Loading patterns…
Continual Learning(CL)
Learning from a sequence of tasks or data distributions while measuring and mitigating loss of earlier capabilities
In 30 seconds
- What
- Trains sequentially on new tasks while measuring and preventing loss of performance on earlier tasks through rehearsal, regularization, or task-specific adapters.
- When to use
- Systems must learn from multiple tasks or data distributions over time without forgetting prior capabilities or requiring access to all historical data.
- Watch out
- Measuring true catastrophic forgetting requires maintaining a full evaluation matrix across all tasks at every stage, which becomes expensive as task count grows.
Ask the AI expert about this pattern
Opens the assistant with your question prefilled. You review it before sending.
Continual Learning: Overview
Learning from a sequence of tasks or data distributions while measuring and mitigating loss of earlier capabilities
- Catastrophic forgetting measurement
- Sequential task learning
- Regularization such as EWC
- Rehearsal and replay
- Task-specific adapters
- Forward and backward transfer evaluation
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.
References
The papers, specifications, and repositories this pattern is based on.
- Overcoming catastrophic forgetting in neural networks (2017)arXiv:1612.00796
From the engineer behind this catalog
Get your agent architecture reviewed
This page documents one pattern. Your system runs dozens, and most failures live in how they fit together. Have the whole design reviewed against the 288 patterns in this catalog: architecture, reliability, evaluation and cost, every finding mapped to the pattern that fixes it.
€750 instead of €1,500, one week, written report and walkthrough call, until 30 September