mirror of
https://github.com/ArthurDanjou/ArtStudies.git
synced 2026-03-16 07:10:13 +01:00
Arthur DANJOU
a8e15d66f7
- Updated execution counts for code cells to reflect changes. - Enhanced output formatting for matched FAQs, including clearer question and answer presentation. - Adjusted the similarity threshold in the `match_faq` function to allow for looser matching (default set to 0.5). - Improved documentation for the threshold parameter to clarify its usage.
2082 lines
424 KiB
Plaintext
2082 lines
424 KiB
Plaintext
{
|
||
"cells": [
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "3ZaYIfQdC8pO"
|
||
},
|
||
"source": [
|
||
"# 🎯 Discovering Embeddings - Text as Vectors!\n",
|
||
"\n",
|
||
"## 🤔 What Are Embeddings?\n",
|
||
"\n",
|
||
"Imagine teaching a computer to understand language by converting words and sentences into **mathematical vectors**!\n",
|
||
"\n",
|
||
"**The Magic:** \n",
|
||
"- 📝 \"The cat is sleeping\" → `[0.23, -0.45, 0.87, ...]` (384 numbers!)\n",
|
||
"- 🧠 Similar meanings → Similar vectors\n",
|
||
"- 🎯 Can measure semantic similarity mathematically\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 🌟 Why Embeddings Are Revolutionary:\n",
|
||
"\n",
|
||
"| Traditional Approach | Embeddings Approach |\n",
|
||
"|---------------------|---------------------|\n",
|
||
"| 🔤 Words as strings | 🔢 Words as vectors |\n",
|
||
"| ❌ Can't compare meanings | ✅ Can measure similarity |\n",
|
||
"| 🚫 No semantic understanding | 🧠 Captures meaning & context |\n",
|
||
"| 📏 Fixed vocabulary | 🌐 Handles any text |\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 🎓 What You'll Learn:\n",
|
||
"\n",
|
||
"1. 🤖 **Load pre-trained models** - Use powerful sentence transformers\n",
|
||
"2. 🔄 **Convert text to vectors** - See embeddings in action\n",
|
||
"3. 📊 **Measure similarity** - Compare semantic meaning\n",
|
||
"4. 🎨 **Visualize embeddings** - See language in 2D space\n",
|
||
"5. 💡 **Real applications** - Discover practical use cases\n",
|
||
"\n",
|
||
"Let's dive in! 🚀"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "BQs7EF55E98S"
|
||
},
|
||
"source": [
|
||
"### 📦 The Model We'll Use:\n",
|
||
"\n",
|
||
"**`text-embedding-3-small`**\n",
|
||
"- Supports multiple languages (English, French, Spanish, etc.)\n",
|
||
"- Creates 1536-dimensional vectors\n",
|
||
"- Optimized for semantic similarity tasks\n",
|
||
"- Perfect balance of speed and accuracy\n",
|
||
"\n",
|
||
"\n",
|
||
"- 🌍 **Multilingual:** Works in 50+ languages\n",
|
||
"- ⚡ **Fast:** Encodes sentences in seconds\n",
|
||
"- 🎯 **Accurate:** State-of-the-art semantic understanding\n",
|
||
"\n",
|
||
"Let's load it!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 1,
|
||
"metadata": {
|
||
"id": "R88Cd1WAp5Br"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"✅ Model loaded successfully!\n",
|
||
" Model can handle: Multiple languages\n",
|
||
" Embedding dimension: 1536\n",
|
||
" Ready to convert text to vectors! 🚀\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"import os\n",
|
||
"\n",
|
||
"from dotenv import load_dotenv\n",
|
||
"from langchain_openai import OpenAIEmbeddings\n",
|
||
"\n",
|
||
"# Load the pre-trained model\n",
|
||
"load_dotenv()\n",
|
||
"os.environ[\"OPENAI_API_KEY\"] = (\n",
|
||
" \"\"\n",
|
||
")\n",
|
||
"\n",
|
||
"# Initialize the embedding model\n",
|
||
"\n",
|
||
"embedding_model = OpenAIEmbeddings(\n",
|
||
" model=\"text-embedding-3-small\",\n",
|
||
" # With the `text-embedding-3` class\n",
|
||
" # of models, you can specify the size\n",
|
||
" # of the embeddings you want returned.\n",
|
||
" # dimensions=1024 # noqa: ERA001\n",
|
||
")\n",
|
||
"\n",
|
||
"\n",
|
||
"print(\"✅ Model loaded successfully!\")\n",
|
||
"print(\" Model can handle: Multiple languages\")\n",
|
||
"print(\" Embedding dimension: 1536\")\n",
|
||
"print(\" Ready to convert text to vectors! 🚀\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 2,
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"from langchain_core.embeddings import Embeddings\n",
|
||
"\n",
|
||
"\n",
|
||
"# 🔄 Creating Your First Embedding\n",
|
||
"def embed(text: str) -> list[float]:\n",
|
||
" \"\"\"Convert a text string into a vector embedding using the pre-trained model.\"\"\"\n",
|
||
" return embedding_model.embed_query(text)\n",
|
||
"\n",
|
||
"\n",
|
||
"\n",
|
||
"class EmbeddingsTest(Embeddings):\n",
|
||
" \"\"\"Test class to implement the Embeddings interface using the OpenAIEmbeddings model.\"\"\"\n",
|
||
"\n",
|
||
" def embed_query(self, text: str) -> list[float]:\n",
|
||
" \"\"\"Embed a single query string into a vector.\"\"\"\n",
|
||
" return embedding_model.embed_query(text)\n",
|
||
"\n",
|
||
" def embed_documents(self, texts: list[str]) -> list[list[float]]:\n",
|
||
" \"\"\"Embed a list of documents into vectors.\"\"\"\n",
|
||
" return embedding_model.embed_documents(texts)\n",
|
||
"\n",
|
||
"\n",
|
||
"e = EmbeddingsTest()\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 3,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/plain": [
|
||
"langchain_core.embeddings.embeddings.Embeddings"
|
||
]
|
||
},
|
||
"execution_count": 3,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"from langchain_core.embeddings import Embeddings\n",
|
||
"\n",
|
||
"Embeddings\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 4,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/plain": [
|
||
"langchain_openai.embeddings.base.OpenAIEmbeddings"
|
||
]
|
||
},
|
||
"execution_count": 4,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"OpenAIEmbeddings\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"# 🔄 Creating Your First Embedding\n",
|
||
"\n",
|
||
"Let's convert a simple sentence into a vector!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 5,
|
||
"metadata": {
|
||
"id": "7qDnnbFn62Vk"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"📝 Input sentence: 'The weather is beautiful'\n",
|
||
"\n",
|
||
"🔢 Embedding (vector representation):\n",
|
||
" First 10 values: [0.010934686288237572, -0.027822494506835938, -0.011565890163183212, 0.010940873995423317, 0.026065023615956306, -0.01712915115058422, 0.0015609929105266929, -0.011058451607823372, -0.025891752913594246, -0.011293605901300907]\n",
|
||
" ...\n",
|
||
" Last 10 values: [0.026461074128746986, 0.011702032759785652, 0.0010303481249138713, -0.006079982500523329, 0.020446067675948143, 0.057130176573991776, -0.004919680301100016, -0.034307811409235, 0.0036139539442956448, 0.013304796069860458]\n",
|
||
"\n",
|
||
"📊 Each number captures different semantic aspects!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Our input sentence\n",
|
||
"input_str = \"The weather is beautiful\"\n",
|
||
"\n",
|
||
"# Convert to embedding (vector)\n",
|
||
"embedding = embedding_model.embed_query(input_str)\n",
|
||
"\n",
|
||
"print(f\"📝 Input sentence: '{input_str}'\")\n",
|
||
"print(\"\\n🔢 Embedding (vector representation):\")\n",
|
||
"print(f\" First 10 values: {embedding[:10]}\")\n",
|
||
"print(\" ...\")\n",
|
||
"print(f\" Last 10 values: {embedding[-10:]}\")\n",
|
||
"print(\"\\n📊 Each number captures different semantic aspects!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 6,
|
||
"metadata": {
|
||
"id": "No35paV7687c"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"📏 Embedding dimension: 1536\n",
|
||
"\n",
|
||
"💡 This means our sentence is represented by 1536 numbers!\n",
|
||
" Each dimension captures different aspects of meaning:\n",
|
||
" • Some dimensions might encode 'weather' concepts\n",
|
||
" • Others might encode positive sentiment\n",
|
||
" • Together, they form a unique 'fingerprint' of the sentence\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Check the dimension\n",
|
||
"dimension = len(embedding)\n",
|
||
"\n",
|
||
"print(f\"📏 Embedding dimension: {dimension}\")\n",
|
||
"print(f\"\\n💡 This means our sentence is represented by {dimension} numbers!\")\n",
|
||
"print(\" Each dimension captures different aspects of meaning:\")\n",
|
||
"print(\" • Some dimensions might encode 'weather' concepts\")\n",
|
||
"print(\" • Others might encode positive sentiment\")\n",
|
||
"print(\" • Together, they form a unique 'fingerprint' of the sentence\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"# 📊 Measuring Similarity\n",
|
||
"\n",
|
||
"Now for the **magic** - let's compare multiple sentences and see which ones are semantically similar!\n",
|
||
"\n",
|
||
"## 🧪 The Experiment:\n",
|
||
"\n",
|
||
"We'll compare three sentences:\n",
|
||
"1. \"The weather is lovely today\" 🌞\n",
|
||
"2. \"It's so sunny outside!\" ☀️\n",
|
||
"3. \"He drove to the stadium\" 🚗\n",
|
||
"\n",
|
||
"**Prediction:** Sentences 1 & 2 should be similar (both about weather), while 3 should be different.\n",
|
||
"\n",
|
||
"Let's test this!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 7,
|
||
"metadata": {
|
||
"id": "QLgxvweMideU"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"🔄 Converting sentences to embeddings...\n",
|
||
"📊 Computing similarity matrix...\n",
|
||
"\n",
|
||
"✅ Complete! Let's visualize the results...\n",
|
||
"\n",
|
||
"📐 Similarity matrix shape: (3, 3)\n",
|
||
" (3 sentences x 3 sentences = 9 similarity scores)\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"import numpy as np\n",
|
||
"\n",
|
||
"# Our test sentences\n",
|
||
"inputs = [\n",
|
||
" \"The weather is lovely today.\",\n",
|
||
" \"It's so sunny outside!\",\n",
|
||
" \"He drove to the stadium.\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"print(\"🔄 Converting sentences to embeddings...\")\n",
|
||
"embeddings = embedding_model.embed_documents(inputs)\n",
|
||
"\n",
|
||
"print(\"📊 Computing similarity matrix...\")\n",
|
||
"embeddings = np.array(embeddings)\n",
|
||
"similarities = np.dot(embeddings, embeddings.T)\n",
|
||
"\n",
|
||
"print(\"\\n✅ Complete! Let's visualize the results...\")\n",
|
||
"print(f\"\\n📐 Similarity matrix shape: {similarities.shape}\")\n",
|
||
"print(\" (3 sentences x 3 sentences = 9 similarity scores)\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"## 🎨 Creating a Visualization Function\n",
|
||
"\n",
|
||
"Let's make a beautiful heatmap to visualize similarities!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 8,
|
||
"metadata": {
|
||
"id": "C8PvTRf1i_i5"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"2.3.5\n",
|
||
"✅ Visualization function created!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"import matplotlib.pyplot as plt\n",
|
||
"import numpy as np\n",
|
||
"import seaborn as sns\n",
|
||
"\n",
|
||
"print(np.__version__)\n",
|
||
"\n",
|
||
"def compute_and_display_similarities(inputs: list[str], title: str = \"Semantic Similarity Heatmap\") -> None:\n",
|
||
" \"\"\"Compute and display a beautiful heatmap of sentence similarities.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" inputs: List of sentences to compare\n",
|
||
" title: Title for the plot\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" # Compute embeddings and similarities\n",
|
||
" embeddings = embedding_model.embed_documents(inputs)\n",
|
||
" embeddings = np.array(embeddings)\n",
|
||
" similarities = np.dot(embeddings, embeddings.T)\n",
|
||
"\n",
|
||
" # Create figure\n",
|
||
" plt.figure(figsize=(12, 10))\n",
|
||
"\n",
|
||
" # Create heatmap\n",
|
||
" _ax = sns.heatmap(\n",
|
||
" similarities,\n",
|
||
" annot=True, # Show values\n",
|
||
" fmt=\".3f\", # 3 decimal places\n",
|
||
" xticklabels=inputs,\n",
|
||
" yticklabels=inputs,\n",
|
||
" cmap=\"YlGnBu\", # Yellow-Green-Blue colormap\n",
|
||
" cbar_kws={\"label\": \"Similarity Score\"},\n",
|
||
" vmin=0, # Minimum value\n",
|
||
" vmax=1, # Maximum value\n",
|
||
" square=True, # Square cells\n",
|
||
" linewidths=0.5, # Grid lines\n",
|
||
" linecolor=\"gray\",\n",
|
||
" )\n",
|
||
"\n",
|
||
" # Styling\n",
|
||
" plt.title(title, fontsize=16, fontweight=\"bold\", pad=20)\n",
|
||
" plt.xlabel(\"Sentences\", fontsize=12, fontweight=\"bold\")\n",
|
||
" plt.ylabel(\"Sentences\", fontsize=12, fontweight=\"bold\")\n",
|
||
" plt.xticks(rotation=45, ha=\"right\")\n",
|
||
" plt.yticks(rotation=0)\n",
|
||
" plt.tight_layout()\n",
|
||
" plt.show()\n",
|
||
"\n",
|
||
" # Print interpretation\n",
|
||
" print(\"\\n📖 How to read this heatmap:\")\n",
|
||
" print(\" • 1.0 (dark blue) = Perfect similarity (same sentence)\")\n",
|
||
" print(\" • 0.8-0.9 = Very similar meaning\")\n",
|
||
" print(\" • 0.5-0.7 = Somewhat related\")\n",
|
||
" print(\" • 0.0-0.4 = Different meanings\")\n",
|
||
" print(\" • 0.0 (light yellow) = Completely different\")\n",
|
||
"\n",
|
||
"\n",
|
||
"print(\"✅ Visualization function created!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 9,
|
||
"metadata": {
|
||
"id": "am4JPM-ojgng"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"🎨 Generating similarity heatmap...\n",
|
||
"\n"
|
||
]
|
||
},
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/var/folders/tp/_ld5_pzs6nx6mv1pbjhq1l740000gn/T/ipykernel_24511/1064958240.py:45: UserWarning: Glyph 127775 (\\N{GLOWING STAR}) missing from font(s) DejaVu Sans.\n",
|
||
" plt.tight_layout()\n",
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/IPython/core/pylabtools.py:170: UserWarning: Glyph 127775 (\\N{GLOWING STAR}) missing from font(s) DejaVu Sans.\n",
|
||
" fig.canvas.print_figure(bytes_io, **kw)\n"
|
||
]
|
||
},
|
||
{
|
||
"data": {
|
||
"image/png": "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",
|
||
"text/plain": [
|
||
"<Figure size 1200x1000 with 2 Axes>"
|
||
]
|
||
},
|
||
"metadata": {},
|
||
"output_type": "display_data"
|
||
},
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"📖 How to read this heatmap:\n",
|
||
" • 1.0 (dark blue) = Perfect similarity (same sentence)\n",
|
||
" • 0.8-0.9 = Very similar meaning\n",
|
||
" • 0.5-0.7 = Somewhat related\n",
|
||
" • 0.0-0.4 = Different meanings\n",
|
||
" • 0.0 (light yellow) = Completely different\n",
|
||
"\n",
|
||
"🔍 Observations:\n",
|
||
" ✅ Sentences 1 & 2 have HIGH similarity (both about weather)\n",
|
||
" ❌ Sentence 3 has LOW similarity with others (different topic)\n",
|
||
" 🎯 The model understands semantic meaning, not just word matching!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Test with our sentences\n",
|
||
"inputs = [\n",
|
||
" \"The weather is lovely today\",\n",
|
||
" \"It's so sunny outside!\",\n",
|
||
" \"He drove to the stadium.\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"print(\"🎨 Generating similarity heatmap...\\n\")\n",
|
||
"compute_and_display_similarities(inputs, \"🌟 Semantic Similarity Between Sentences\")\n",
|
||
"\n",
|
||
"print(\"\\n🔍 Observations:\")\n",
|
||
"print(\" ✅ Sentences 1 & 2 have HIGH similarity (both about weather)\")\n",
|
||
"print(\" ❌ Sentence 3 has LOW similarity with others (different topic)\")\n",
|
||
"print(\" 🎯 The model understands semantic meaning, not just word matching!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"### 🌍 Multilingual Magic!\n",
|
||
"\n",
|
||
"Our model works across languages! Let's test with French and English:"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 10,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"🌍 Testing multilingual understanding...\n",
|
||
"\n"
|
||
]
|
||
},
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/var/folders/tp/_ld5_pzs6nx6mv1pbjhq1l740000gn/T/ipykernel_24511/1064958240.py:45: UserWarning: Glyph 127760 (\\N{GLOBE WITH MERIDIANS}) missing from font(s) DejaVu Sans.\n",
|
||
" plt.tight_layout()\n",
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/IPython/core/pylabtools.py:170: UserWarning: Glyph 127760 (\\N{GLOBE WITH MERIDIANS}) missing from font(s) DejaVu Sans.\n",
|
||
" fig.canvas.print_figure(bytes_io, **kw)\n"
|
||
]
|
||
},
|
||
{
|
||
"data": {
|
||
"image/png": "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",
|
||
"text/plain": [
|
||
"<Figure size 1200x1000 with 2 Axes>"
|
||
]
|
||
},
|
||
"metadata": {},
|
||
"output_type": "display_data"
|
||
},
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"📖 How to read this heatmap:\n",
|
||
" • 1.0 (dark blue) = Perfect similarity (same sentence)\n",
|
||
" • 0.8-0.9 = Very similar meaning\n",
|
||
" • 0.5-0.7 = Somewhat related\n",
|
||
" • 0.0-0.4 = Different meanings\n",
|
||
" • 0.0 (light yellow) = Completely different\n",
|
||
"\n",
|
||
"🤯 Amazing! The model recognizes that:\n",
|
||
" • French 'Il fait beau' ≈ English 'sunny outside'\n",
|
||
" • Despite different languages, it captures the SAME meaning!\n",
|
||
" • This is the power of multilingual embeddings! 🚀\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Mix French and English!\n",
|
||
"multilingual_inputs = [\n",
|
||
" \"Il fait beau aujourd'hui\", # French: \"The weather is nice today\"\n",
|
||
" \"It's so sunny outside!\", # English\n",
|
||
" \"He drove to the stadium.\", # English (different topic)\n",
|
||
"]\n",
|
||
"\n",
|
||
"print(\"🌍 Testing multilingual understanding...\\n\")\n",
|
||
"compute_and_display_similarities(\n",
|
||
" multilingual_inputs, \"🌐 Cross-Language Semantic Similarity\",\n",
|
||
")\n",
|
||
"\n",
|
||
"print(\"\\n🤯 Amazing! The model recognizes that:\")\n",
|
||
"print(\" • French 'Il fait beau' ≈ English 'sunny outside'\")\n",
|
||
"print(\" • Despite different languages, it captures the SAME meaning!\")\n",
|
||
"print(\" • This is the power of multilingual embeddings! 🚀\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "PpUEegq2Hd4s"
|
||
},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"# 🧪 Hands-On Exercise: Understanding Similarity Metrics\n",
|
||
"\n",
|
||
"## 🎯 Goal\n",
|
||
"\n",
|
||
"Explore how **cosine similarity** and **Euclidean distance** behave with embeddings.\n",
|
||
"\n",
|
||
"### 📐 Two Ways to Compare Vectors:\n",
|
||
"\n",
|
||
"| Metric | Formula | What it measures |\n",
|
||
"|--------|---------|------------------|\n",
|
||
"| **Cosine Similarity** | $\\frac{u \\cdot v}{\\|u\\| \\|v\\|}$ | Angle between vectors (0-1) |\n",
|
||
"| **Euclidean Distance** | $\\|u - v\\|$ | Straight-line distance |\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 🤔 Why Does This Matter?\n",
|
||
"\n",
|
||
"**Cosine similarity** is preferred for text because:\n",
|
||
"- 🎯 Focuses on **direction** (meaning) not magnitude (length)\n",
|
||
"- 📏 Normalized to 0-1 range (easy to interpret)\n",
|
||
"- 🎨 More robust to text length variations\n",
|
||
"\n",
|
||
"Let's see this in action!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "NqWKiIJ8GOC7"
|
||
},
|
||
"source": [
|
||
"## 🔧 Implementing Similarity Functions"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 11,
|
||
"metadata": {
|
||
"id": "ibFv2FWOGgU4"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"✅ Similarity functions defined!\n",
|
||
"\n",
|
||
"📊 Available metrics:\n",
|
||
" 1. cosine_similarity(u, v) → 0 to 1 (higher = more similar)\n",
|
||
" 2. euclidean_distance(u, v) → 0 to ∞ (lower = more similar)\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"import numpy as np\n",
|
||
"\n",
|
||
"\n",
|
||
"def cosine_similarity(u: list[float] | np.ndarray, v: list[float] | np.ndarray) -> float:\n",
|
||
" \"\"\"Calculate cosine similarity between two vectors.\n",
|
||
"\n",
|
||
" Returns a value between 0 and 1:\n",
|
||
" - 1.0 = vectors point in same direction (very similar)\n",
|
||
" - 0.0 = vectors are perpendicular (unrelated)\n",
|
||
" \"\"\"\n",
|
||
" u, v = np.array(u), np.array(v)\n",
|
||
" # Normalize vectors to unit length\n",
|
||
" u_norm = u / np.linalg.norm(u)\n",
|
||
" v_norm = v / np.linalg.norm(v)\n",
|
||
" # Dot product of normalized vectors\n",
|
||
" return float(u_norm @ v_norm)\n",
|
||
"\n",
|
||
"\n",
|
||
"def euclidean_distance(u: list[float] | np.ndarray, v: list[float] | np.ndarray) -> float:\n",
|
||
" \"\"\"Calculate Euclidean distance between two vectors.\n",
|
||
"\n",
|
||
" Returns the straight-line distance:\n",
|
||
" - 0.0 = identical vectors\n",
|
||
" - Larger values = more different\n",
|
||
" \"\"\"\n",
|
||
" u, v = np.array(u), np.array(v)\n",
|
||
" return float(np.linalg.norm(u - v))\n",
|
||
"\n",
|
||
"\n",
|
||
"print(\"✅ Similarity functions defined!\")\n",
|
||
"print(\"\\n📊 Available metrics:\")\n",
|
||
"print(\" 1. cosine_similarity(u, v) → 0 to 1 (higher = more similar)\")\n",
|
||
"print(\" 2. euclidean_distance(u, v) → 0 to ∞ (lower = more similar)\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"## 🐱 Test Sentences: The Cat Family\n",
|
||
"\n",
|
||
"We'll compare three sentences of increasing length, all about cats:"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 12,
|
||
"metadata": {
|
||
"id": "44YoBMZMh7Ro"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"🐱 Our test sentences:\n",
|
||
"\n",
|
||
"1️⃣ SHORT: 'the cat is a feline'\n",
|
||
" Length: 19 characters\n",
|
||
"\n",
|
||
"2️⃣ MEDIUM: 'the cat is a feline, the cat is a feline'\n",
|
||
" Length: 40 characters\n",
|
||
"\n",
|
||
"3️⃣ LONG: 'the cat is a feline, the cat is a feline, the cat is a feline, the cat is a feline'\n",
|
||
" Length: 82 characters\n",
|
||
"\n",
|
||
"💡 All three sentences are about cats, but with different lengths!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Three sentences about cats, with increasing detail\n",
|
||
"phrase_chat_1 = \"the cat is a feline\"\n",
|
||
"phrase_chat_2 = \"the cat is a feline, the cat is a feline\"\n",
|
||
"phrase_chat_3 = (\n",
|
||
" \"the cat is a feline, the cat is a feline, the cat is a feline, the cat is a feline\"\n",
|
||
")\n",
|
||
"\n",
|
||
"print(\"🐱 Our test sentences:\\n\")\n",
|
||
"print(f\"1️⃣ SHORT: '{phrase_chat_1}'\")\n",
|
||
"print(f\" Length: {len(phrase_chat_1)} characters\\n\")\n",
|
||
"\n",
|
||
"print(f\"2️⃣ MEDIUM: '{phrase_chat_2}'\")\n",
|
||
"print(f\" Length: {len(phrase_chat_2)} characters\\n\")\n",
|
||
"\n",
|
||
"print(f\"3️⃣ LONG: '{phrase_chat_3}'\")\n",
|
||
"print(f\" Length: {len(phrase_chat_3)} characters\\n\")\n",
|
||
"\n",
|
||
"print(\"💡 All three sentences are about cats, but with different lengths!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 13,
|
||
"metadata": {
|
||
"id": "UZWhMtD8uWdK"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"🔄 Converting sentences to embeddings...\n",
|
||
"\n",
|
||
"✅ Embeddings created!\n",
|
||
" Each is a 1536-dimensional vector\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Generate embeddings\n",
|
||
"print(\"🔄 Converting sentences to embeddings...\\n\")\n",
|
||
"\n",
|
||
"phrase_chat_1_emb = embedding_model.embed_query(phrase_chat_1)\n",
|
||
"phrase_chat_2_emb = embedding_model.embed_query(phrase_chat_2)\n",
|
||
"phrase_chat_3_emb = embedding_model.embed_query(phrase_chat_3)\n",
|
||
"\n",
|
||
"print(\"✅ Embeddings created!\")\n",
|
||
"print(f\" Each is a {len(phrase_chat_1_emb)}-dimensional vector\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"## 📏 Task 1: Calculate Euclidean Distance\n",
|
||
"\n",
|
||
"Now let's compare using **Euclidean distance**.\n",
|
||
"\n",
|
||
"**What to expect:** Lower values mean more similar!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 14,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"📏 Euclidean Distance Results:\n",
|
||
"\n",
|
||
"============================================================\n",
|
||
"📐 Short vs Long: 0.5506\n",
|
||
"📐 Short vs Medium: 0.4474\n",
|
||
"📐 Medium vs Long: 0.3820\n",
|
||
"============================================================\n",
|
||
"\n",
|
||
"🔍 Observations:\n",
|
||
" • Distances INCREASE with text length difference\n",
|
||
" • Short vs Long has LARGEST distance: 0.5506\n",
|
||
" • Medium vs Long has SMALLEST distance: 0.3820\n",
|
||
"\n",
|
||
"⚠️ Problem: Euclidean distance is affected by vector magnitude!\n",
|
||
" Longer texts → larger vectors → larger distances\n",
|
||
" Even if the MEANING is similar!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Calculate all pairwise Euclidean distances\n",
|
||
"print(\"📏 Euclidean Distance Results:\\n\")\n",
|
||
"print(\"=\" * 60)\n",
|
||
"\n",
|
||
"\n",
|
||
"# Sentence 1 vs Sentence 3\n",
|
||
"dist_1_3 = euclidean_distance(phrase_chat_1_emb, phrase_chat_3_emb)\n",
|
||
"print(f\"📐 Short vs Long: {dist_1_3:.4f}\")\n",
|
||
"\n",
|
||
"dist_1_2 = euclidean_distance(phrase_chat_1_emb, phrase_chat_2_emb)\n",
|
||
"print(f\"📐 Short vs Medium: {dist_1_2:.4f}\")\n",
|
||
"\n",
|
||
"# Sentence 2 vs Sentence 3\n",
|
||
"dist_2_3 = euclidean_distance(phrase_chat_2_emb, phrase_chat_3_emb)\n",
|
||
"print(f\"📐 Medium vs Long: {dist_2_3:.4f}\")\n",
|
||
"\n",
|
||
"print(\"=\" * 60)\n",
|
||
"\n",
|
||
"# Analysis\n",
|
||
"print(\"\\n🔍 Observations:\")\n",
|
||
"print(\" • Distances INCREASE with text length difference\")\n",
|
||
"print(f\" • Short vs Long has LARGEST distance: {dist_1_3:.4f}\")\n",
|
||
"print(f\" • Medium vs Long has SMALLEST distance: {dist_2_3:.4f}\")\n",
|
||
"print(\"\\n⚠️ Problem: Euclidean distance is affected by vector magnitude!\")\n",
|
||
"print(\" Longer texts → larger vectors → larger distances\")\n",
|
||
"print(\" Even if the MEANING is similar!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "O179uX3OGi_9"
|
||
},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"## 📊 Task 2: Calculate Cosine Similarity\n",
|
||
"\n",
|
||
"Compare each pair of sentences using **cosine similarity**.\n",
|
||
"\n",
|
||
"**What to expect:** Values close to 1 mean very similar meanings!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 15,
|
||
"metadata": {
|
||
"id": "-hyq4pfMGtS8"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"📊 Cosine Similarity Results:\n",
|
||
"\n",
|
||
"============================================================\n",
|
||
"🐱 Short vs Medium: 0.8999\n",
|
||
"🐱 Short vs Long: 0.8484\n",
|
||
"🐱 Medium vs Long: 0.9270\n",
|
||
"============================================================\n",
|
||
"\n",
|
||
"🔍 Observations:\n",
|
||
" • Average similarity: 0.8918\n",
|
||
" • All values are HIGH (close to 1.0)\n",
|
||
" • The model recognizes they're all about CATS! 🐱\n",
|
||
" • Text length doesn't drastically affect similarity\n",
|
||
"\n",
|
||
"💡 Cosine similarity focuses on MEANING, not LENGTH!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Calculate all pairwise cosine similarities\n",
|
||
"print(\"📊 Cosine Similarity Results:\\n\")\n",
|
||
"print(\"=\" * 60)\n",
|
||
"\n",
|
||
"# Sentence 1 vs Sentence 2\n",
|
||
"sim_1_2 = cosine_similarity(phrase_chat_1_emb, phrase_chat_2_emb)\n",
|
||
"print(f\"🐱 Short vs Medium: {sim_1_2:.4f}\")\n",
|
||
"\n",
|
||
"# Sentence 1 vs Sentence 3\n",
|
||
"sim_1_3 = cosine_similarity(phrase_chat_1_emb, phrase_chat_3_emb)\n",
|
||
"print(f\"🐱 Short vs Long: {sim_1_3:.4f}\")\n",
|
||
"\n",
|
||
"# Sentence 2 vs Sentence 3\n",
|
||
"sim_2_3 = cosine_similarity(phrase_chat_2_emb, phrase_chat_3_emb)\n",
|
||
"print(f\"🐱 Medium vs Long: {sim_2_3:.4f}\")\n",
|
||
"\n",
|
||
"print(\"=\" * 60)\n",
|
||
"\n",
|
||
"# Analysis\n",
|
||
"print(\"\\n🔍 Observations:\")\n",
|
||
"avg_similarity = (sim_1_2 + sim_1_3 + sim_2_3) / 3\n",
|
||
"print(f\" • Average similarity: {avg_similarity:.4f}\")\n",
|
||
"print(\" • All values are HIGH (close to 1.0)\")\n",
|
||
"print(\" • The model recognizes they're all about CATS! 🐱\")\n",
|
||
"print(\" • Text length doesn't drastically affect similarity\")\n",
|
||
"print(\"\\n💡 Cosine similarity focuses on MEANING, not LENGTH!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "E-xtDp93HNDX"
|
||
},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"## 🤔 Task 3: Why Cosine Similarity?\n",
|
||
"\n",
|
||
"### 💡 Key Insight:\n",
|
||
"\n",
|
||
"Let's compare the two metrics side by side:"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 16,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"📊 Metric Comparison:\n",
|
||
"\n",
|
||
" Comparison Cosine Similarity Euclidean Distance\n",
|
||
"Short vs Medium 0.8999 0.4474\n",
|
||
" Short vs Long 0.8484 0.5506\n",
|
||
" Medium vs Long 0.9270 0.3820\n",
|
||
"\n",
|
||
"======================================================================\n",
|
||
"🎯 CONCLUSION: Why Cosine Similarity Wins for Text\n",
|
||
"======================================================================\n",
|
||
"\n",
|
||
"✅ Cosine Similarity Advantages:\n",
|
||
" 1. 📐 Length-invariant: Not affected by text length\n",
|
||
" 2. 🎯 Semantic focus: Measures meaning, not magnitude\n",
|
||
" 3. 📊 Normalized: Always between 0 and 1 (easy to interpret)\n",
|
||
" 4. 🔄 Consistent: Similar meanings = similar scores\n",
|
||
"\n",
|
||
"❌ Euclidean Distance Disadvantages:\n",
|
||
" 1. 📏 Length-sensitive: Longer texts → larger distances\n",
|
||
" 2. 🎭 Misleading: Can be large even for similar meanings\n",
|
||
" 3. ⚖️ Not normalized: Hard to interpret absolute values\n",
|
||
" 4. 📉 Inconsistent: Same meaning, different distance based on length\n",
|
||
"\n",
|
||
"🌟 This is why NLP systems use COSINE SIMILARITY! 🌟\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"import pandas as pd\n",
|
||
"\n",
|
||
"# Create comparison table\n",
|
||
"comparison_data = {\n",
|
||
" \"Comparison\": [\"Short vs Medium\", \"Short vs Long\", \"Medium vs Long\"],\n",
|
||
" \"Cosine Similarity\": [f\"{sim_1_2:.4f}\", f\"{sim_1_3:.4f}\", f\"{sim_2_3:.4f}\"],\n",
|
||
" \"Euclidean Distance\": [f\"{dist_1_2:.4f}\", f\"{dist_1_3:.4f}\", f\"{dist_2_3:.4f}\"],\n",
|
||
"}\n",
|
||
"\n",
|
||
"df = pd.DataFrame(comparison_data)\n",
|
||
"print(\"📊 Metric Comparison:\\n\")\n",
|
||
"print(df.to_string(index=False))\n",
|
||
"\n",
|
||
"print(\"\\n\" + \"=\" * 70)\n",
|
||
"print(\"🎯 CONCLUSION: Why Cosine Similarity Wins for Text\")\n",
|
||
"print(\"=\" * 70)\n",
|
||
"\n",
|
||
"print(\"\\n✅ Cosine Similarity Advantages:\")\n",
|
||
"print(\" 1. 📐 Length-invariant: Not affected by text length\")\n",
|
||
"print(\" 2. 🎯 Semantic focus: Measures meaning, not magnitude\")\n",
|
||
"print(\" 3. 📊 Normalized: Always between 0 and 1 (easy to interpret)\")\n",
|
||
"print(\" 4. 🔄 Consistent: Similar meanings = similar scores\")\n",
|
||
"\n",
|
||
"print(\"\\n❌ Euclidean Distance Disadvantages:\")\n",
|
||
"print(\" 1. 📏 Length-sensitive: Longer texts → larger distances\")\n",
|
||
"print(\" 2. 🎭 Misleading: Can be large even for similar meanings\")\n",
|
||
"print(\" 3. ⚖️ Not normalized: Hard to interpret absolute values\")\n",
|
||
"print(\" 4. 📉 Inconsistent: Same meaning, different distance based on length\")\n",
|
||
"\n",
|
||
"print(\"\\n🌟 This is why NLP systems use COSINE SIMILARITY! 🌟\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "Aznk0AOrIRFQ"
|
||
},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"# 🎨 Visualizing Embeddings in 2D\n",
|
||
"\n",
|
||
"## 🤯 The Challenge\n",
|
||
"\n",
|
||
"Our embeddings have **384 dimensions** - impossible to visualize directly!\n",
|
||
"\n",
|
||
"**Solution:** Use **UMAP** (Uniform Manifold Approximation and Projection)\n",
|
||
"- 🎯 Reduces 384D → 2D while preserving relationships\n",
|
||
"- 🎨 Keeps similar items close together\n",
|
||
"- 👀 Makes patterns visible to human eyes\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"## 🧪 The Experiment\n",
|
||
"\n",
|
||
"We'll embed words from **two different domains**:\n",
|
||
"- 🐾 **Animals:** Cat, Dog, Elephant, Tiger, Lion\n",
|
||
"- 💻 **Technology:** Phone, Antenna, Modem, Router, Cable\n",
|
||
"\n",
|
||
"**Hypothesis:** Words from the same domain should cluster together!"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 17,
|
||
"metadata": {
|
||
"id": "xCrpwv63_m9K"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"📋 Test Words:\n",
|
||
"\n",
|
||
"🐾 Animals:\n",
|
||
" Cat, Dog, Elephant, Tiger, Lion\n",
|
||
"\n",
|
||
"💻 Technology:\n",
|
||
" Phone, Antenna, Modem, Router, Cable\n",
|
||
"\n",
|
||
"🎯 Goal: See if embeddings naturally separate these domains!\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Our test words\n",
|
||
"elements = [\n",
|
||
" # Animals\n",
|
||
" \"Cat\",\n",
|
||
" \"Dog\",\n",
|
||
" \"Elephant\",\n",
|
||
" \"Tiger\",\n",
|
||
" \"Lion\",\n",
|
||
" # Technology\n",
|
||
" \"Phone\",\n",
|
||
" \"Antenna\",\n",
|
||
" \"Modem\",\n",
|
||
" \"Router\",\n",
|
||
" \"Cable\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"# Categories for coloring\n",
|
||
"categories = [\n",
|
||
" \"Animals\",\n",
|
||
" \"Animals\",\n",
|
||
" \"Animals\",\n",
|
||
" \"Animals\",\n",
|
||
" \"Animals\",\n",
|
||
" \"Technology\",\n",
|
||
" \"Technology\",\n",
|
||
" \"Technology\",\n",
|
||
" \"Technology\",\n",
|
||
" \"Technology\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"print(\"📋 Test Words:\")\n",
|
||
"print(\"\\n🐾 Animals:\")\n",
|
||
"print(\" \" + \", \".join(elements[:5]))\n",
|
||
"print(\"\\n💻 Technology:\")\n",
|
||
"print(\" \" + \", \".join(elements[5:]))\n",
|
||
"print(\"\\n🎯 Goal: See if embeddings naturally separate these domains!\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 18,
|
||
"metadata": {
|
||
"id": "sT6m2Wkv2gZx"
|
||
},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"🔄 Step 1: Creating 1536D embeddings...\n",
|
||
" ✅ Created 10 embeddings of 1536 dimensions each\n",
|
||
"\n",
|
||
"📉 Step 2: Reducing 1536D → 2D with UMAP...\n"
|
||
]
|
||
},
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/umap/umap_.py:1952: UserWarning: n_jobs value 1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
|
||
" warn(\n"
|
||
]
|
||
},
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
" ✅ Reduced to 2D coordinates\n",
|
||
"\n",
|
||
"🎨 Step 3: Creating visualization...\n",
|
||
"\n"
|
||
]
|
||
},
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/var/folders/tp/_ld5_pzs6nx6mv1pbjhq1l740000gn/T/ipykernel_24511/980470006.py:68: UserWarning: Glyph 127919 (\\N{DIRECT HIT}) missing from font(s) DejaVu Sans.\n",
|
||
" plt.tight_layout()\n",
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/IPython/core/pylabtools.py:170: UserWarning: Glyph 127919 (\\N{DIRECT HIT}) missing from font(s) DejaVu Sans.\n",
|
||
" fig.canvas.print_figure(bytes_io, **kw)\n"
|
||
]
|
||
},
|
||
{
|
||
"data": {
|
||
"image/png": "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",
|
||
"text/plain": [
|
||
"<Figure size 1400x1000 with 1 Axes>"
|
||
]
|
||
},
|
||
"metadata": {},
|
||
"output_type": "display_data"
|
||
},
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"🔍 Analysis:\n",
|
||
" 🐾 RED cluster = Animals (Cat, Dog, Elephant, Tiger, Lion)\n",
|
||
" 💻 TEAL cluster = Technology (Phone, Modem, Router, etc.)\n",
|
||
"\n",
|
||
"✨ Magic Observation:\n",
|
||
" • The model AUTOMATICALLY groups related concepts together!\n",
|
||
" • No human told it that 'Cat' and 'Dog' are related\n",
|
||
" • It learned these relationships from massive text data\n",
|
||
" • This is the power of embeddings! 🚀\n",
|
||
"\n",
|
||
"💡 Why This Matters:\n",
|
||
" 1. Search: Find similar items even with different words\n",
|
||
" 2. Clustering: Automatically group related content\n",
|
||
" 3. Recommendations: Suggest similar items\n",
|
||
" 4. Classification: Categorize new content\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"import umap\n",
|
||
"\n",
|
||
"rng = np.random.default_rng(42)\n",
|
||
"# Set random seed for reproducibilit\n",
|
||
"\n",
|
||
"print(\"🔄 Step 1: Creating 1536D embeddings...\")\n",
|
||
"embeddings_1536d = embedding_model.embed_documents(elements)\n",
|
||
"print(\n",
|
||
" f\" ✅ Created {len(embeddings_1536d)} embeddings of {len(embeddings_1536d[0])} dimensions each\",\n",
|
||
")\n",
|
||
"\n",
|
||
"print(\"\\n📉 Step 2: Reducing 1536D → 2D with UMAP...\")\n",
|
||
"reducer = umap.UMAP(\n",
|
||
" n_neighbors=3, # Local neighborhood size\n",
|
||
" min_dist=0.1, # Minimum distance between points\n",
|
||
" metric=\"cosine\", # Use cosine similarity\n",
|
||
" random_state=42,\n",
|
||
")\n",
|
||
"embeddings_2d = reducer.fit_transform(embeddings_1536d)\n",
|
||
"print(\" ✅ Reduced to 2D coordinates\")\n",
|
||
"\n",
|
||
"print(\"\\n🎨 Step 3: Creating visualization...\\n\")\n",
|
||
"\n",
|
||
"# Create beautiful plot\n",
|
||
"plt.figure(figsize=(14, 10))\n",
|
||
"\n",
|
||
"# Scatter plot with categories\n",
|
||
"scatter = sns.scatterplot(\n",
|
||
" x=embeddings_2d[:, 0],\n",
|
||
" y=embeddings_2d[:, 1],\n",
|
||
" hue=categories,\n",
|
||
" palette=[\"#FF6B6B\", \"#4ECDC4\"], # Red for animals, teal for tech\n",
|
||
" s=300,\n",
|
||
" alpha=0.7,\n",
|
||
" edgecolor=\"black\",\n",
|
||
" linewidth=2,\n",
|
||
")\n",
|
||
"\n",
|
||
"# Add labels for each word\n",
|
||
"for i, word in enumerate(elements):\n",
|
||
" plt.annotate(\n",
|
||
" word,\n",
|
||
" (embeddings_2d[i, 0], embeddings_2d[i, 1]),\n",
|
||
" xytext=(10, 10),\n",
|
||
" textcoords=\"offset points\",\n",
|
||
" fontsize=12,\n",
|
||
" fontweight=\"bold\",\n",
|
||
" bbox={\n",
|
||
" \"boxstyle\": \"round,pad=0.5\", \"facecolor\": \"white\", \"alpha\": 0.7, \"edgecolor\": \"gray\",\n",
|
||
" },\n",
|
||
" arrowprops={\n",
|
||
" \"arrowstyle\": \"->\", \"connectionstyle\": \"arc3,rad=0\", \"color\": \"gray\", \"lw\": 1.5,\n",
|
||
" },\n",
|
||
" )\n",
|
||
"\n",
|
||
"# Styling\n",
|
||
"plt.title(\n",
|
||
" \"🎯 Semantic Space: Animals vs Technology\\n(1536D embeddings projected to 2D)\",\n",
|
||
" fontsize=16,\n",
|
||
" fontweight=\"bold\",\n",
|
||
" pad=20,\n",
|
||
")\n",
|
||
"plt.xlabel(\"Dimension 1\", fontsize=13, fontweight=\"bold\")\n",
|
||
"plt.ylabel(\"Dimension 2\", fontsize=13, fontweight=\"bold\")\n",
|
||
"plt.legend(title=\"Category\", fontsize=12, title_fontsize=13, loc=\"best\")\n",
|
||
"plt.grid(visible=True, alpha=0.3, linestyle=\"--\")\n",
|
||
"\n",
|
||
"plt.tight_layout()\n",
|
||
"plt.show()\n",
|
||
"\n",
|
||
"# Analysis\n",
|
||
"print(\"\\n🔍 Analysis:\")\n",
|
||
"print(\" 🐾 RED cluster = Animals (Cat, Dog, Elephant, Tiger, Lion)\")\n",
|
||
"print(\" 💻 TEAL cluster = Technology (Phone, Modem, Router, etc.)\")\n",
|
||
"print(\"\\n✨ Magic Observation:\")\n",
|
||
"print(\" • The model AUTOMATICALLY groups related concepts together!\")\n",
|
||
"print(\" • No human told it that 'Cat' and 'Dog' are related\")\n",
|
||
"print(\" • It learned these relationships from massive text data\")\n",
|
||
"print(\" • This is the power of embeddings! 🚀\")\n",
|
||
"\n",
|
||
"print(\"\\n💡 Why This Matters:\")\n",
|
||
"print(\" 1. Search: Find similar items even with different words\")\n",
|
||
"print(\" 2. Clustering: Automatically group related content\")\n",
|
||
"print(\" 3. Recommendations: Suggest similar items\")\n",
|
||
"print(\" 4. Classification: Categorize new content\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {
|
||
"id": "ua8DR4AbIpOq"
|
||
},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"# 💡 Real-World Applications\n",
|
||
"\n",
|
||
"## 🚀 Where Are Embeddings Used?\n",
|
||
"\n",
|
||
"Embeddings power many modern AI applications:\n",
|
||
"\n",
|
||
"### 1. 🔍 Semantic Search\n",
|
||
"**Use Case:** Search by meaning, not just keywords\n",
|
||
"- Query: \"how to fix a broken faucet\"\n",
|
||
"- Finds: \"plumbing repair for leaky tap\" ✅\n",
|
||
"- Traditional search would miss this!\n",
|
||
"\n",
|
||
"**Example:** Google Search, Elasticsearch with vector search\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 2. 🤖 Chatbots & Virtual Assistants\n",
|
||
"**Use Case:** Understanding user intent\n",
|
||
"- \"I'm hungry\" ≈ \"I'd like something to eat\" ≈ \"Feed me\"\n",
|
||
"- All trigger the same response\n",
|
||
"\n",
|
||
"**Example:** Customer service bots, Siri, Alexa\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 3. 📚 Recommendation Systems\n",
|
||
"**Use Case:** Find similar content\n",
|
||
"- Watched \"The Matrix\"? Try \"Inception\" (similar themes)\n",
|
||
"- Read article about AI? Here are related papers\n",
|
||
"\n",
|
||
"**Example:** Netflix, YouTube, Amazon recommendations\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 4. 📝 Document Clustering\n",
|
||
"**Use Case:** Automatically organize documents\n",
|
||
"- Group news articles by topic\n",
|
||
"- Organize customer support tickets\n",
|
||
"- Cluster research papers by subject\n",
|
||
"\n",
|
||
"**Example:** Google News topic grouping\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 5. 🔐 Duplicate Detection\n",
|
||
"**Use Case:** Find similar or duplicate content\n",
|
||
"- Detect plagiarism\n",
|
||
"- Remove duplicate customer queries\n",
|
||
"- Find copyright violations\n",
|
||
"\n",
|
||
"**Example:** Turnitin, Stack Overflow duplicate detection\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 6. 🌐 Translation & Cross-Lingual Search\n",
|
||
"**Use Case:** Search across languages\n",
|
||
"- Query in English, find French documents\n",
|
||
"- Multilingual customer support\n",
|
||
"\n",
|
||
"**Example:** Wikipedia cross-language search\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 7. 📊 Text Classification\n",
|
||
"**Use Case:** Automatically categorize text\n",
|
||
"- Email spam detection\n",
|
||
"- Sentiment analysis (positive/negative reviews)\n",
|
||
"- Topic classification (sports, politics, tech)\n",
|
||
"\n",
|
||
"**Example:** Gmail spam filter, Twitter sentiment analysis\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 8. 🎯 RAG (Retrieval-Augmented Generation)\n",
|
||
"**Use Case:** Power AI assistants with relevant context\n",
|
||
"- Embed company documents\n",
|
||
"- Find relevant passages for LLM context\n",
|
||
"- Ground AI responses in real data\n",
|
||
"\n",
|
||
"**Example:** ChatGPT with custom knowledge bases\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 9. 🔗 Entity Linking\n",
|
||
"**Use Case:** Connect mentions to real entities\n",
|
||
"- \"Apple\" → Apple Inc. (company) or apple (fruit)?\n",
|
||
"- Disambiguate based on context\n",
|
||
"\n",
|
||
"**Example:** Wikipedia entity linking\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 10. 🎨 Creative Applications\n",
|
||
"**Use Case:** AI art & content generation\n",
|
||
"- Text-to-image (DALL-E, Midjourney)\n",
|
||
"- Music generation based on mood\n",
|
||
"- Story continuation\n",
|
||
"\n",
|
||
"**Example:** DALL-E converts text embeddings to image space"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"---\n",
|
||
"\n",
|
||
"# 🎉 Congratulations!\n",
|
||
"\n",
|
||
"You've completed the embeddings tutorial! Here's what you've mastered:\n",
|
||
"\n",
|
||
"## 📚 Key Concepts:\n",
|
||
"\n",
|
||
"✅ **Embeddings** - Converting text to semantic vectors \n",
|
||
"✅ **Sentence Transformers** - Pre-trained models for embeddings \n",
|
||
"✅ **Similarity Metrics** - Cosine similarity vs Euclidean distance \n",
|
||
"✅ **Dimensionality Reduction** - Visualizing high-D data with UMAP \n",
|
||
"✅ **Semantic Clustering** - Finding natural groups in text \n",
|
||
"✅ **Real Applications** - Where embeddings power modern AI \n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"## 🎯 Key Takeaways:\n",
|
||
"\n",
|
||
"1. **Embeddings = Semantic Fingerprints**\n",
|
||
" - Each text gets a unique vector representation\n",
|
||
" - Similar meanings → Similar vectors\n",
|
||
"\n",
|
||
"2. **Cosine Similarity FTW**\n",
|
||
" - Best for comparing text embeddings\n",
|
||
" - Length-invariant, normalized, semantic-focused\n",
|
||
"\n",
|
||
"3. **Pre-trained Models Are Powerful**\n",
|
||
" - No need to train from scratch\n",
|
||
" - Multilingual capabilities out of the box\n",
|
||
" - Ready for production use\n",
|
||
"\n",
|
||
"4. **Dimensionality Reduction Makes Patterns Visible**\n",
|
||
" - UMAP/t-SNE reveal structure in high-D data\n",
|
||
" - Great for exploration and understanding\n",
|
||
"\n",
|
||
"5. **Embeddings Power Modern AI**\n",
|
||
" - From search to recommendations\n",
|
||
" - From chatbots to content generation\n",
|
||
" - The foundation of most NLP systems\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"## 🚀 Next Steps:\n",
|
||
"\n",
|
||
"1. **🧪 Experiment**\n",
|
||
" - Try different sentences and languages\n",
|
||
" - Test various domains (sports, science, art)\n",
|
||
" - Compare different models from Sentence Transformers\n",
|
||
"\n",
|
||
"2. **🔨 Build Something**\n",
|
||
" - Create a semantic search engine\n",
|
||
" - Build a recommendation system\n",
|
||
" - Make a duplicate detector\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"<div style=\"text-align: center; background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 30px; border-radius: 15px; margin: 30px 0;\">\n",
|
||
" <h2 style=\"margin: 0; font-size: 2em;\">🎓 You Now Understand Embeddings!</h2>\n",
|
||
" <p style=\"margin: 15px 0 0 0; font-size: 1.2em;\">The foundation of modern NLP is in your hands! 🚀</p>\n",
|
||
"</div>\n",
|
||
"\n",
|
||
"---\n",
|
||
"\n",
|
||
"### 🌟 Remember:\n",
|
||
"\n",
|
||
"> *\"Embeddings turn words into geometry, making language computable and comparable.\"*\n",
|
||
"\n",
|
||
"Now go build something amazing! 💪✨"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"## 📚 Exercise Set 1: Basic Embedding Operations"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 30,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Most similar pair: 'I love programming in Python' and 'Coding with Python is my favorite hobby'\n",
|
||
"Cosine similarity score: 0.6924\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"sentences = [\n",
|
||
" \"I love programming in Python\",\n",
|
||
" \"Coding with Python is my favorite hobby\",\n",
|
||
" \"The sunset was beautiful yesterday\",\n",
|
||
" \"I enjoy playing basketball\",\n",
|
||
" \"Python is a great programming language\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"# Your code here:\n",
|
||
"# 1. Generate embeddings for all sentences\n",
|
||
"embeddings = embedding_model.embed_documents(sentences)\n",
|
||
"# 2. Calculate cosine similarity between all pairs\n",
|
||
"similarities = np.dot(embeddings, np.array(embeddings).T)\n",
|
||
"np.fill_diagonal(similarities, -1) # Ignore self-similarity\n",
|
||
"max_sim_indices = np.unravel_index(np.argmax(similarities), similarities.shape)\n",
|
||
"most_similar_pair = (sentences[max_sim_indices[0]], sentences[max_sim_indices[1]])\n",
|
||
"similarity_score = similarities[max_sim_indices]\n",
|
||
"# 3. Print the most similar pair and their similarity score\n",
|
||
"print(f\"Most similar pair: '{most_similar_pair[0]}' and '{most_similar_pair[1]}'\")\n",
|
||
"print(f\"Cosine similarity score: {similarity_score:.4f}\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 22,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Score: 0.507 | Doc: Neural networks mimic the human brain\n",
|
||
"Score: 0.376 | Doc: Deep learning is a subset of machine learning\n",
|
||
"Score: 0.342 | Doc: Machine learning uses statistical techniques\n",
|
||
"Most similar pair: 'I love programming in Python' and 'Coding with Python is my favorite hobby'\n",
|
||
"Cosine similarity score: 0.6924\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"documents = [\n",
|
||
" \"Python is a high-level programming language\",\n",
|
||
" \"Machine learning uses statistical techniques\",\n",
|
||
" \"Dogs are loyal pets and companions\",\n",
|
||
" \"Neural networks mimic the human brain\",\n",
|
||
" \"Cats are independent animals\",\n",
|
||
" \"Deep learning is a subset of machine learning\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"def semantic_search(query: str, documents: list, top_k: int = 3) -> list[tuple[str, int | float]]:\n",
|
||
" \"\"\"Find the most relevant documents for a query.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" query: Search query string\n",
|
||
" documents: List of documents to search\n",
|
||
" top_k: Number of top results to return\n",
|
||
"\n",
|
||
" Returns:\n",
|
||
" List of (document, similarity_score) tuples\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" # Your code here:\n",
|
||
" # 1. Embed the query\n",
|
||
" # 2. Embed all documents\n",
|
||
" # 3. Calculate similarities\n",
|
||
" # 4. Return top_k most similar documents\n",
|
||
" query_emb = embedding_model.embed_query(query)\n",
|
||
" doc_embs = embedding_model.embed_documents(documents)\n",
|
||
" similarities = np.dot(doc_embs, query_emb)\n",
|
||
" top_indices = np.argsort(similarities)[-top_k:][::-1]\n",
|
||
" return [(documents[i], similarities[i]) for i in top_indices]\n",
|
||
"\n",
|
||
"# Test your function\n",
|
||
"query = \"What is AI and neural networks?\"\n",
|
||
"results = semantic_search(query, documents, top_k=3)\n",
|
||
"for doc, score in results:\n",
|
||
" print(f\"Score: {score:.3f} | Doc: {doc}\")\n",
|
||
"\n",
|
||
"embeddings = embedding_model.embed_documents(sentences)\n",
|
||
"similarities = np.dot(embeddings, np.array(embeddings).T)\n",
|
||
"np.fill_diagonal(similarities, -1) # Ignore self-similarity\n",
|
||
"max_sim_indices = np.unravel_index(np.argmax(similarities), similarities.shape)\n",
|
||
"most_similar_pair = (sentences[max_sim_indices[0]], sentences[max_sim_indices[1]])\n",
|
||
"similarity_score = similarities[max_sim_indices]\n",
|
||
"\n",
|
||
"print(f\"Most similar pair: '{most_similar_pair[0]}' and '{most_similar_pair[1]}'\")\n",
|
||
"print(f\"Cosine similarity score: {similarity_score:.4f}\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 31,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"English: 'Good morning' | Best French Match: 'Bonjour' | Similarity: 0.4983\n",
|
||
"English: 'Thank you' | Best French Match: 'Merci' | Similarity: 0.5890\n",
|
||
"English: 'How are you?' | Best French Match: 'Comment allez-vous?' | Similarity: 0.5474\n",
|
||
"English: 'Goodbye' | Best French Match: 'Au revoir' | Similarity: 0.7500\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"english_phrases = [\n",
|
||
" \"Good morning\",\n",
|
||
" \"Thank you\",\n",
|
||
" \"How are you?\",\n",
|
||
" \"Goodbye\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"french_phrases = [\n",
|
||
" \"Au revoir\",\n",
|
||
" \"Bonjour\",\n",
|
||
" \"Merci\",\n",
|
||
" \"Comment allez-vous?\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"# Your code here:\n",
|
||
"# 1. Create embeddings for both lists\n",
|
||
"# 2. For each English phrase, find its best French match\n",
|
||
"# 3. Print the matches with similarity scores\n",
|
||
"embeddings_en = embedding_model.embed_documents(english_phrases)\n",
|
||
"embeddings_fr = embedding_model.embed_documents(french_phrases)\n",
|
||
"similarities = np.dot(embeddings_en, np.array(embeddings_fr).T)\n",
|
||
"for i, eng_phrase in enumerate(english_phrases):\n",
|
||
" best_match_idx = np.argmax(similarities[i])\n",
|
||
" best_match_fr = french_phrases[best_match_idx]\n",
|
||
" similarity_score = similarities[i][best_match_idx]\n",
|
||
" print(f\"English: '{eng_phrase}' | Best French Match: '{best_match_fr}' | Similarity: {similarity_score:.4f}\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"## 📊 Exercise Set 2: Advanced Similarity Analysis\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 25,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"False: 'The cat is on the mat' ↔ 'A feline sits on the rug'\n",
|
||
"False: 'I enjoy pizza' ↔ 'Pizza is my favorite food'\n",
|
||
"False: 'The sky is blue' ↔ 'Elephants are large animals'\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"def is_paraphrase(sentence1: str, sentence2: str, threshold: float = 0.85) -> bool:\n",
|
||
" \"\"\"Determine if two sentences are paraphrases of each other.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" sentence1: First sentence\n",
|
||
" sentence2: Second sentence\n",
|
||
" threshold: Minimum similarity to consider paraphrases\n",
|
||
"\n",
|
||
" Returns:\n",
|
||
" True if paraphrases, False otherwise\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" emb1 = embedding_model.embed_query(sentence1)\n",
|
||
" emb2 = embedding_model.embed_query(sentence2)\n",
|
||
" similarity = cosine_similarity(emb1, emb2)\n",
|
||
" return similarity >= threshold\n",
|
||
"\n",
|
||
"# Test cases\n",
|
||
"test_pairs = [\n",
|
||
" (\"The cat is on the mat\", \"A feline sits on the rug\"),\n",
|
||
" (\"I enjoy pizza\", \"Pizza is my favorite food\"),\n",
|
||
" (\"The sky is blue\", \"Elephants are large animals\"),\n",
|
||
"]\n",
|
||
"\n",
|
||
"for sent1, sent2 in test_pairs:\n",
|
||
" result = is_paraphrase(sent1, sent2)\n",
|
||
" print(f\"{result}: '{sent1}' ↔ '{sent2}'\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 26,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Group 1 outlier: The car needs gasoline\n",
|
||
"Group 2 outlier: Bananas are yellow fruits\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"groups = [\n",
|
||
" [\n",
|
||
" \"I love reading books\",\n",
|
||
" \"Books are my passion\",\n",
|
||
" \"Reading is enjoyable\",\n",
|
||
" \"The car needs gasoline\", # Outlier\n",
|
||
" \"I enjoy a good novel\",\n",
|
||
" ],\n",
|
||
" [\n",
|
||
" \"Python is a programming language\",\n",
|
||
" \"Java is used for software development\",\n",
|
||
" \"JavaScript runs in browsers\",\n",
|
||
" \"Bananas are yellow fruits\", # Outlier\n",
|
||
" \"C++ is a compiled language\",\n",
|
||
" ],\n",
|
||
"]\n",
|
||
"\n",
|
||
"def find_outlier(sentences: list) -> str:\n",
|
||
" \"\"\"Find the sentence that is least similar to others in the group.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" sentences: List of sentences\n",
|
||
"\n",
|
||
" Returns:\n",
|
||
" The outlier sentence\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" embeddings = embedding_model.embed_documents(sentences)\n",
|
||
" similarities = np.dot(embeddings, np.array(embeddings).T)\n",
|
||
" np.fill_diagonal(similarities, 0) # Ignore self-similarity\n",
|
||
" avg_similarities = similarities.sum(axis=1) / (len(sentences) - 1)\n",
|
||
" outlier_index = np.argmin(avg_similarities)\n",
|
||
" return sentences[outlier_index]\n",
|
||
"\n",
|
||
"for i, group in enumerate(groups, 1):\n",
|
||
" outlier = find_outlier(group)\n",
|
||
" print(f\"Group {i} outlier: {outlier}\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"## 🎨 Exercise Set 3: Visualization Challenges"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 41,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/umap/umap_.py:1952: UserWarning: n_jobs value 1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
|
||
" warn(\n",
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/umap/umap_.py:2462: UserWarning: n_neighbors is larger than the dataset size; truncating to X.shape[0] - 1\n",
|
||
" warn(\n"
|
||
]
|
||
},
|
||
{
|
||
"data": {
|
||
"image/png": "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",
|
||
"text/plain": [
|
||
"<Figure size 800x600 with 1 Axes>"
|
||
]
|
||
},
|
||
"metadata": {},
|
||
"output_type": "display_data"
|
||
}
|
||
],
|
||
"source": [
|
||
"domains = {\n",
|
||
" \"Sports\": [\"Football\", \"Basketball\", \"Tennis\", \"Swimming\", \"Running\"],\n",
|
||
" \"Food\": [\"Pizza\", \"Sushi\", \"Burger\", \"Pasta\", \"Salad\"],\n",
|
||
" \"Music\": [\"Guitar\", \"Piano\", \"Drums\", \"Violin\", \"Saxophone\"],\n",
|
||
"}\n",
|
||
"\n",
|
||
"# Your code here:\n",
|
||
"# 1. Flatten the words and track their domains\n",
|
||
"all_words = []\n",
|
||
"domain_labels = []\n",
|
||
"for domain, words in domains.items():\n",
|
||
" all_words.extend(words)\n",
|
||
" domain_labels.extend([domain] * len(words))\n",
|
||
"\n",
|
||
"# 2. Generate embeddings\n",
|
||
"embeddings = embedding_model.embed_documents(all_words)\n",
|
||
"\n",
|
||
"# 3. Reduce to 2D with UMAP\n",
|
||
"umap_model = umap.UMAP(n_components=2, random_state=42)\n",
|
||
"reduced_embeddings = umap_model.fit_transform(embeddings)\n",
|
||
"\n",
|
||
"# 4. Create a scatter plot with different colors per domain\n",
|
||
"import matplotlib.pyplot as plt\n",
|
||
"\n",
|
||
"plt.figure(figsize=(8, 6))\n",
|
||
"for domain in domains:\n",
|
||
" indices = [i for i, label in enumerate(domain_labels) if label == domain]\n",
|
||
" x_coords = [reduced_embeddings[i][0] for i in indices]\n",
|
||
" y_coords = [reduced_embeddings[i][1] for i in indices]\n",
|
||
" plt.scatter(x_coords, y_coords, label=domain, alpha=0.7)\n",
|
||
"plt.title(\"UMAP Projection of Word Embeddings by Domain\")\n",
|
||
"# 5. Add labels to each point\n",
|
||
"for i, word in enumerate(all_words):\n",
|
||
" plt.annotate(word, (reduced_embeddings[i][0], reduced_embeddings[i][1]), fontsize=8, alpha=0.7)\n",
|
||
"\n",
|
||
"plt.legend()\n",
|
||
"plt.grid(visible=True, alpha=0.3)\n",
|
||
"plt.show()\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 42,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/umap/umap_.py:1952: UserWarning: n_jobs value 1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
|
||
" warn(\n",
|
||
"/Users/arthurdanjou/Workspace/studies/.venv/lib/python3.13/site-packages/umap/umap_.py:2462: UserWarning: n_neighbors is larger than the dataset size; truncating to X.shape[0] - 1\n",
|
||
" warn(\n"
|
||
]
|
||
},
|
||
{
|
||
"data": {
|
||
"image/png": "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",
|
||
"text/plain": [
|
||
"<Figure size 1000x800 with 1 Axes>"
|
||
]
|
||
},
|
||
"metadata": {},
|
||
"output_type": "display_data"
|
||
}
|
||
],
|
||
"source": [
|
||
"start_word = \"Summer\"\n",
|
||
"end_word = \"Winter\"\n",
|
||
"related_words = [\n",
|
||
" \"Summer\", \"Warm\", \"Sunny\", \"Hot\", \"Beach\",\n",
|
||
" \"Autumn\", \"Cool\", \"Falling leaves\", \"Harvest\",\n",
|
||
" \"Winter\", \"Cold\", \"Snow\", \"Ice\", \"Freezing\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"# Your code here:\n",
|
||
"# 1. Generate embeddings for all words\n",
|
||
"embeddings = embedding_model.embed_documents(related_words)\n",
|
||
"\n",
|
||
"# 2. Reduce to 2D\n",
|
||
"umap_model = umap.UMAP(n_components=2, random_state=42)\n",
|
||
"reduced_embeddings = umap_model.fit_transform(embeddings)\n",
|
||
"# 3. Create a path visualization connecting summer → autumn → winter\n",
|
||
"plt.figure(figsize=(10, 8))\n",
|
||
"for i, word in enumerate(related_words):\n",
|
||
" plt.scatter(reduced_embeddings[i][0], reduced_embeddings[i][1], alpha=0.7)\n",
|
||
" plt.annotate(word, (reduced_embeddings[i][0], reduced_embeddings[i][1]), fontsize=9)\n",
|
||
"plt.title(\"Semantic Journey: Summer → Autumn → Winter\")\n",
|
||
"# 4. Use arrows to show the semantic journey\n",
|
||
"for i in range(len(related_words) - 1):\n",
|
||
" plt.arrow(\n",
|
||
" reduced_embeddings[i][0],\n",
|
||
" reduced_embeddings[i][1],\n",
|
||
" reduced_embeddings[i + 1][0] - reduced_embeddings[i][0],\n",
|
||
" reduced_embeddings[i + 1][1] - reduced_embeddings[i][1],\n",
|
||
" head_width=0.05,\n",
|
||
" head_length=0.1,\n",
|
||
" fc=\"gray\",\n",
|
||
" ec=\"gray\",\n",
|
||
" alpha=0.5,\n",
|
||
" )\n",
|
||
"plt.show()\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"## 🔨 Exercise Set 4: Real-World Applications"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 45,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"Question: I forgot my password, what should I do?\n",
|
||
" - Match: How do I reset my password? (similarity: 0.6909)\n",
|
||
" - Answer: Go to Settings > Security > Reset Password\n",
|
||
"\n",
|
||
"Question: How much does delivery cost?\n",
|
||
" - Match: What are your shipping costs? (similarity: 0.5593)\n",
|
||
" - Answer: Shipping is free for orders over $50\n",
|
||
"\n",
|
||
"Question: Can I return items I purchased?\n",
|
||
" - Match: Do you accept returns? (similarity: 0.6974)\n",
|
||
" - Answer: Yes, within 30 days with original receipt\n",
|
||
"\n",
|
||
"Question: What is the meaning of life?\n",
|
||
" No matching FAQ found (below threshold)\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"faqs = {\n",
|
||
" \"How do I reset my password?\": \"Go to Settings > Security > Reset Password\",\n",
|
||
" \"What are your shipping costs?\": \"Shipping is free for orders over $50\",\n",
|
||
" \"How can I track my order?\": \"Use the tracking number sent to your email\",\n",
|
||
" \"Do you accept returns?\": \"Yes, within 30 days with original receipt\",\n",
|
||
" \"What payment methods do you accept?\": \"We accept credit cards, PayPal, and bank transfers\",\n",
|
||
"}\n",
|
||
"\n",
|
||
"def match_faq(user_question: str, faqs: dict, threshold: float = 0.5) -> tuple[str, str, float] | None:\n",
|
||
" \"\"\"Find the best matching FAQ for a user question.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" user_question: Question from user\n",
|
||
" faqs: Dictionary of FAQ questions and answers\n",
|
||
" threshold: Minimum similarity to return a match (default 0.5 for loose matching, use 0.7+ for strict)\n",
|
||
"\n",
|
||
" Returns:\n",
|
||
" Tuple of (matching_question, answer, similarity) or None\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" # Generate embedding for user question\n",
|
||
" user_emb = embedding_model.embed_query(user_question)\n",
|
||
"\n",
|
||
" # Find best matching FAQ\n",
|
||
" best_match = None\n",
|
||
" best_similarity = -1\n",
|
||
"\n",
|
||
" for faq_question, answer in faqs.items():\n",
|
||
" faq_emb = embedding_model.embed_query(faq_question)\n",
|
||
" similarity = cosine_similarity(user_emb, faq_emb)\n",
|
||
"\n",
|
||
" if similarity > best_similarity:\n",
|
||
" best_similarity = similarity\n",
|
||
" best_match = (faq_question, answer, similarity)\n",
|
||
"\n",
|
||
" # Return match only if above threshold\n",
|
||
" if best_similarity >= threshold:\n",
|
||
" return best_match\n",
|
||
" return None\n",
|
||
"\n",
|
||
"# Test queries\n",
|
||
"test_queries = [\n",
|
||
" \"I forgot my password, what should I do?\",\n",
|
||
" \"How much does delivery cost?\",\n",
|
||
" \"Can I return items I purchased?\",\n",
|
||
" \"What is the meaning of life?\", # Should not match\n",
|
||
"]\n",
|
||
"\n",
|
||
"for query in test_queries:\n",
|
||
" result = match_faq(query, faqs)\n",
|
||
" if result:\n",
|
||
" print(f\"\\nQuestion: {query}\")\n",
|
||
" print(f\" - Match: {result[0]} (similarity: {result[2]:.4f})\")\n",
|
||
" print(f\" - Answer: {result[1]}\")\n",
|
||
" else:\n",
|
||
" print(f\"\\nQuestion: {query}\")\n",
|
||
" print(\" No matching FAQ found (below threshold)\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 36,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Duplicate groups found:\n",
|
||
"\n",
|
||
"Group 1:\n",
|
||
" [0] The quick brown fox jumps over the lazy dog\n",
|
||
" [5] The quick brown fox jumps over the lazy dog\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"documents = [\n",
|
||
" \"The quick brown fox jumps over the lazy dog\",\n",
|
||
" \"A fast brown fox leaps over a sleepy dog\", # Near-duplicate of #1\n",
|
||
" \"Python is a popular programming language\",\n",
|
||
" \"The weather is nice today\",\n",
|
||
" \"Python is widely used for programming\", # Near-duplicate of #3\n",
|
||
" \"The quick brown fox jumps over the lazy dog\", # Exact duplicate of #1\n",
|
||
"]\n",
|
||
"\n",
|
||
"def find_duplicates(documents: list, similarity_threshold: float = 0.85) -> list[list[int]]:\n",
|
||
" \"\"\"Find groups of near-duplicate documents.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" documents: List of text documents\n",
|
||
" similarity_threshold: Minimum similarity to consider duplicates\n",
|
||
"\n",
|
||
" Returns:\n",
|
||
" List of duplicate groups, each group is a list of indices\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" # Your code here:\n",
|
||
" # 1. Generate embeddings\n",
|
||
" embeddings = embedding_model.embed_documents(documents)\n",
|
||
" # 2. Compute pairwise similarities\n",
|
||
" n = len(embeddings)\n",
|
||
" similarities = np.zeros((n, n))\n",
|
||
" for i in range(n):\n",
|
||
" for j in range(n):\n",
|
||
" similarities[i, j] = cosine_similarity(embeddings[i], embeddings[j])\n",
|
||
" # 3. Group documents with similarity > threshold\n",
|
||
" visited = set()\n",
|
||
" duplicate_groups = []\n",
|
||
" for i in range(n):\n",
|
||
" if i in visited:\n",
|
||
" continue\n",
|
||
" group = [i]\n",
|
||
" visited.add(i)\n",
|
||
" for j in range(i + 1, n):\n",
|
||
" if j not in visited and similarities[i, j] > similarity_threshold:\n",
|
||
" group.append(j)\n",
|
||
" visited.add(j)\n",
|
||
" if len(group) > 1:\n",
|
||
" duplicate_groups.append(group)\n",
|
||
" # 4. Return groups of duplicate indices\n",
|
||
" return duplicate_groups\n",
|
||
"\n",
|
||
"duplicate_groups = find_duplicates(documents)\n",
|
||
"print(\"Duplicate groups found:\")\n",
|
||
"for i, group in enumerate(duplicate_groups, 1):\n",
|
||
" print(f\"\\nGroup {i}:\")\n",
|
||
" for idx in group:\n",
|
||
" print(f\" [{idx}] {documents[idx]}\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 37,
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"User has read:\n",
|
||
" • Introduction to Machine Learning with Python\n",
|
||
" • Understanding Neural Networks and Deep Learning\n",
|
||
"\n",
|
||
"Recommended articles:\n",
|
||
" • Building AI Applications with TensorFlow\n",
|
||
" • Advanced Python Programming Techniques\n",
|
||
" • How to Train Your Dog: A Complete Guide\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"articles = [\n",
|
||
" \"Introduction to Machine Learning with Python\",\n",
|
||
" \"10 Best Pasta Recipes for Beginners\",\n",
|
||
" \"Understanding Neural Networks and Deep Learning\",\n",
|
||
" \"How to Train Your Dog: A Complete Guide\",\n",
|
||
" \"Advanced Python Programming Techniques\",\n",
|
||
" \"The Ultimate Guide to Italian Cooking\",\n",
|
||
" \"Cat Behavior: Understanding Your Feline Friend\",\n",
|
||
" \"Building AI Applications with TensorFlow\",\n",
|
||
"]\n",
|
||
"\n",
|
||
"def recommend_articles(user_history: list, all_articles: list, n_recommendations: int = 3) -> list[int]:\n",
|
||
" \"\"\"Recommend articles based on reading history.\n",
|
||
"\n",
|
||
" Args:\n",
|
||
" user_history: List of article indices the user has read\n",
|
||
" all_articles: List of all available articles\n",
|
||
" n_recommendations: Number of recommendations to return\n",
|
||
"\n",
|
||
" Returns:\n",
|
||
" List of recommended article indices (excluding already read)\n",
|
||
"\n",
|
||
" \"\"\"\n",
|
||
" # Your code here:\n",
|
||
" # 1. Get embeddings for articles in user history\n",
|
||
" user_embeddings = embedding_model.embed_documents([all_articles[i] for i in user_history])\n",
|
||
" # 2. Compute \"user profile\" (average embedding of read articles)\n",
|
||
" user_profile = np.mean(user_embeddings, axis=0)\n",
|
||
" # 3. Find articles most similar to user profile\n",
|
||
" all_embeddings = embedding_model.embed_documents(all_articles)\n",
|
||
" similarities = [cosine_similarity(user_profile, emb) for emb in all_embeddings]\n",
|
||
" # 4. Exclude already-read articles\n",
|
||
" recommendations = [(i, similarities[i]) for i in range(len(all_articles)) if i not in user_history]\n",
|
||
" # 5. Return top N recommendations\n",
|
||
" return [idx for idx, _ in sorted(recommendations, key=lambda x: x[1], reverse=True)[:n_recommendations]]\n",
|
||
"\n",
|
||
"# Test: User has read articles about ML/AI\n",
|
||
"user_read = [0, 2] # Machine Learning and Neural Networks articles\n",
|
||
"recommendations = recommend_articles(user_read, articles, n_recommendations=3)\n",
|
||
"\n",
|
||
"print(\"User has read:\")\n",
|
||
"for idx in user_read:\n",
|
||
" print(f\" • {articles[idx]}\")\n",
|
||
"\n",
|
||
"print(\"\\nRecommended articles:\")\n",
|
||
"for idx in recommendations:\n",
|
||
" print(f\" • {articles[idx]}\")\n"
|
||
]
|
||
}
|
||
],
|
||
"metadata": {
|
||
"colab": {
|
||
"provenance": []
|
||
},
|
||
"kernelspec": {
|
||
"display_name": "studies (3.13.9)",
|
||
"language": "python",
|
||
"name": "python3"
|
||
},
|
||
"language_info": {
|
||
"codemirror_mode": {
|
||
"name": "ipython",
|
||
"version": 3
|
||
},
|
||
"file_extension": ".py",
|
||
"mimetype": "text/x-python",
|
||
"name": "python",
|
||
"nbconvert_exporter": "python",
|
||
"pygments_lexer": "ipython3",
|
||
"version": "3.13.9"
|
||
}
|
||
},
|
||
"nbformat": 4,
|
||
"nbformat_minor": 4
|
||
}
|