Decision-Focused Learning
PulseAugur coverage of Decision-Focused Learning — every cluster mentioning Decision-Focused Learning across labs, papers, and developer communities, ranked by signal.
-
New DFL method for MDPs tackles scalability with occupancy measure approach
Researchers have developed a new approach to decision-focused learning (DFL) for Markov decision processes (MDPs) that addresses scalability limitations of existing methods. By reformulating the MDP as an occupancy meas…
-
AutoEnergy algorithm boosts energy forecasting accuracy and speed
A new thesis introduces AutoEnergy, a domain-tailored automated feature engineering algorithm designed to improve energy consumption forecasting. This algorithm reduces reliance on expert-driven feature design by genera…
-
Jacobian Rank Collapse identified in decision-focused learning
Researchers have identified a phenomenon called Jacobian Rank Collapse in decision-focused learning (DFL), where the predictor's Jacobian matrix exhibits low rank. This condition means that the gradients used by the opt…
-
New Adversarial DFL method improves network interdiction game learning
Researchers have developed a new method called Adversarial DFL (A-DFL) to address a fundamental failure in Decision-Focused Learning (DFL) when applied to network interdiction games. In these games, an interdictor stren…
-
New DFL method boosts efficiency and scalability for complex problems
Researchers have developed a new method for decision-focused learning (DFL) that significantly improves efficiency and scalability. The approach reframes the problem as cost-sensitive multi-output regression, incorporat…
-
New method projects prediction error onto tangent space for decision-focused learning
Researchers have developed a new method called PEAR (Projected Error As Regret-gradient) for Decision-Focused Learning. This technique simplifies the computation of regret gradients by projecting prediction errors onto …