🚨 Limited Offer: First 50 users get 500 credits for free — only ... spots left!
Free Artificial Intelligence & Machine Learning flashcards, exportable to Notion
Learn faster with 47 Artificial Intelligence & Machine Learning flashcards. One-click export to Notion.
Learn fast, memorize everything, master Artificial Intelligence & Machine Learning. No credit card required.
Want to create flashcards from your own textbooks and notes?
Let AI create automatically flashcards from your own textbooks and notes. Upload your PDF, select the pages you want to memorize fast, and let AI do the rest. One-click export to Notion.
Artificial Intelligence & Machine Learning
47 flashcards
Transfer learning is a technique that involves using knowledge gained from one task to improve learning performance on a related task, often by reusing parts of a pre-trained model.
Explainable AI refers to techniques and methods that aim to make machine learning models more interpretable and their decisions more understandable to humans.
Supervised learning is a type of machine learning where the algorithm learns from labeled training data to produce a mapping function that can make predictions on new, unseen data.
Unsupervised learning is a type of machine learning where the algorithm tries to find patterns and relationships in unlabeled data without any guidance.
Reinforcement learning is a type of machine learning where an agent learns by trial-and-error to make decisions and take actions in an environment to maximize a reward signal.
A neural network is a machine learning model inspired by the human brain, consisting of interconnected nodes that process information and learn to perform tasks like classification or prediction.
Backpropagation is an algorithm used to train artificial neural networks by calculating gradients and adjusting weights to minimize the output error signal.
Overfitting occurs when a model learns the training data too well, including the noise, and fails to generalize well to new, unseen data.
Regularization is a technique used to prevent overfitting by adding a penalty term to the loss function to constrain the model complexity.
Classification is a machine learning task that predicts discrete, categorical outputs, while regression predicts continuous, numerical outputs.
Ensemble learning combines multiple machine learning models to improve predictive performance and robustness compared to a single model.
A decision tree is a machine learning model that makes predictions by traversing a tree-like structure based on feature values, with nodes representing tests on features and leaves representing outcomes or class labels.
Gradient descent is an optimization algorithm used to find the minimum of a function by iteratively adjusting parameters in the direction of the negative gradient.
The bias-variance tradeoff describes the relationship between a model's ability to minimize bias (learning the true signal) and variance (sensitivity to fluctuations in the training data).
Dimensionality reduction refers to techniques for reducing the number of features in a dataset while retaining most of the relevant information, often to improve model performance or visualization.
The curse of dimensionality refers to the challenges that arise when analyzing high-dimensional data, such as increased sparsity and the need for more training data as the number of dimensions grows.
An SVM is a supervised learning algorithm that finds the optimal hyperplane that separates data points of different classes with the maximum margin.
K-means clustering is an unsupervised learning algorithm that partitions data into K clusters by iteratively assigning data points to the nearest cluster centroid and updating the centroids.
PCA is a dimensionality reduction technique that transforms data into a new coordinate system of linearly uncorrelated variables called principal components, ordered by decreasing variance.
A generative model learns the joint probability distribution of inputs and outputs, while a discriminative model learns the conditional probability of outputs given inputs directly.
A CNN is a type of neural network designed for processing grid-like data such as images, using convolutional layers to detect local patterns and pooling layers to reduce spatial dimensions.
An RNN is a type of neural network designed to process sequential data such as text or time series data, with connections between nodes that form a directed cycle, allowing it to exhibit dynamic temporal behavior.
LSTM is a type of RNN architecture designed to model long-range dependencies more effectively by incorporating memory cells and gates that regulate information flow.
The attention mechanism is a technique used in neural networks to focus on the most relevant parts of the input data when making predictions, by assigning weights to different parts of the input.
A GAN is a type of generative model consisting of two neural networks competing against each other, with one network generating samples and the other discriminating between real and generated samples.
The AI control problem refers to the challenge of ensuring that advanced AI systems behave in alignment with human values and intentions, and remain under meaningful human control.
The Turing test is a test proposed by Alan Turing to determine if a machine can exhibit intelligent behavior equivalent to or indistinguishable from a human.
The AI alignment problem refers to the challenge of ensuring that advanced AI systems reliably pursue goals and behaviors that are aligned with human values and intentions.
A genetic algorithm is a search and optimization technique inspired by natural selection and evolution, which generates solutions to a problem by iteratively applying bio-inspired operators like mutation and crossover to a population of candidate solutions.
Simulated annealing is a probabilistic optimization algorithm that is inspired by the annealing process in metallurgy, where it tries to find the global optimum of a given function by temporarily accepting worse solutions to avoid getting stuck in local optima.
The TSP is a classic optimization problem in computer science and operations research, where the goal is to find the shortest possible route that visits a set of cities exactly once and returns to the starting point.
The knapsack problem is a combinatorial optimization problem where the goal is to fill a knapsack with the most valuable items without exceeding the knapsack's weight capacity.
The minimax algorithm is a recursive decision rule used in game theory and decision-making to minimize the maximum possible loss, assuming the opponent is playing optimally.
A* is an informed search algorithm used for finding the shortest path between nodes in a graph, by combining information from the heuristic function and the cost function.
DFS explores as far as possible along each branch before backtracking, while BFS explores all nodes at the present depth before moving on to nodes at the next level.
Adversarial machine learning is a field that studies how to make machine learning models robust against adversarial attacks, where an adversary intentionally introduces perturbations to input data to mislead the model.
Federated learning is a decentralized machine learning approach where models are trained on distributed data across multiple devices or silos, without directly sharing the data.
Few-shot learning refers to machine learning techniques that aim to learn new concepts from very few examples, by leveraging prior knowledge and meta-learning.
Imitation learning is a machine learning paradigm where an agent learns to perform a task by observing demonstrations from another agent or expert, rather than through explicit instructions or reward signals.
Inverse reinforcement learning is a technique used to infer the reward function or preferences of an expert agent by observing its behavior, with the goal of learning a policy that mimics the expert's behavior.
Multi-agent reinforcement learning is a branch of reinforcement learning that focuses on developing algorithms and techniques for multiple agents to learn and interact in a shared environment, often with mixed cooperative and competitive objectives.
Hierarchical reinforcement learning is an approach that decomposes a complex task into a hierarchy of subtasks, where higher-level policies guide the selection of lower-level policies or skills to solve the overall task more efficiently.
Meta-learning, also known as 'learning to learn', refers to techniques that enable machine learning models to adapt and learn new tasks more efficiently by leveraging knowledge acquired from previous learning experiences.
Self-supervised learning is a machine learning paradigm where the model learns representations from the data itself, without relying on explicit labels, by solving pretext tasks or auxiliary prediction problems.
Contrastive learning is a self-supervised learning approach that trains models to learn representations by pulling together similar data points (positive pairs) and pushing apart dissimilar data points (negative pairs) in the embedding space.
The explore-exploit dilemma refers to the tradeoff between exploring unfamiliar actions to gather more information, and exploiting current knowledge by taking actions that are expected to yield high rewards.
Reward shaping is a technique used in reinforcement learning to modify the reward signal by adding additional rewards or penalties, with the goal of guiding the agent towards desired behaviors and improving learning efficiency.