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: $δ$-CLUE: Diverse Sets of Explanations for Uncertainty Estimates

Abstract: To interpret uncertainty estimates from differentiable probabilistic models, recent work has proposed generating Counterfactual Latent Uncertainty Explanations (CLUEs). However, for a single input, such approaches could output a variety of explanations due to the lack of constraints placed on the explanation. Here we augment the original CLUE approach, to provide what we call $\delta$-CLUE. CLUE indicates $\it{one}$ way to change an input, while remaining on the data manifold, such that the model becomes more confident about its prediction. We instead return a $\it{set}$ of plausible CLUEs: multiple, diverse inputs that are within a $\delta$ ball of the original input in latent space, all yielding confident predictions.
Comments: Appeared as a workshop paper at ICLR 2021 (Responsible AI | Secure ML | Robust ML)
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
Cite as: arXiv:2104.06323 [cs.LG]
  (or arXiv:2104.06323v2 [cs.LG] for this version)

Submission history

From: Dan Ley [view email]
[v1] Tue, 13 Apr 2021 16:03:27 GMT (10902kb,D)
[v2] Wed, 14 Apr 2021 08:10:33 GMT (10902kb,D)
[v3] Sat, 24 Apr 2021 14:08:57 GMT (10902kb,D)
[v4] Tue, 27 Apr 2021 15:23:09 GMT (10902kb,D)
[v5] Sat, 8 May 2021 09:29:40 GMT (11008kb,D)
[v6] Fri, 3 Dec 2021 16:52:01 GMT (11008kb,D)

Link back to: arXiv, form interface, contact.