Read this Medium article for full discussion.

Semantic Models for Ethiopic

The semantic models resources are added to Lanfrica

amharic-corpus

Semantic models

Amharic Segmenter, toknizer, and translitrator

Announcements

:tada: :tada: :tada: The Amharic RoBERTa model is uploaded in Huggingface Amharic RoBERTa Model :tada: :tada: :tada:

:tada: :tada: The Amharic FLAIR embedding model is integrated into the FLAIR library as am-forward :tada: :tada: The model will be accessible on the next FLAIR release. Details

:tada: :tada: The Amharic Segmenter, Toknizer, and Translitrator is released and can be installed as pip install amseg :tada: :tada:

:tada: :tada: The Flair based Amharic NER classifier model is now released am-flair-ner :tada: :tada:

:tada: :tada: The Flair based Amharic Sentiment classifier model is now released am-flair-sent :tada: :tada:

:tada: :tada: The Flair based Amharic POS tagger is now released am-flair-pos :tada: :tada:

Different semantic models and applications for Amharic

• Introduction

• Static word2Vec Embeddings

• Network Embeddings

• Contextual Embeddings

Tasks, Datasets and Preprocessing tools

• Here, we have described the different NLP tasks for which we built models using the benchmark datasets Tasks
• NER
• Sentiment
• POS tagging
• Question classification

• Machine Translation

• The different datsets and resources are available under: Datasets
• Named Entity recognition dataset
• POS dataset
• Sentiment Dataset
• Question Classification Dataset
• Machine Translation Dataset

• For Etiopic word segmentation, tokenization, and translitration check this project: Segmentation

Publications

To cite the different Amharic NLP models and resources, use the following paper

@Article{fi13110275,
AUTHOR = {Yimam, Seid Muhie and Ayele, Abinew Ali and Venkatesh, Gopalakrishnan and Gashaw, Ibrahim and Biemann, Chris},
TITLE = {Introducing Various Semantic Models for Amharic: Experimentation and Evaluation with Multiple Tasks and Datasets},
JOURNAL = {Future Internet},
VOLUME = {13},
YEAR = {2021},
NUMBER = {11},
ARTICLE-NUMBER = {275},
URL = {https://www.mdpi.com/1999-5903/13/11/275},
ISSN = {1999-5903},
DOI = {10.3390/fi13110275}
}

To cite the impacts of homophone normalization, use the the following paper

@inproceedings{belay2021impacts,
  title={Impacts of Homophone Normalization on Semantic Models for Amharic},
  author={Belay, Tadesse Destaw and Ayele, Abinew Ali and Gelaye, Getie and Yimam, Seid Muhie and Biemann, Chris},
  booktitle={2021 International Conference on Information and Communication Technology for Development for Africa (ICT4DA)},
  pages={101-106},
  year={2021},
  ISSN = {978-1-6654-3666-3},
  DOI = (10.1109/ICT4DA53266.2021.9672229},
  publisher={IEEE}
}

To cite the Question Answering Classification for Amharic, use the the following paper

@inproceedings{belay2022question,
  title={Question Answering Classification for Amharic Social Media Community Based Questions},
  author={Belay, Tadesse Destaw and Yimam, Seid Muhie and Gelaye, Getie and Ayele, Abinew Ali  and Biemann, Chris},
  booktitle={2022  1st Annual Meeting of the ELRA/ISCA Special Interest Group on Under-Resourced Languages (SIGUL)},
  pages={137-145},
  year={2022},
  publisher={aclanthology}
}

To cite biderectional Amharic to English Machine Translation, use the the following paper

@INPROCEEDINGS{9971385,  
author={Belay, Tadesse Destaw and Tonja, Atnafu Lambebo and Kolesnikova, Olga and Yimam, Seid Muhie and Ayele, Abinew Ali and Haile, Silesh Bogale and Sidorov, Grigori and Gelbukh, Alexander},  
booktitle={2022 International Conference on Information and Communication Technology for Development for Africa (ICT4DA)},   
title={The Effect of Normalization for Bi-directional Amharic-English Neural Machine Translation},   
year={2022},  
pages={84-89},  
doi={10.1109/ICT4DA56482.2022.9971385}
}