In this paper we present our open-source neural machine translation (NMT) toolkit called "Yet Another Neural Machine Translation Toolkit" abbreviated as YANMTT which is built on top of the Transformers library. Despite the growing importance of sequence to sequence pre-training there surprisingly few, if not none, well established toolkits that allow users to easily do pre-training. Toolkits such as Fairseq which do allow pre-training, have very large codebases and thus they are not beginner friendly. With regards to transfer learning via fine-tuning most toolkits do not explicitly allow the user to have control over what parts of the pre-trained models can be transferred. YANMTT aims to address these issues via the minimum amount of code to pre-train large scale NMT models, selectively transfer pre-trained parameters and fine-tune them, perform translation as well as extract representations and attentions for visualization and analyses. Apart from these core features our toolkit also provides other advanced functionalities such as but not limited to document/multi-source NMT, simultaneous NMT and model compression via distillation which we believe are relevant to the purpose behind our toolkit.

Human evaluation has always been expensive while researchers struggle to trust the automatic metrics. To address this, we propose to customise traditional metrics by taking advantages of the pre-trained language models (PLMs) and the limited available human labelled scores. We first re-introduce the hLEPOR metric factors, followed by the Python portable version we developed which achieved the automatic tuning of the weighting parameters in hLEPOR metric. Then we present the customised hLEPOR (cushLEPOR) which uses LABSE distilled knowledge model to improve the metric agreement with human judgements by automatically optimised factor weights regarding the exact MT language pairs that cushLEPOR is deployed to. We also optimise cushLEPOR towards human evaluation data based on MQM and pSQM framework on English-German and Chinese-English language pairs. The experimental investigations show cushLEPOR boosts hLEPOR performances towards better agreements to PLMs like LABSE with much lower cost, and better agreements to human evaluations including MQM and pSQM scores, and yields much better performances than BLEU (data available at \url{https://github.com/poethan/cushLEPOR}).

Machine learning based compilation has been explored for over a decade [1]. Early work focused on learning profitability heuristics while more recently, deep learning models have been used to build code-to-code models, for translating or decompiling code. However, to the best of our knowledge, there has been no prior work on using machine learning to entirely automate compilation i.e given a high level source code program generate the equivalent assembler code. In this paper, we investigate whether it is possible to learn an end-to-end machine compiler using neural machine translation. In particular, we focus on the translation of small C functions to x86 assembler We use an existing function-level C corpus, Anghabench [2], to build a parallel C-x86 assembler corpus.

NMT is a popular and widely used translation service that incorporates an end-to-end approach for automatic translation which overcomes the weaknesses of RBMT and SMT methods. NMT uses the most recent deep learning methods to produce better translation output than other traditional Machine Translation solutions. It is the most recent type of machine translation that employs a neural network that is closely related to the neurons of the human brain, allowing it to categorize data into various groups and layers. NMT is a language translation approach that tries to incorporate the context of the sentences or paragraphs rather than individual words. The NMT system is made up of current multilingual databases and automated learning mechanisms that contribute to continuous improvement.

Advancements in natural language processing (NLP) enable computers to understand what humans say and help people communicate through tools like machine translation, voice-controlled assistants and chatbots.

Etsy is seeking a Senior Machine Learning Engineer to join our International Search engineering organization in the International Customer Experience initiative. We believe that code is craft and the work we do is part of a larger creative culture connecting 81M buyers and over 4M sellers to millions of items in the diverse marketplace. We believe that small, empowered, self-motivated teams can do big, impactful things. Do you have a passion for natural language processing, machine learning and working in a multi-functional engineering team? If so, this is the role for you!

Amazon's suite of ML services is constantly expanding. From having capabilities of building custom ML pipelines in SageMaker to a versatile set of AutoML services, options to deploy and tackle ML problems are limitless. Neural Machine Translation is a theoretically intense field and requires deep knowledge of LSTMs and Deep Learning frameworks such as TensorFlow and PyTorch. For this article we will explore AWS Translate, a Neural Machine Translation tool that supports 71 languages and lets you build applications with a simple API call. This article is a continuation of the Auto-ML on AWS series, check out the Rekognition and Comprehend articles for the first two parts.

Aspect Sentiment Triplet Extraction (ASTE) is the task of extracting triplets of aspect terms, their associated sentiments, and the opinion terms that provide evidence for the expressed sentiments. Previous approaches to ASTE usually simultaneously extract all three components or first identify the aspect and opinion terms, then pair them up to predict their sentiment polarities. In this work, we present a novel paradigm, ASTE-RL, by regarding the aspect and opinion terms as arguments of the expressed sentiment in a hierarchical reinforcement learning (RL) framework. We first focus on sentiments expressed in a sentence, then identify the target aspect and opinion terms for that sentiment. This takes into account the mutual interactions among the triplet's components while improving exploration and sample efficiency. Furthermore, this hierarchical RLsetup enables us to deal with multiple and overlapping triplets. In our experiments, we evaluate our model on existing datasets from laptop and restaurant domains and show that it achieves state-of-the-art performance. The implementation of this work is publicly available at https://github.com/declare-lab/ASTE-RL.

Moving towards human-like linguistic performance is often argued to require compositional generalisation. Whether neural networks exhibit this ability is typically studied using artificial languages, for which the compositionality of input fragments can be guaranteed and their meanings algebraically composed. However, compositionality in natural language is vastly more complex than this rigid, arithmetics-like version of compositionality, and as such artificial compositionality tests do not allow us to draw conclusions about how neural models deal with compositionality in more realistic scenarios. In this work, we re-instantiate three compositionality tests from the literature and reformulate them for neural machine translation (NMT). The results highlight two main issues: the inconsistent behaviour of NMT models and their inability to (correctly) modulate between local and global processing. Aside from an empirical study, our work is a call to action: we should rethink the evaluation of compositionality in neural networks of natural language, where composing meaning is not as straightforward as doing the math.

In the field of machine learning, as in the most unrelated fields as well, we need some sort of evaluation. You can think of a student taking an exam, a car in a crash test, a web server on load test, and performance evaluation of a model in AI. Evaluation methods differ among these fields and evolution criteria designed marginally. This procedure is needed mainly to assess the quality of outputs of a model, and also to compare them among different models or with different setups, etc. Natural Language Generation (NLG), a field in Natural Language Processing (NLP), is an applied subfield of artificial intelligence, where the goal is to produce a textual output. It has a vast amount of subtasks like machine translation (MT), question answering (QA), summarization, question generation (QG), etc. Here, the discussion is around the performance of the models whose outputs are text.