We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Learning to Optimize in Swarms

Abstract: Learning to optimize has emerged as a powerful framework for various optimization and machine learning tasks. Current such "meta-optimizers" often learn in the space of continuous optimization algorithms that are point-based and uncertainty-unaware. To overcome the limitations, we propose a meta-optimizer that learns in the algorithmic space of both point-based and population-based optimization algorithms. The meta-optimizer targets at a meta-loss function consisting of both cumulative regret and entropy. Specifically, we learn and interpret the update formula through a population of LSTMs embedded with sample- and feature-level attentions. Meanwhile, we estimate the posterior directly over the global optimum and use an uncertainty measure to help guide the learning process. Empirical results over non-convex test functions and the protein-docking application demonstrate that this new meta-optimizer outperforms existing competitors.
Comments: Accepted to Neural Information Processing Systems (NeurIPS2019)
Subjects: Machine Learning (cs.LG); Biomolecules (q-bio.BM); Machine Learning (stat.ML)
Cite as: arXiv:1911.03787 [cs.LG]
  (or arXiv:1911.03787v2 [cs.LG] for this version)

Submission history

From: Yue Cao [view email]
[v1] Sat, 9 Nov 2019 22:25:05 GMT (542kb,D)
[v2] Sat, 16 Nov 2019 19:16:42 GMT (542kb,D)

Link back to: arXiv, form interface, contact.