Machine Learning and AWS AI Services Lecture Review

A comprehensive set of vocabulary flashcards covering basic machine learning concepts, model training processes, deep learning architectures, and specific AWS AI/ML services.

Labeled data

A dataset where each instance or example is accompanied by a label or target variable that represents the desired output or classification.

Unlabeled data

A dataset where the instances or examples do not have any associated labels or target variables, consisting only of input features.

Structured data

Data that is organized and formatted in a predefined manner, typically in the form of tables or databases with rows and columns.

Tabular data

Data stored in spreadsheets, databases, or CSV files, with rows representing instances and columns representing features.

Time-series data

A type of data consisting of sequences of values measured at successive points in time, such as sensor readings or stock prices.

Unstructured data

Data that lacks a predefined structure or format, such as text, images, audio, and video.

Supervised learning

A machine learning process where algorithms are trained on labeled data to learn a mapping function that can predict output for new input data.

Unsupervised learning

Algorithms that learn from unlabeled data to discover inherent patterns, structures, or relationships within the input data.

Reinforcement learning

A learning process where the machine is given a performance score as guidance and learns from feedback in the form of rewards or penalties.

Semi-supervised learning

A variation of machine learning where only a portion of the training data is labeled.

Inferencing

The process of using information that a model has learned to make predictions or decisions after it has been trained.

Batch inferencing

When a computer takes a large amount of data and analyzes it all at once to provide a set of results.

Real-time inferencing

When a computer makes decisions quickly in response to new information as it comes in, such as in chatbots or self-driving cars.

Deep learning

A field inspired by the structure and function of the brain that involves the use of artificial neural networks.

Neural networks

Computational models designed to mimic the brain, consisting of units called nodes organized into input, hidden, and output layers.

Nodes

Tiny units within neural networks that are connected together and organized into layers to identify patterns.

Computer vision

A field of artificial intelligence that makes it possible for computers to interpret and understand digital images and videos.

Natural language processing (NLP)

A branch of artificial intelligence that deals with the interaction between computers and human languages.

Foundation models (FMs)

Models pretrained on internet-scale data that can be adapted to perform multiple tasks like text generation or summarization.

Self-supervised learning

A pre-training method for FMs that does not require labeled examples and uses the structure within the data to autogenerate labels.

Continuous pre-training

A stage in the model lifecycle where a model is further pre-trained on additional data to expand its knowledge base.

Prompt engineering

The process of developing, designing, and optimizing instructions (prompts) to guide and enhance the output of foundation models.

Large language models (LLMs)

Models trained on vast amounts of text data, commonly based on the transformer architecture, designed to generate human-like text.

Tokens

The basic units of text, such as words, phrases, or individual characters, that an LLM processes.

Diffusion models

A deep learning architecture that starts with random noise and gradually adds meaningful information through forward and reverse steps.

Multimodal models

Models that can process and generate multiple modes of data simultaneously, such as taking in an image and text to generate a caption.

Generative adversarial networks (GANs)

A generative model framework involving two neural networks, a generator and a discriminator, competing against each other.

Variational autoencoders (VAEs)

A generative model that uses an encoder to map data to a latent space and a decoder to reconstruct the original data.

Fine-tuning

A supervised learning process that involves taking a pre-trained model and adding specific, smaller datasets to modify model weights.

Retrieval-augmented generation (RAG)

A technique that supplies domain-relevant data as context for user prompts without changing the foundation model's weights.

Amazon SageMaker AI

A toolset offered by AWS to build, train, and run LLMs and other foundation models efficiently with managed infrastructure.

Amazon Comprehend

An AI service that uses ML and NLP to uncover insights and relationships in unstructured data.

Amazon Translate

A neural machine translation service that uses deep learning models to deliver natural-sounding translations.

Amazon Textract

A service that automatically extracts text, data from forms, and information from tables in scanned documents.

Amazon Lex

A managed AI service for building conversational interfaces using automatic speech recognition and natural language understanding.

Amazon Polly

An AI service that uses deep learning technologies to synthesize speech that sounds like a human voice from text.

Amazon Transcribe

An automatic speech recognition (ASR) service for converting speech to text from audio files or live streams.

Amazon Rekognition

A service that facilitates adding image and video analysis, including facial analysis and object detection, to applications.

Amazon Kendra

An ML-powered intelligent search service that allows users to find content scattered across multiple enterprise locations.

Amazon Personalize

An ML service used by developers to create individualized recommendations for articles, products, or videos for their customers.

AWS DeepRacer

A 1/18th scale race car used as a fun way to get started with reinforcement learning.

Amazon Bedrock

A fully managed service that makes foundation models from leading AI startups and Amazon available via an API.

Amazon Q

A generative AI assistant that provides answers, solves problems, and generates content using company information repositories.

Amazon Q Developer

A service that provides ML-powered code recommendations for languages like Python, Java, and JavaScript to improve developer productivity.