How deeply does Deep Learning with R, Third Edition cover data curation?

Connects curation to manifold coverage and interpolation, but stops short of a procedural curation workflow.

Data curation

As explained in Deep Learning with R, Third Edition

Data curation is the selection, cleaning, organization, and quality management of data for analysis or machine learning.

Data curation selects, cleans, organizes, and quality-checks data so models learn from informative, low-noise examples. For deep learning, better data often improves generalization more reliably than changing the model. section 5.1.2

Prioritize data that makes the learning problem easier to interpolate:

Increase coverage of the input space, especially near decision boundaries. section 5.1.2
Preserve informative features and remove overly noisy or inaccurate data. section 5.1.2

Randomly shuffled MNIST labels

An MNIST classifier is trained after randomly shuffling the training labels. The runnable experiment keeps the images but destroys the input–label relationship, illustrating why model capacity cannot compensate for poor curation.

How Deep Learning with R, Third Edition covers it

Cover of Deep Learning with R, Third Edition

liveBook manning.com

Read Deep Learning with R, Third Edition on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Introduction to Generative AI, Second EditionDataset curation is the process of selecting, organizing, documenting, reviewing, maintaining, and quality-controlling data for a particular purpose.

Compare how 2 books explain data curation →

Related concepts

rejection samplingincludes · Rejection sampling is a method that generates candidate samples and accepts those meeting a specified criterion while discarding the others.
pipelinepart of · A pipeline is an ordered or coordinated sequence of connected processing stages or tasks that passes data, work, or results toward an outcome.
model architecturecontrasts with · The structural organization of a computational system’s components, relationships, and operations that determines how it processes information or behaves.
datasetuses · A dataset is an organized collection of related data items, records, or observations treated as a unit for storage, processing, analysis, or computation.
fine-tuninguses · Fine-tuning is adapting a pretrained machine-learning model by training it further on targeted data or objectives for a task, domain, or behavior.
lifecyclepart of · A lifecycle is the sequence of stages or states an entity, process, or artifact passes through during its existence.

This explainer was generated by Manning's knowledge graph from the text of Deep Learning with R, Third Edition. The section links point to the parts of the book it draws on.