AI vs ML vs Deep Learning

# Comparing Classical AI vs Traditional ML vs Deep Learning in Python
# ─────────────────────────────────────────────────────────────────────
 
# 1. CLASSICAL SYMBOLIC AI: Handcrafted deterministic if-else rules
def classify_sentiment_classical(text):
    positive_words = {"superb", "brilliant", "great", "excellent", "love"}
    negative_words = {"terrible", "awful", "horrible", "worst", "hate"}
 
    words = set(text.lower().split())
    pos_count = len(words.intersection(positive_words))
    neg_count = len(words.intersection(negative_words))
 
    if pos_count > neg_count:
        return "Positive"
    elif neg_count > pos_count:
        return "Negative"
    return "Neutral"
 
print(f"Classical AI Output: {classify_sentiment_classical('The movie was brilliant and superb!')}")
 
 
# 2. TRADITIONAL MACHINE LEARNING: Hand-extracted features + Logistic Regression
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.linear_model import LogisticRegression
 
train_texts = ["Great film love it", "Terrible waste of time", "Excellent acting", "Worst movie ever"]
train_labels = [1, 0, 1, 0] # 1 = Positive, 0 = Negative
 
# Step A: Human manual feature engineering (Bag of Words)
vectorizer = CountVectorizer()
X_features = vectorizer.fit_transform(train_texts)
 
# Step B: Statistical Classifier
ml_model = LogisticRegression()
ml_model.fit(X_features, train_labels)
 
test_sample = vectorizer.transform(["I love this masterpiece"])
ml_pred = ml_model.predict(test_sample)[0]
print(f"Traditional ML Prediction: {'Positive' if ml_pred == 1 else 'Negative'}")
 
 
# 3. DEEP LEARNING: End-to-End Neural Representation (PyTorch / Keras conceptual)
# The network takes raw token embeddings and learns hierarchical attention weights
# Model: Embedding Layer (128d) -> 4x Transformer Blocks -> Dense Classification Head
print(f"Deep Learning Pipeline: [Raw Text] -> [Token Embeddings] -> [Self-Attention Layers] -> [Sentiment Score: 0.985]")