AI Dictionary › AI Fundamentals
Memoization is an optimization in which a function remembers the results of calls it has already performed: when invoked again with the same inputs, it returns the stored result instead of recomputing it. It is caching applied to functions, and it shines when computation is expensive and inputs repeat frequently.
The mechanism relies on a lookup table (typically a hash map) that maps a function's arguments to its output. On every call the function checks the table: if the input-result pair exists, it reuses it; otherwise it computes, stores and returns. It only works reliably with pure functions, meaning functions that always produce the same output for the same input and have no side effects.
In AI, memoization appears in agent frameworks that avoid re-running tool calls already executed with identical parameters, in evaluation systems that reuse previously computed scores, and in feature engineering pipelines. In classic programming it is the technique that makes dynamic programming tractable, from computing Fibonacci numbers to aligning DNA sequences in bioinformatics.
The term was coined by Donald Michie, the British AI pioneer, in his 1968 Nature paper Memo Functions and Machine Learning. It derives from the Latin memorandum, a thing to be remembered: Michie deliberately chose memoization rather than memorization to describe turning a function into a memo function that remembers its own results.
Grace's problem solving trains you to recognize when a computation needs to be redone and when it can be reused: the same principle as memoization, applied to your AI workflows. The daily gym with its 0-100 score shows how much efficiency you gain with each iteration.
From our network
Kaimaki Web: Websites That Win Customers
Custom websites, web apps and digital marketing for growing businesses.
Visit kaimakiweb.com →From the Agora Intelligence blog
📱 Download the Android app (beta) iOS coming soon
Say what you mean. Get what you need.
Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.