Pith. sign in

REVIEW 3 cited by

Lenient Evaluation of Japanese Speech Recognition: Modeling Naturally Occurring Spelling Inconsistency

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 2306.04530 v1 pith:F5JXMRJX submitted 2023-06-07 cs.CL

Lenient Evaluation of Japanese Speech Recognition: Modeling Naturally Occurring Spelling Inconsistency

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

Word error rate (WER) and character error rate (CER) are standard metrics in Speech Recognition (ASR), but one problem has always been alternative spellings: If one's system transcribes adviser whereas the ground truth has advisor, this will count as an error even though the two spellings really represent the same word. Japanese is notorious for ``lacking orthography'': most words can be spelled in multiple ways, presenting a problem for accurate ASR evaluation. In this paper we propose a new lenient evaluation metric as a more defensible CER measure for Japanese ASR. We create a lattice of plausible respellings of the reference transcription, using a combination of lexical resources, a Japanese text-processing system, and a neural machine translation model for reconstructing kanji from hiragana or katakana. In a manual evaluation, raters rated 95.4% of the proposed spelling variants as plausible. ASR results show that our method, which does not penalize the system for choosing a valid alternate spelling of a word, affords a 2.4%-3.1% absolute reduction in CER depending on the task.

discussion (0)

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

Forward citations

Cited by 3 Pith papers

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

  1. Voice of India: A Large-Scale Benchmark for Real-World Speech Recognition in India

    cs.CL 2026-04 unverdicted novelty 7.0

    Voice of India is a new 536-hour benchmark of real telephonic conversations in 15 Indian languages with variant-aware transcripts for more realistic ASR evaluation.

  2. Voice of India: A Large-Scale Benchmark for Real-World Speech Recognition in India

    cs.CL 2026-04 conditional novelty 6.0

    A 536-hour unscripted telephonic ASR benchmark covering 15 Indian languages and 139 regional clusters, with multi-reference transcripts for spelling variation and district-level performance analysis.

  3. Voice of India: A Large-Scale Benchmark for Real-World Speech Recognition in India

    cs.CL 2026-04 unverdicted novelty 5.0

    A 536-hour, 15-language, 139-cluster telephonic ASR benchmark for Indian languages with spelling-variation-aware transcripts and geographic performance analysis.