# Intelligence Reborn

> A digital archive of machine learning, AI concepts, and the evolution of artificial intelligence. Published by 3IVIS GmbH (https://3ivis.com).

Intelligence Reborn is an AI knowledge platform organized into two sections: **Dispatches** (long-form essays on AI engineering and practice) and **The Library** (structured reference entries for core ML/AI concepts). All content is produced by the 3IVIS research team.

## Dispatches

Long-form essays on AI engineering, architecture, and practice.

- [Working Alongside AI: Writing and Coding](https://intelligencereborn.com/blog/working-alongside-ai): A practical framework for collaborating with AI, understanding its strengths, limitations, and how humans can complement it.
- [AI Agents: The Layered Architecture Behind Every Response](https://intelligencereborn.com/blog/ai-agents): Inside every AI agent is a layered system of models, prompts, memory, tools, and control logic. Understanding the architecture changes how you use, trust, and build with these systems.
- [Building AI That Remembers You](https://intelligencereborn.com/blog/ai-agent-memory): Language models are stateless by design. A deep look at the four types of AI memory, how vector retrieval actually works, and why building systems that genuinely remember is harder than it looks.
- [From Loop to Harness: Building Reliable AI Agents by Solving Real Failures](https://intelligencereborn.com/blog/harness-engineering): The eleven-layer architecture inside Build A Harness didn't come from a whiteboard — it came from watching agents fail in specific, reproducible ways. This is the story of each component and the problem that forced it into existence.
- [Coding Agents: How They Actually Work — and When to Trust Them](https://intelligencereborn.com/blog/coding-agents): From autocomplete to autonomous pull requests — a practical guide to the coding agent landscape, how these tools actually work, where they fail, and how to use them without losing control of your codebase.

## The Library

Structured reference entries for core machine learning and AI concepts.

### Foundations
- [Artificial Intelligence](https://intelligencereborn.com/library/artificial-intelligence): The field concerned with building systems capable of intelligent behavior such as reasoning, learning, and decision-making.
- [Machine Learning](https://intelligencereborn.com/library/machine-learning): Algorithms and systems that learn patterns from data to make predictions or decisions.
- [Linear Algebra for ML](https://intelligencereborn.com/library/linear-algebra): Vectors, matrices, and linear transformations underlying most machine learning models.
- [Probability & Statistics](https://intelligencereborn.com/library/probability-statistics): Probabilistic modeling and statistical inference forming the backbone of ML.

### Learning Paradigms
- [Supervised Learning](https://intelligencereborn.com/library/supervised-learning): Learning mappings from inputs to labeled outputs.
- [Unsupervised Learning](https://intelligencereborn.com/library/unsupervised-learning): Discovering structure in unlabeled data.
- [Reinforcement Learning](https://intelligencereborn.com/library/reinforcement-learning): Agents learning through interaction and reward maximization.

### Core Algorithms
- [Linear Regression](https://intelligencereborn.com/library/linear-regression): A simple yet powerful model for predicting continuous values.
- [Logistic Regression](https://intelligencereborn.com/library/logistic-regression): A probabilistic linear classifier for binary outcomes.
- [Decision Trees](https://intelligencereborn.com/library/decision-trees): Tree-based models that split data using interpretable rules.
- [Random Forests](https://intelligencereborn.com/library/random-forests): Ensembles of decision trees for robust predictions.
- [Gradient Boosting](https://intelligencereborn.com/library/gradient-boosting): Sequential ensemble methods that optimize predictive performance.
- [Support Vector Machines](https://intelligencereborn.com/library/support-vector-machines): Margin-based classifiers using kernel methods.
- [K-Means Clustering](https://intelligencereborn.com/library/k-means): A distance-based algorithm for clustering data.

### Deep Learning
- [Neural Networks](https://intelligencereborn.com/library/neural-networks): Multi-layer networks trained via backpropagation.
- [Convolutional Neural Networks](https://intelligencereborn.com/library/convolutional-neural-networks): Neural networks specialized for spatial data.
- [Recurrent Neural Networks](https://intelligencereborn.com/library/recurrent-neural-networks): Neural networks for sequential data.
- [Transformers](https://intelligencereborn.com/library/transformers): Attention-based architectures powering modern language models.

### Optimization & Training
- [Gradient Descent](https://intelligencereborn.com/library/gradient-descent): Iterative optimization of loss functions.
- [Loss Functions](https://intelligencereborn.com/library/loss-functions): Objective functions used to train models.
- [Regularization](https://intelligencereborn.com/library/regularization): Techniques to reduce overfitting and improve generalization.
- [Hyperparameter Tuning](https://intelligencereborn.com/library/hyperparameter-tuning): Strategies for selecting optimal hyperparameters to improve model performance.

### Evaluation & Validation
- [Model Evaluation & Validation](https://intelligencereborn.com/library/model-evaluation): Metrics and strategies for assessing model performance.
- [Bias–Variance Tradeoff](https://intelligencereborn.com/library/bias-variance-tradeoff): The fundamental tradeoff between underfitting and overfitting.
- [Overfitting & Underfitting](https://intelligencereborn.com/library/overfitting-underfitting): Common failure modes where models generalize poorly due to bias or variance.
- [Cross-Validation](https://intelligencereborn.com/library/cross-validation): Techniques for estimating model performance using repeated data splits.
- [Train / Validation / Test Split](https://intelligencereborn.com/library/train-validation-test-split): Standard data partitioning strategy for training and evaluating models.

### Representation Learning
- [Feature Engineering](https://intelligencereborn.com/library/feature-engineering): Transforming raw data into informative features.
- [Dimensionality Reduction](https://intelligencereborn.com/library/dimensionality-reduction): Reducing feature space while preserving structure.
- [Embeddings](https://intelligencereborn.com/library/embeddings): Dense vector representations that capture semantic relationships in data.

### Interpretability
- [Model Interpretability](https://intelligencereborn.com/library/model-interpretability): Techniques for understanding and explaining model predictions.

### Applications
- [Natural Language Processing](https://intelligencereborn.com/library/natural-language-processing): Machine learning techniques for text and language.
- [Computer Vision](https://intelligencereborn.com/library/computer-vision): Learning from images and video.

### Systems
- [Machine Learning Frameworks](https://intelligencereborn.com/library/ml-frameworks): Core libraries and ecosystems for building ML systems.
- [Model Deployment & Monitoring](https://intelligencereborn.com/library/model-deployment-monitoring): Serving, monitoring, and maintaining ML models in production.