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

Download:

Current browse context:

cs.AI

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

Computer Science > Artificial Intelligence

Title: A Flawed Dataset for Symbolic Equation Verification

Authors: Ernest Davis
Abstract: Arabshahi, Singh, and Anandkumar (2018) propose a method for creating a dataset of symbolic mathematical equations for the tasks of symbolic equation verification and equation completion. Unfortunately, a dataset constructed using the method they propose will suffer from two serious flaws. First, the class of true equations that the procedure can generate will be very limited. Second, because true and false equations are generated in completely different ways, there are likely to be artifactual features that allow easy discrimination.
Moreover, over the class of equations they consider, there is an extremely simple probabilistic procedure that solves the problem of equation verification with extremely high reliability. The usefulness of this problem in general as a testbed for AI systems is therefore doubtful.
Subjects: Artificial Intelligence (cs.AI); Symbolic Computation (cs.SC)
Cite as: arXiv:2105.11479 [cs.AI]
  (or arXiv:2105.11479v4 [cs.AI] for this version)

Submission history

From: Ernest Davis [view email]
[v1] Mon, 24 May 2021 18:05:38 GMT (6kb)
[v2] Wed, 26 May 2021 14:46:53 GMT (8kb)
[v3] Thu, 27 May 2021 01:34:45 GMT (8kb)
[v4] Fri, 28 May 2021 03:05:46 GMT (8kb)

Link back to: arXiv, form interface, contact.