A continuous relaxation of the acquisition process is introduced that enables non-myopic pathwise policy gradients (NM-PPG) through the full acquisition trajectory, avoiding the high variance of standard score-function policy gradients while allowing end-to-end optimization of the acquisition policy.
Linus Aronsson, M. Chehreghani· arXiv.org· 3 citations
This work splits the routing policy into a node permutation stage and an edge selection stage, and introduces a pre-encoding edge aggregation scheme and a non-autoregressive architecture for the edge stage, as well as a hierarchical reinforcement learning method to train the stages jointly.
Filip Rydin, M. Chehreghani, Balázs Kulcsár· arXiv.org· 0 citations
AFA with Incomplete Training Data is studied, showing that under missing completely at random (MCAR) data, one-step acquisition values remain unchanged, whereas multi-step values can decrease and that generative restoration is able to recover lost performance in many experiments.
Reza Rezvan, Valter Schütz, Han Wu et al.· 0 citations
The results show that Bayesian linear regression based on this physics-aware model can improve the reliability of the expected energy consumption, as compared with standard linear regression, and it is shown that more complex machine learning models such as neural networks and gradient boosted regression trees can furt...
Hannes Nilsson, R. Basso, Bal'azs Kulcs'ar et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
