SYSTEMS AND METHODS FOR CUSTOMIZING MACHINE LEARNING MODELS FOR PERMITTING DIFFERENT TYPES OF INFERENCES

• Main Authors: COX, Henrik Thomas Oftedahl, KELLY, Samuel James, DEHGAN, Alex, BUNJE, Paul, GALLINAT, Chad Stephen, COMBE, Patrick Edward Charles
• Format: Patent
• Language: English; French; German
• Published: 05.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

A process for facilitating environmental conservation is described. The process includes: (i) creating a model for describing phenomena relevant to environmental conservation; (ii) training the model using machine-learning to produce a candidate model; (iii) determining whether the candidate model satisfies predetermined model statistics, and if not, repeating previous step until they are satisfied; (iv) deeming the candidate model as a deployable model if the predetermined model statistics are satisfied; (v) deploying the candidate model to permit an inference; (vi) determining whether the inference satisfies predefined inference criteria; (vii) deeming the deployable model as the final model if the predefined inference criteria are satisfied, and if not, modifying the deployable model until they are satisfied to produce the final model; (viii) implementing the final model on an AI adapter device remote to the user to draw an inference; and (ix) taking action using the AI adapter device to facilitate environmental conservation.