research emphasized symbolic reasoning, search, and expert systems. Researchers hoped that intelligence could be built by encoding human knowledge directly into formal rules.
AI Winters
Progress often slowed when systems failed to scale, data was limited, or expectations exceeded reality. These periods of disappointment are called AI winters and helped reshape research priorities.
Why the field revived
The rise of large datasets, cheaper computation, and better optimization made data-driven methods far more effective.
Key AI Eras
| Era | Main Idea | Common Limit |
|---|---|---|
| Symbolic AI | Rules and logic | Poor robustness |
| Expert Systems | Human-crafted knowledge bases | Hard to maintain |
| Machine Learning | Learn from data | Needs data and tuning |
| Deep Learning | Multi-layer neural networks | Compute and data hungry |
What was a major reason for the success of modern AI?
Large-scale data and compute made learning-based methods much more effective in many tasks.
Correct answer: More data and more computing power
What is an AI winter?
The term refers to reduced enthusiasm, investment, or progress in the field.
Correct answer: A period when AI progress and funding slow after exaggerated expectations are not met.
How AI Progress Typically Happens
-
1
Step 1: Identify a task that humans perform well.
-
2
Step 2: Define measurable success criteria.
-
3
Step 3: Build a baseline model or rule-based system.
-
4
Step 4: Improve performance using data, features, or architecture changes.
-
5
Step 5: Evaluate failures and iterate.
Expert systems were primarily built using:
Expert systems depended on human experts translating domain knowledge into rules.
Correct answer: Manually encoded knowledge and rules