The part of an AI system that generates answers. An inference engine comprises the hardware and software that provides analyses, makes predictions or generates unique content. In other words, the reason for the AI in the first place.
Human Rules Were the First AI
Years ago and relying entirely on human rules, "expert systems" were the first inference engines. However, the capabilities of today's neural networks and GPT architectures are light years ahead of expert systems.
Inference vs. Training
The inference engine is the processing, or runtime, component of an AI in contrast to the fact gathering or learning side of the system, which uses considerably more computer power. Large language models that take in trillions of data examples can take weeks and months to be fully trained.
After the training phase is over, the inference engine does work for the user (the purpose of the AI). The inference side requires less compute power than the training phase, and it is considerably faster.
A Term With Wiggle Room
The English word "inference" implies assumption and conjecture. Apparently, in the early days of AI, "inferring" an answer seemed a safer bet than "generating" or "processing" the answer, which implies accuracy in the tech world. Thus, even today, an AI does not generate a result; it "infers" the result. See
AI training vs. inference,
AI processing methods,
AI training,
neural network,
GPT,
deep learning and
expert system.