Semi-supervised learning
Semi-supervised learning is a machine-learning approach that trains models using both labeled and unlabeled data.
Outlier Detection in Python
Semisupervised learning is a machine-learning paradigm that trains with labeled data while also exploiting substantially larger amounts of unlabeled data.
- How it is used
- Improves scoring of familiar, relevant outliers after labeled records become available in an outlier detection system.
- How deep it goes
- Defines the approach and its place in a detection workflow, but does not specify an algorithm or implementation.
- Examples
- Illustrates label collection with flagged records and records just below detection thresholds in ongoing or one-off projects.
- What you can do after
- Evaluate/understand it as a way to improve familiar-outlier scores without abandoning unsupervised detection of novel records.
Read how Outlier Detection in Python explains semi-supervised learning →
Financial AI in Practice
Semi-supervised learning uses a limited amount of labeled data together with a larger pool of unlabeled data.
- How it is used
- Addresses scarce fraud labels and class imbalance by combining labeled fraud data with a larger pool of unlabeled transactions.
- How deep it goes
- Defines the approach and outlines self-training and co-training procedures, without specifying an implementation or evaluation protocol.
- Examples
- Uses fraud transactions to illustrate how high-confidence predictions can expand a labeled set.
- What you can do after
- Evaluate/understand when scarce fraud labels and abundant unlabeled transactions make self-training or co-training worth considering.
Read how Financial AI in Practice explains semi-supervised learning →
Causal AI
Semi-supervision is a learning setting that combines labeled and unlabeled examples to train a model.
- How it is used
- Combines labeled and unlabeled examples to improve prediction when labeling is costly and causal structure supports the approach.
- How deep it goes
- Explains data roles, distributional motivation, and causal success conditions without detailing a specific training algorithm.
- Examples
- Uses unlabeled digit images in a VAE-based causal model and violence-image moderation with costly human labels.
- What you can do after
- Evaluate whether unlabeled examples are likely to improve prediction by examining causal direction and labeling costs.
Graph Neural Networks in Action
Semi-supervised learning is a learning setting that combines labeled and unlabeled examples during model development.
- How it is used
- Uses a small labeled subset of graph nodes to train classifiers and infer labels for the unlabeled majority.
- How deep it goes
- Details label encoding, masking, embedding alignment, and classifier training, but does not derive the methods mathematically.
- Examples
- Uses the Political Books co-purchase graph, political-orientation labels, and comparisons of embedding-based and GNN classifiers.
- What you can do after
- Apply Boolean masks to train classifiers on labeled nodes and predict labels for the remaining graph nodes.
Read how Graph Neural Networks in Action explains semi-supervised learning →
Knowledge Graphs and LLMs in Action
Semi-supervised learning is a machine-learning paradigm that learns from a combination of labeled and unlabeled examples.
- How it is used
- Uses graph structure alongside partial labels to improve node classification and link prediction.
- How deep it goes
- Defines the paradigm and its graph-specific limitation but stops short of algorithmic implementation.
- Examples
- Uses Flickr users, social connections, interest groups, and a node-classification training-prediction flow.
- What you can do after
- Evaluate when unlabeled graph nodes can provide useful structural information despite missing labels.
Read how Knowledge Graphs and LLMs in Action explains semi-supervised learning →
Related concepts
- unsupervised learninguses · Unsupervised learning is a machine-learning approach that discovers patterns, structure, or representations in data without labeled target outputs.
- supervised learninguses · Supervised learning is a machine-learning approach that learns to map inputs to target outputs from examples paired with known labels or values.
- labeled dataused by · Labeled data is a set of examples or observations paired with target values, categories, or other task-relevant annotations.
- node classificationused by · Node classification is a graph-learning task that predicts one or more discrete category labels for individual nodes in a graph.
- link predictionused by · Link prediction is a graph-learning task that estimates whether a connection exists or may occur between nodes, often by scoring candidate links.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.