Inference model training method and device, electronic device and storage medium

Abstract

The embodiment of the invention discloses an inference model training method and device, an electronic device and a storage medium, and the inference model training method comprises the steps: obtaining the current question state of a current question by aiming at the current question in a training sample; selecting a current to-be-executed action according to the current question state; executingthe current to-be-executed action to obtain a new topic state; taking the new topic state as the current topic state of the current topic; according to the technical scheme, the inference model can continuously accumulate the problem solving experience, the inference accuracy is optimized, and the trained inference model can achieve automatic inference of complex elementary mathematical problemsin actual use.

Description

technical field [0001] The present disclosure relates to the technical field of the Internet, and in particular to a training method, device, electronic equipment and storage medium of an inference model. Background technique [0002] Automatic reasoning is one of the core technologies for machine problem-solving. It is based on the given conditions, problems and a series of actions (inference rules) defined inside the system, and continuously changes the conditions and problems by executing actions until the new conditions are met. A new problem is to find a feasible solution path. [0003] Current reasoning techniques are mainly based on forward push, backward push and two-way search reasoning methods. The main problem of these three reasoning methods is that the search space grows geometrically with the increase of the number of conditions, questions and reasoning rules, so it can only be used to solve some simple reasoning problems. Although in some specific fields, it...

AI Technical Summary

Problems solved by technology

The main problem of these three reasoning methods is that the search space grows geometrically with the increase of the number of conditions, questions and reasoning rules, so it can only solve some simple reasoning problems with

However, these methods are limited to their own fields and cannot be promoted. At the same time, these methods rely on specific mechanical methods and strategies, and cannot achieve the same flexibility as human reasoning.

Images