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


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


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

Computer Science > Machine Learning

Title: NP-PROV: Neural Processes with Position-Relevant-Only Variances

Abstract: Neural Processes (NPs) families encode distributions over functions to a latent representation, given context data, and decode posterior mean and variance at unknown locations. Since mean and variance are derived from the same latent space, they may fail on out-of-domain tasks where fluctuations in function values amplify the model uncertainty. We present a new member named Neural Processes with Position-Relevant-Only Variances (NP-PROV). NP-PROV hypothesizes that a target point close to a context point has small uncertainty, regardless of the function value at that position. The resulting approach derives mean and variance from a function-value-related space and a position-related-only latent space separately. Our evaluation on synthetic and real-world datasets reveals that NP-PROV can achieve state-of-the-art likelihood while retaining a bounded variance when drifts exist in the function value.
Comments: 10 pages, 5 figures
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (stat.ML)
Cite as: arXiv:2007.00767 [cs.LG]
  (or arXiv:2007.00767v1 [cs.LG] for this version)

Submission history

From: Xuesong Wang [view email]
[v1] Mon, 15 Jun 2020 06:11:21 GMT (4420kb,D)

Link back to: arXiv, form interface, contact.