SignON (https://signon-project.eu/) is a Horizon 2020 project, running from 2021 until the end of 2023, which addresses the lack of technology and services for the automatic translation between sign languages (SLs) and spoken languages, through an inclusive, human-centric solution, hence contributing to the repertoire of communication media for deaf, hard of hearing (DHH) and hearing individuals. In this paper, we present an update of the status of the project, describing the approaches developed to address the challenges and peculiarities of SL machine translation (SLMT).

The SignON project (www.signon-project.eu) focuses on the research and development of a Sign Language (SL) translation mobile application and an open communications framework. SignON rectifies the lack of technology and services for the automatic translation between signed and spoken languages, through an inclusive, humancentric solution which facilitates communication between deaf, hard of hearing (DHH) and hearing individuals. We present an overview of the current status of the project, describing the milestones reached to date and the approaches that are being developed to address the challenges and peculiarities of Sign Language Machine Translation (SLMT).

One of the major challenges in sign language translation from a sign language to a spoken language is the lack of parallel corpora. Recent works have achieved promising results on the RWTH-PHOENIX-Weather 2014T dataset, which consists of over eight thousand parallel sentences between German sign language and German. However, from the perspective of neural machine translation, this is still a tiny dataset. To improve the performance of models trained on small datasets, transfer learning can be used. While this has been previously applied in sign language translation for feature extraction, to the best of our knowledge, pretrained language models have not yet been investigated. We use pretrained BERT-base and mBART-50 models to initialize our sign language video to spoken language text translation model. To mitigate overfitting, we apply the frozen pretrained transformer technique: we freeze the majority of parameters during training. Using a pretrained BERT model, we outperform a baseline trained from scratch by 1 to 2 BLEU-4. Our results show that pretrained language models can be used to improve sign language translation performance and that the self-attention patterns in BERT transfer in zero-shot to the encoder and decoder of sign language translation models.

Sign languages are complex languages. Research into them is ongoing, supported by large video corpora of which only small parts are annotated. Sign language recognition can be used to speed up the annotation process of these corpora, in order to aid research into sign languages and sign language recognition. Previous research has approached sign language recognition in various ways, using feature extraction techniques or end-to-end deep learning. In this work, we apply a combination of feature extraction using OpenPose for human keypoint estimation and end-to-end feature learning with Convolutional Neural Networks. The proven multi-head attention mechanism used in transformers is applied to recognize isolated signs in the Flemish Sign Language corpus. Our proposed method significantly outperforms the previous state of the art of sign language recognition on the Flemish Sign Language corpus: we obtain an accuracy of 74.7% on a vocabulary of 100 classes. Our results will be implemented as a suggestion system for sign language corpus annotation.

Character-level features are currently used in different neural network-based natural language processing algorithms. However, little is known about the character-level patterns those models learn. Moreover, models are often compared only quantitatively while a qualitative analysis is missing. In this paper, we investigate which character-level patterns neural networks learn and if those patterns coincide with manually-defined word segmentations and annotations. To that end, we extend the contextual decomposition technique (Murdoch et al. 2018) to convolutional neural networks which allows us to compare convolutional neural networks and bidirectional long short-term memory networks. We evaluate and compare these models for the task of morphological tagging on three morphologically different languages and show that these models implicitly discover understandable linguistic rules.

In this paper, we introduce the novel concept of densely connected layers into recurrent neural networks. We evaluate our proposed architecture on the Penn Treebank language modeling task. We show that we can obtain similar perplexity scores with six times fewer parameters compared to a standard stacked 2-layer LSTM model trained with dropout (Zaremba et al., 2014). In contrast with the current usage of skip connections, we show that densely connecting only a few stacked layers with skip connections already yields significant perplexity reductions.