This is an on-point book on Artificial Intelligence (AI) and is meant for technical audience. You will not see lengthy introductions or digressions and instead, we will focus on the crux of AI algorithms, models, and related machinery. We will be assuming good knowledge of Python and linear algebra.

### Linear Regression

This is likely the simplest form of statistical AI. The goal is to approximate date with a line.

## Implementation (Linear Regression)

```
import numpy as np
from dataclasses import dataclass
@dataclass
class LinearRegression:
"""Linear Regression."""
epochs: int
learning_rate: float
logging: bool
def fit(self, features: np.ndarray, labels: np.ndarray) -> None:
"""Fits the Linear Regression model."""
num_samples, num_features = features.shape
self.weights, self.bias = np.zeros(num_features), 0
for epoch in range(self.epochs):
residuals = labels - self.predict(features)
d_weights = -2 / num_samples * features.T.dot(residuals)
d_bias = -2 / num_samples * residuals.sum()
self.weights -= self.learning_rate * d_weights
self.bias -= self.learning_rate * d_bias
if self.logging:
print(f"MSE Loss [{epoch}]: {(residuals ** 2).mean():.3f}")
def predict(self, features: np.ndarray) -> np.ndarray:
"""Performs inference using the given features."""
return features.dot(self.weights) + self.bias
```

### Logistic Regression

TODO

### K-Nearest Neighbors (k-NN)

TODO

### K-Means Clustering (KMeans)

TODO