Despite the recent advancements of attention-based deep learning architectures across a majority of Natural Language Processing tasks, their application remains limited in a low-resource setting because of a lack of pre-trained models for such languages. In this study, we make the first attempt to investigate the challenges of adapting these techniques to an extremely low-resource language – Sumerian cuneiform – one of the world’s oldest written languages attested from at least the beginning of the 3rd millennium BC. Specifically, we introduce the first cross-lingual information extraction pipeline for Sumerian, which includes part-of-speech tagging, named entity recognition, and machine translation. We introduce InterpretLR, an interpretability toolkit for low-resource NLP and use it alongside human evaluations to gauge the trained models. Notably, all our techniques and most components of our pipeline can be generalised to any low-resource language. We publicly release all our implementations including a novel data set with domain-specific pre-processing to promote further research in this domain.

The Sumerian cuneiform script was invented more than 5,000 years ago and represents one of the oldest in history. We present the first attempt to translate Sumerian texts into English automatically. We publicly release high-quality corpora for standardized training and evaluation and report results on experiments with supervised, phrase-based, and transfer learning techniques for machine translation. Quantitative and qualitative evaluations indicate the usefulness of the translations. Our proposed methodology provides a broader audience of researchers with novel access to the data, accelerates the costly and time-consuming manual translation process, and helps them better explore the relationships between Sumerian cuneiform and Mesopotamian culture.

This paper describes our contribution to the Third Shared Task on Translation Inference across Dictionaries (TIAD-2020). We describe an approach on translation inference based on symbolic methods, the propagation of concepts over a graph of interconnected dictionaries: Given a mapping from source language words to lexical concepts (e.g., synsets) as a seed, we use bilingual dictionaries to extrapolate a mapping of pivot and target language words to these lexical concepts. Translation inference is then performed by looking up the lexical concept(s) of a source language word and returning the target language word(s) for which these lexical concepts have the respective highest score. We present two instantiations of this system: One using WordNet synsets as concepts, and one using lexical entries (translations) as concepts. With a threshold of 0, the latter configuration is the second among participant systems in terms of F1 score. We also describe additional evaluation experiments on Apertium data, a comparison with an earlier approach based on embedding projection, and an approach for constrained projection that outperforms the TIAD-2020 vanilla system by a large margin.

We introduce an attention-based Bi-LSTM for Chinese implicit discourse relations and demonstrate that modeling argument pairs as a joint sequence can outperform word order-agnostic approaches. Our model benefits from a partial sampling scheme and is conceptually simple, yet achieves state-of-the-art performance on the Chinese Discourse Treebank. We also visualize its attention activity to illustrate the model’s ability to selectively focus on the relevant parts of an input sequence.

We present a resource-lean neural recognizer for modeling coherence in commonsense stories. Our lightweight system is inspired by successful attempts to modeling discourse relations and stands out due to its simplicity and easy optimization compared to prior approaches to narrative script learning. We evaluate our approach in the Story Cloze Test demonstrating an absolute improvement in accuracy of 4.7% over state-of-the-art implementations.