SECAI Core · phase 2 of 15 · noun

Federated learning

A machine learning approach where a central model is trained collaboratively across many decentralized devices or servers using their local data, so the raw data never leaves the devices—only model updates (like gradients or parameters) are shared and aggregated to improve the global model while enhancing privacy.

The Explain card

Plain English
Federated learning trains one shared model across many devices or servers. Each participant trains on its own local data and sends back only model updates, never the raw data.
Example
Thousands of phones each improve a keyboard's next-word prediction locally and send gradient updates to a central server that merges them into the global model.
Why it matters
Raw data stays home, which helps privacy, but updates can still leak information about it, and a malicious participant can send poisoned updates to corrupt or backdoor the global model.
Hook
The model travels to the data, not the data to the model.

Word knowledge

How the word is built, where it came from, and what it sits beside in memory.

In a sentence

Each Android phone trains its Gboard next-word model with federated learning, then uploads a weight update that the aggregator averages into the next global model.

Why these words
  • federated Latin foederatus, allied by treaty, from foedus, a league supplies the league adjective, so the learning is of allied members rather than one site
  • learning Old English leornung, the act of getting knowledge supplies the head noun, the classroom activity the phrase assigns to that league
Where it came from
Origin
English compound from Latin foederatus (allied by treaty, from foedus, league) and Old English leornung (the act of getting knowledge); as an ML label, coined by Brendan McMahan and colleagues at Google
Entered the language
2016
What changed
Federated had meant bound by a treaty, then joined as a league of states, then, in computing, a union of systems that keep local control. Learning had meant a person acquiring knowledge. McMahan's 2016 paper joined the two as a research title, and the pair became the ordinary name for that training method.
How it is spelled
Pattern
feder- from Latin foedus keeps the d
Pattern
-ate from Latin -atus stays ate
Pattern
-ed marks the past participle used as an adjective
Pattern
-ing turns a verb into a noun of activity
Breaks the pattern
the noun phrase stays two words; the hyphen appears only when it modifies another noun
Breaks the pattern
federated dropped the Latin oe of foedus, writing e where the source had a diphthong

Spelled like

  • federal
  • confederate
  • generated
  • training
Broken into chunks
  • feder
    • federal
    • federation
    • confederate
  • ate
    • generate
    • separate
    • moderate
  • ed
    • trained
    • marked
    • unexpected
  • learn
    • learner
    • unlearn
    • relearn
  • ing
    • training
    • computing
    • reasoning
What it sits beside

Same subject

  • differential privacy
  • secure aggregation
  • split learning
  • on-device training

Same shape

Learning compounds

Federate family

  • federation
  • federal
  • confederate

Privacy-preserving training

  • differential privacy
  • secure aggregation
  • split learning

Where it sits in the deck

Phase 2: How Models Are Built: Architectures and Learning Mechanics

With the paradigms named, zoom in on the architectural building blocks that turn data into trained artifacts.

Neighbours in this phase