| Title: | Sequence-to-Sequence models and their evaluation for spoken language normalization of Slovenian |
|---|
| Authors: | ID Sepesy Maučec, Mirjam (Author)
ID Verdonik, Darinka (Author)
ID Donaj, Gregor (Author) |
|---|
| Files: |  applsci-14-09515.pdf (437,99 KB)
MD5: 19D67562D60D137292969407CD398506
 https://www.mdpi.com/2076-3417/14/20/9515
|
|---|
| Language: | English |
|---|
| Work type: | Article |
|---|
| Typology: | 1.01 - Original Scientific Article |
|---|
| Organization: | FERI - Faculty of Electrical Engineering and Computer Science
|
|---|
| Abstract: | Sequence-to-sequence models have been applied to many challenging problems, including
those in text and speech technologies. Normalization is one of them. It refers to transforming
non-standard language forms into their standard counterparts. Non-standard language forms
come from different written and spoken sources. This paper deals with one such source, namely
speech from the less-resourced highly inflected Slovenian language. The paper explores speech
corpora recently collected in public and private environments. We analyze the efficiencies of three
sequence-to-sequence models for automatic normalization from literal transcriptions to standard
forms. Experiments were performed using words, subwords, and characters as basic units for
normalization. In the article, we demonstrate that the superiority of the approach is linked to the
choice of the basic modeling unit. Statistical models prefer words, while neural network-based
models prefer characters. The experimental results show that the best results are obtained with
neural architectures based on characters. Long short-term memory and transformer architectures
gave comparable results. We also present a novel analysis tool, which we use for in-depth error
analysis of results obtained by character-based models. This analysis showed that systems with
similar overall results can differ in the performance for different types of errors. Errors obtained with
the transformer architecture are easier to correct in the post-editing process. This is an important
insight, as creating speech corpora is a time-consuming and costly process. The analysis tool also
incorporates two statistical significance tests: approximate randomization and bootstrap resampling.
Both statistical tests confirm the improved results of neural network-based models compared to
statistical ones. |
|---|
| Keywords: | low-resource language, applications, spoken language, normalization, character unit, subword unit, statistical model, long short-term memory, transformer, error analysis |
|---|
| Publication status: | Published |
|---|
| Publication version: | Version of Record |
|---|
| Submitted for review: | 06.09.2024 |
|---|
| Article acceptance date: | 16.10.2024 |
|---|
| Publication date: | 18.10.2024 |
|---|
| Publisher: | MDPI |
|---|
| Year of publishing: | 2024 |
|---|
| Number of pages: | 24 str. |
|---|
| Numbering: | let. 14, št. 20, št. članka 9515 |
|---|
| PID: | 20.500.12556/DKUM-91741  |
|---|
| UDC: | 004.8 |
|---|
| ISSN on article: | 2076-3417 |
|---|
| COBISS.SI-ID: | 213048067 |
|---|
| DOI: | 10.3390/app14209515 |
|---|
| Copyright: | © 2024 by the authors |
|---|
| Publication date in DKUM: | 31.01.2025 |
|---|
| Views: | 125 |
|---|
| Downloads: | 17 |
|---|
| Metadata: |    |
|---|
| Categories: | Misc.
|
|---|
|
:
|
Copy citation |
|---|
| | | | Average score: | (0 votes) |
|---|
| Your score: | Voting is allowed only for logged in users. |
|---|
| Share: |  |
|---|
Hover the mouse pointer over a document title to show the abstract or click
on the title to get all document metadata. |
