Lesan -- Machine Translation for Low Resource Languages

Millions of people around the world can not access content on the Web because most of the content is not readily available in their language. Machine translation (MT) systems have the potential to change this for many languages. Current MT systems provide very accurate results for high resource language pairs, e.g., German and English. However, for many low resource languages, MT is still under active research. The key challenge is lack of datasets to build these systems. We present Lesan, an MT system for low resource languages. Our pipeline solves the key bottleneck to low resource MT by leveraging online and offline sources, a custom OCR system for Ethiopic and an automatic alignment module. The final step in the pipeline is a sequence to sequence model that takes parallel corpus as input and gives us a translation model. Lesan's translation model is based on the Transformer architecture. After constructing a base model, back translation, is used to leverage monolingual corpora. Currently Lesan supports translation to and from Tigrinya, Amharic and English. We perform extensive human evaluation and show that Lesan outperforms state-of-the-art systems such as Google Translate and Microsoft Translator across all six pairs. Lesan is freely available and has served more than 10 million translations so far. At the moment, there are only 217 Tigrinya and 15,009 Amharic Wikipedia articles. We believe that Lesan will contribute towards democratizing access to the Web through MT for millions of people.

Evaluating Amharic Machine Translation

Machine translation (MT) systems are now able to provide very accurate results for high resource language pairs. However, for many low resource languages, MT is still under active research. In this paper, we develop and share a dataset to automatically evaluate the quality of MT systems for Amharic. We compare two commercially available MT systems that support translation of Amharic to and from English to assess the current state of MT for Amharic. The BLEU score results show that the results for Amharic translation are promising but still low. We hope that this dataset will be useful to the research community both in academia and industry as a benchmark to evaluate Amharic MT systems.

Cross-lingual Short-text Matching with Deep Learning

The problem of short text matching is formulated as follows: given a pair of sentences or questions, a matching model determines whether the input pair mean the same or not. Models that can automatically identify questions with the same meaning have a wide range of applications in question answering sites and modern chatbots. In this article, we describe the approach by team hahu to solve this problem in the context of the "CIKM AnalytiCup 2018 - Cross-lingual Short-text Matching of Question Pairs" that is sponsored by Alibaba. Our solution is an end-to-end system based on current advances in deep learning which avoids heavy feature-engineering and achieves improved performance over traditional machine-learning approaches. The log-loss scores for the first and second rounds of the contest are 0.35 and 0.39 respectively. The team was ranked 7th from 1027 teams in the overall ranking scheme by the organizers that consisted of the two contest scores as well as: innovation and system integrity, understanding data as well as practicality of the solution for business.