Understand
Modern cognitive computing is organized into three nested domains:
- Artificial Intelligence (AI): The overarching scientific discipline aimed at engineering computational systems capable of performing tasks that traditionally require human intelligence—such as visual perception, speech recognition, decision-making under uncertainty, and translation. Coined by John McCarthy in 1956.
- Machine Learning (ML): A core subset of AI where mathematical algorithms learn predictive statistical patterns directly from empirical training data without being explicitly hand-coded with deterministic rules. Categorized into:
- Supervised Learning: Trained on labeled input-output pairs (e.g., classification of benign vs malignant tumors; linear regression predicting house prices).
- Unsupervised Learning: Discovers hidden intrinsic patterns, clusters, or dimensionality reductions in unlabeled datasets (e.g., customer segmentation via K-Means).
- Reinforcement Learning: An autonomous agent learns optimal behavioral policies through trial-and-error interactions with an environment, maximizing cumulative rewards (e.g., AlphaGo, self-driving vehicle navigation).
- Deep Learning (DL): A specialized subset of ML based on Artificial Neural Networks (ANNs) with multiple hidden layers (deep architectures). Inspired by biological neural networks in the human brain, DL automatically learns hierarchical feature representations (e.g., raw pixels → edges → shapes → facial recognition) without manual feature engineering.
AI Sub-Disciplines
- Natural Language Processing (NLP): Algorithms enabling computers to comprehend, analyze, generate, and manipulate human natural languages (sentiment analysis, speech-to-text, machine translation, large language models).
- Computer Vision: Computational methods that extract high-level semantic understanding from digital images and video streams (object detection, facial recognition, medical imaging diagnostics).