Pith. sign in

REVIEW 2 cited by

Unsupervised Expressive Rules Provide Explainability and Assist Human Experts Grasping New Domains

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2010.09459 v1 pith:3MS6JLNB submitted 2020-10-19 cs.CL cs.HCcs.LG

Unsupervised Expressive Rules Provide Explainability and Assist Human Experts Grasping New Domains

classification cs.CL cs.HCcs.LG
keywords rulescategoriesassistcorpusdatadomainexpertsexplainability
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Approaching new data can be quite deterrent; you do not know how your categories of interest are realized in it, commonly, there is no labeled data at hand, and the performance of domain adaptation methods is unsatisfactory. Aiming to assist domain experts in their first steps into a new task over a new corpus, we present an unsupervised approach to reveal complex rules which cluster the unexplored corpus by its prominent categories (or facets). These rules are human-readable, thus providing an important ingredient which has become in short supply lately - explainability. Each rule provides an explanation for the commonality of all the texts it clusters together. We present an extensive evaluation of the usefulness of these rules in identifying target categories, as well as a user study which assesses their interpretability.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. An LLM-Based System for Argument Mining

    cs.CL 2026-05 unverdicted novelty 6.0

    An LLM pipeline turns natural language arguments into structured graphs of premises, conclusions, and support/attack/undercut relations.

  2. An LLM-Based System for Argument Mining

    cs.CL 2026-05 unverdicted novelty 4.0

    An LLM pipeline converts natural-language arguments into abstract graphs of premises, conclusions, and support/attack/undercut relations, with manual and benchmark evaluations showing adequate recovery of structure.