{"slug": "ml-fundamentals", "title": "ML Fundamentals", "summary": "Machine learning, deep learning, and reinforcement learning are three distinct approaches that enable computers to learn tasks without explicit instructions. Machine learning uses algorithms like linear regression and decision trees to find parameters that minimize prediction loss from training data, while deep learning employs neural networks with activation functions to learn optimal features and approximate complex functions. Reinforcement learning differs by having computers learn through trial and error rather than from data, expanding the ways artificial intelligence can model and predict the world.", "body_md": "# ML Foundations\n\n* Intelligence* require understanding how the world works. We\n\n**Model** the world to make prediction, i.e a\n\n**Model** is something that lets us make prediction.\n\nOur brain is a model, remember Mental Model?\n\nHow can we make these models? i.e organisms that can predict something. Turns out Computers are good base for it. We can leverage Computer to predict via making * Software Models*.\n\n3 Difference Ways Computers can * Learn*:\n\n- Machine Learning\n- Deep Learning\n- Reinforcement Learning\n\nComputer Learning: Computer can do things without explicit instructions.\n\n## ML: Machine Learning\n\nAllow computers to learn tasks directly from data.\n\nGiven\n\n```\nTraining Input Data\nTraining Expected Ouput Data\n```\n\nWe need to find parameters, I something think about parameters as knobs (see [Perceptron](https://www.youtube.com/watch?v=l-9ALe3U-Fg)), that will help the model predict the closest values to the expected values\n\n```\nLoss = Expected Value - Predictoin Value\n```\n\nThe Goal of training phase is to find such knobs that will give lower loss.\n\nLoss value can be negative is prediction value is bigger then expected value, so lets square it.\n\n```\nLoss = (Expected Value - Prediction Value) ^ 2\nLoss = (Y - X0) ^ 2\n```\n\nThe better models fits the * Prediction to Reality* very well.\n\n### ML Techniques:\n\n- Linear Regression\n- Logistics regression\n- Decision Tree\n- Random Forest\n- XGBoost\n- SVM (Support Vector Machine)\n\nSo **Machine learning** is learning from Data, data points are also called * features*, i.e if a model takes 4 inputs, then it has 4 features.\n\nFor example, in our linear regression simple example given above, our model was:\n\n```\nY = mX + c\n```\n\ni.e it has 1 feature.\n\nWhat if we have multiple features (inputs) then are we giving equal wattage (weights) to each features, is one feature more important then other?\n\n## DL: Deep Learning\n\nNeural Networks that learn * optimal* features on their own.\n\nRemember:\n\nFeature is the data (input information)\n\nParameters are weights that are learned.\n\n### Neural Networks (NN)\n\nIn Deep Learning we use Neural Networks to learn optimal features,\n\nwhich features are more important then others.\n\nA series of operations that can approximate (practically) any function.\n\nThe `Sum(wixi) + b`\n\nis also seen in Linear regression. but the `g`\n\nthe Activation function is a non-linear function which gives the neural network the ability to approximate any function.\n\n### Activation Function:\n\nActivation Functions is a **Decision Maker **of a neural network. It sits at the end of a neuron, and decides whether that neural should “fire” (pass information forward) or stay dormant.\n\nThis non-linearity of Activation function make the neural network learn complex decision boundary, since it forces the model to become non-linear.\n\nArtificial neural networks are inspired by the human brain. In your brain, a biological neuron receives electrical signals from neighbors. It doesn't just pass every tiny spark along; it waits until the signal is strong enough to cross a threshold, and then it \"fires\" an action potential.\n\nThese types of Activation functions are used in different types of Neural networks:\n\n**Vanilla**: Weighted Sum + Activation. FFNNs and CNNs.** LSTM**: Long Short Term Memory, advance neuron with memory and gates for long-term dependencies in sequences.\n\nTypes of layers structure:\n\nRNNs, CNNs are type of neural networks because they have different type of Layer structure.\n\n[Attention](https://arxiv.org/abs/1706.03762) is also a type of neural network layer structure.\n\n### Types of Networks:\n\nHyperparameters:\n\nValues that guide the training process\n\n## Reinforcement Learning\n\nComputer learn from trail and error, not from data.\n\nReferences:\n\n[← Back to All Blogs](/blog)", "url": "https://wpnews.pro/news/ml-fundamentals", "canonical_source": "https://knl.co.in/blog/ml-foundations/", "published_at": "2026-06-06 10:30:45+00:00", "updated_at": "2026-06-06 11:18:22.507730+00:00", "lang": "en", "topics": ["machine-learning"], "entities": [], "alternates": {"html": "https://wpnews.pro/news/ml-fundamentals", "markdown": "https://wpnews.pro/news/ml-fundamentals.md", "text": "https://wpnews.pro/news/ml-fundamentals.txt", "jsonld": "https://wpnews.pro/news/ml-fundamentals.jsonld"}}