Papers by Jack Rueter

4 papers
Finnish Dialect Identification: The Effect of Audio and Text (2021.emnlp-main)

Copied to clipboard

Challenge: Finnish is a language with multiple dialects that differ in accent, morphological forms and lexical choice.
Approach: They propose an approach to automatically detect the dialect of a speaker based on a transcript and transcript with audio recording in a dataset consisting of 23 different dialects.
Outcome: The proposed method achieves 57% accuracy, compared to 85% accuracy for text and audio.
The Low Saxon LSDC Dataset at Universal Dependencies (2024.lrec-main)

Copied to clipboard

Challenge: Low Saxon is a low-resource language that lacks a common standard . dialectal variation in morphological categories can cause problems .
Approach: They extend the Low Saxon Universal Dependencies dataset to include 8 of the 9 major dialects.
Outcome: The proposed dataset covers the last 200 years and 8 of the 9 major dialects.
Combining Concepts and Their Translations from Structured Dictionaries of Uralic Minority Languages (L18-1)

Copied to clipboard

Challenge: a new method to expand the knowledge in existing dictionaries is proposed . small Uralic languages are facing a problem of limited language resources .
Approach: They propose to combine conceptually divided translations from multilingual dictionaries for small Uralic languages into a single lexical entry.
Outcome: The proposed method can be used to expand existing dictionaries and provide translations when adding new entries.
Ve’rdd. Narrowing the Gap between Paper Dictionaries, Low-Resource NLP and Community Involvement (2020.coling-demos)

Copied to clipboard

Challenge: Existing tools for Skolt Sami are limited due to its pluricentric nature and limited resources.
Approach: They propose to integrate community activities into a finite-state language description of a seriously endangered minority language, Skolt Sami.
Outcome: The proposed system integrates with existing tools and infrastructures for Uralic language masking the technical complexities behind a user-friendly UI.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations