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: HardCoRe-NAS: Hard Constrained diffeRentiable Neural Architecture Search

Abstract: Realistic use of neural networks often requires adhering to multiple constraints on latency, energy and memory among others. A popular approach to find fitting networks is through constrained Neural Architecture Search (NAS), however, previous methods enforce the constraint only softly. Therefore, the resulting networks do not exactly adhere to the resource constraint and their accuracy is harmed. In this work we resolve this by introducing Hard Constrained diffeRentiable NAS (HardCoRe-NAS), that is based on an accurate formulation of the expected resource requirement and a scalable search method that satisfies the hard constraint throughout the search. Our experiments show that HardCoRe-NAS generates state-of-the-art architectures, surpassing other NAS methods, while strictly satisfying the hard resource constraints without any tuning required.
Comments: Niv Nayman and Yonathan Aflalo contributed equally. An implementation of HardCoRe-NAS is available at: this https URL
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Optimization and Control (math.OC); Machine Learning (stat.ML)
MSC classes: 68T09, 68T45
ACM classes: G.1.6; G.3; I.2.8; I.2.10; I.5.1
Cite as: arXiv:2102.11646 [cs.LG]
  (or arXiv:2102.11646v1 [cs.LG] for this version)

Submission history

From: Niv Nayman [view email]
[v1] Tue, 23 Feb 2021 11:56:30 GMT (932kb,D)

Link back to: arXiv, form interface, contact.