Current browse context:
cs.LG
Change to browse by:
References & Citations
Computer Science > Machine Learning
Title: Adaptive Partial Scanning Transmission Electron Microscopy with Reinforcement Learning
(Submitted on 6 Apr 2020 (v1), revised 7 Apr 2020 (this version, v2), latest version 11 Mar 2021 (v8))
Abstract: Compressed sensing is applied to scanning transmission electron microscopy to decrease electron dose and scan time. However, established methods use static sampling strategies that do not adapt to samples. We have extended recurrent deterministic policy gradients to train deep LSTMs and differentiable neural computers to adaptively sample scan path segments. Recurrent agents cooperate with a convolutional generator to complete partial scans. We show that our approach outperforms established algorithms based on spiral scans, and we expect our results to be generalizable to other scan systems. Source code, pretrained models and training data is available at this https URL
Submission history
From: Jeffrey Ede BSc MPhys [view email][v1] Mon, 6 Apr 2020 16:25:38 GMT (2906kb,D)
[v2] Tue, 7 Apr 2020 13:05:52 GMT (2971kb,D)
[v3] Thu, 3 Dec 2020 16:06:00 GMT (3356kb,D)
[v4] Tue, 22 Dec 2020 14:26:08 GMT (3357kb,D)
[v5] Mon, 1 Mar 2021 11:11:30 GMT (3401kb,D)
[v6] Wed, 3 Mar 2021 12:15:56 GMT (3401kb,D)
[v7] Mon, 8 Mar 2021 18:09:48 GMT (3401kb,D)
[v8] Thu, 11 Mar 2021 16:55:09 GMT (3401kb,D)
Link back to: arXiv, form interface, contact.