Intelligence artificielle adaptative

Définition

XXXXXXXXX

Français

XXXXXXXXX

Anglais

Adaptive AI

Adaptive AI is a method of training artificial intelligence to respond to a variety of situations and improve its performance. It does this by utilizing machine learning, which is the study of patterns in data and learns to act accordingly. Machine learning has different types, such as reinforcement learning, supervised learning, and unsupervised learning. While traditional AI trains an AI system with massive amounts of data collected in a stationary environment, adaptive AI incorporates a retraining process.

For example, financial institutions often develop credit risk models every two years, monitor their performance, and redevelop the models using newer data. The process takes six to 12 months, and during that time, the models are suboptimal. However, with Adaptive AI, lenders are able to deploy new models every month, outperforming their current production.

Adaptive AI is capable of outperforming traditional machine learning models and encourage better results with fewer assets and less effort. Unlike traditional machine learning models, adaptive AI can be used in a dynamic software environment and can play a pivotal role in digital transformation.

Source : limitlessarticles