Mathematical solvers use parametrized Optimization Problems (OPs) as inputs to yield optimal decisions. In many real-world settings, some of these parameters are unknown or uncertain. Recent research focuses on predicting the value of these unknown parameters using available contextual features, aiming to decrease decision regret by adopting end-to-end learning approaches. However, these approaches disregard prediction uncertainty and therefore make the mathematical solver susceptible to provide erroneous decisions in case of low-confidence predictions. We propose a novel framework that models prediction uncertainty with Bayesian Neural Networks (BNNs) and propagates this uncertainty into the mathematical solver with a Stochastic Programming technique. The differentiable nature of BNNs and differentiable mathematical solvers allow for two different learning approaches: In the Decoupled learning approach, we update the BNN weights to increase the quality of the predictions' distribution of the OP parameters, while in the Combined learning approach, we update the weights aiming to directly minimize the expected OP's cost function in a stochastic end-to-end fashion. We do an extensive evaluation using synthetic data with various noise properties and a real dataset, showing that decisions regret are generally lower (better) with both proposed methods.

Artificial intelligence, machine learning, and deep learning solutions are some of the hottest buzzwords in today's corporate landscape. These technologies are reshaping the corporate landscape with their capability to provide innovative solutions to some long-standing problems. In today's quickly-evolving corporate landscape, companies must often engage in intense competition to secure users and customers. In the age of big data and in-depth analysis of customer behavior, artificial intelligence (AI) and machine learning (ML) solutions are emerging as the de facto way for companies to gain a competitive edge. Today, it is easier to harvest large amounts of data from the customer. The advancement of the AI field has resulted in the creation and adoption of machine learning. Machine learning was then discovered to be a good fit for the corporate landscape, providing cost-effective solutions to problems that previously required a lot of resources.

Artificial intelligence solutions are inundating the market right now: They're penetrating and transforming industries from transportation and agriculture to finance (paywall) and business development. And sales and marketing are no exception to this trend. Many tout AI's unique ability to recognize and track nuances that a human eye can't as the "golden ticket" that promises to skyrocket any company's attempts to understand their customers on a whole new level. I believe AI essentially presents companies with an opportunity to establish longer-lasting, deeper and more meaningful connections with their customers. How does this happen, exactly? At its very core, artificial intelligence operates as a brain.