A variety of natural language tasks require processing of textual data which contains a mix of natural language and formal languages such as mathematical expressions. In this paper, we take unit conversions as an example and propose a data augmentation technique which leads to models learning both translation and conversion tasks as well as how to adequately switch between them for end-to-end localization.

Wikipedia's vision is a world in which everyone can share in the sum of all knowledge. In its first two decades, this vision has been very unevenly achieved. One of the largest hindrances is the sheer number of languages Wikipedia needs to cover in order to achieve that goal. We argue that we need a new approach to tackle this problem more effectively, a multilingual Wikipedia where content can be shared between language editions. This paper proposes an architecture for a system that fulfills this goal. It separates the goal in two parts: creating and maintaining content in an abstract notation within a project called Abstract Wikipedia, and creating an infrastructure called Wikilambda that can translate this notation to natural language. Both parts are fully owned and maintained by the community, as is the integration of the results in the existing Wikipedia editions. This architecture will make more encyclopedic content available to more people in their own language, and at the same time allow more people to contribute knowledge and reach more people with their contributions, no matter what their respective language backgrounds. Additionally, Wikilambda will unlock a new type of knowledge asset people can share in through the Wikimedia projects, functions, which will vastly expand what people can do with knowledge from Wikimedia, and provide a new venue to collaborate and to engage the creativity of contributors from all around the world. These two projects will considerably expand the capabilities of the Wikimedia platform to enable every single human being to freely share in the sum of all knowledge.

In training deep learning networks, the optimizer and related learning rate are often used without much thought or with minimal tuning, even though it is crucial in ensuring a fast convergence to a good quality minimum of the loss function that can also generalize well on the test dataset. Drawing inspiration from the successful application of cyclical learning rate policy for computer vision related convolutional networks and datasets, we explore how cyclical learning rate can be applied to train transformer-based neural networks for neural machine translation. From our carefully designed experiments, we show that the choice of optimizers and the associated cyclical learning rate policy can have a significant impact on the performance. In addition, we establish guidelines when applying cyclical learning rates to neural machine translation tasks. Thus with our work, we hope to raise awareness of the importance of selecting the right optimizers and the accompanying learning rate policy, at the same time, encourage further research into easy-to-use learning rate policies.

We cast neural machine translation (NMT) as a classification task in an autoregressive setting and analyze the limitations of both classification and autoregression components. Classifiers are known to perform better with balanced class distributions during training. Since the Zipfian nature of languages causes imbalanced classes, we explore the effect of class imbalance on NMT. We analyze the effect of vocabulary sizes on NMT performance and reveal an explanation for 'why' certain vocabulary sizes are better than others.

A specific use case worth exploring in this regard is MT for User Generated Content (UGC). Because of the speed with which UGC (comments, feedback, reviews) is being created and the corresponding costs of its professional translation, many organizations turn to MT. Popular examples of such companies are Skype (in addition to text translation, Microsoft developed the Automatic Speech Recognition (ASR) for audio speech translation in Skype) and Facebook. The social network is aiming to solve the challenge of fine-tuning each system relating to a specific language pair, using neural machine translation (NMT) and benefiting from various contexts for translations. One solution that tackles this issue is the technology developed by Language I/O. It takes into account the client's glossaries and TMs, selects the best MT engine output and then improves on the results using cultural intelligence and/or human linguists who compare machine translations post-facto to ensure that their MT Optimizer engine learns over time.

Non-autoregressive machine translation models significantly speed up decoding by allowing for parallel prediction of the entire target sequence. However, modeling word order is more challenging due to the lack of autoregressive factors in the model. This difficultly is compounded during training with cross entropy loss, which can highly penalize small shifts in word order. In this paper, we propose aligned cross entropy (AXE) as an alternative loss function for training of non-autoregressive models. AXE uses a differentiable dynamic program to assign loss based on the best possible monotonic alignment between target tokens and model predictions. AXE-based training of conditional masked language models (CMLMs) substantially improves performance on major WMT benchmarks, while setting a new state of the art for non-autoregressive models.

The quality of machine translation has increased remarkably over the past years, to the degree that it was found to be indistinguishable from professional human translation in a number of empirical investigations. We reassess Hassan et al.'s 2018 investigation into Chinese to English news translation, showing that the finding of human-machine parity was owed to weaknesses in the evaluation design - which is currently considered best practice in the field. We show that the professional human translations contained significantly fewer errors, and that perceived quality in human evaluation depends on the choice of raters, the availability of linguistic context, and the creation of reference translations. Our results call for revisiting current best practices to assess strong machine translation systems in general and human-machine parity in particular, for which we offer a set of recommendations based on our empirical findings.

Roman Urdu is an informal form of the Urdu language written in Roman script, which is widely used in South Asia for online textual content. It lacks standard spelling and hence poses several normalization challenges during automatic language processing. In this article, we present a feature-based clustering framework for the lexical normalization of Roman Urdu corpora, which includes a phonetic algorithm UrduPhone, a string matching component, a feature-based similarity function, and a clustering algorithm Lex-Var. UrduPhone encodes Roman Urdu strings to their pronunciation-based representations. The string matching component handles character-level variations that occur when writing Urdu using Roman script.

New machine translations (MT) technologies are emerging rapidly and with them, bold claims of achieving human parity such as: (i) the results produced approach "accuracy achieved by average bilingual human translators" (Wu et al., 2017b) or (ii) the "translation quality is at human parity when compared to professional human translators" (Hassan et al., 2018) have seen the light of day (Laubli et al., 2018). Aside from the fact that many of these papers craft their own definition of human parity, these sensational claims are often not supported by a complete analysis of all aspects involved in translation. Establishing the discrepancies between the strengths of statistical approaches to MT and the way humans translate has been the starting point of our research. By looking at MT output and linguistic theory, we were able to identify some remaining issues. The problems range from simple number and gender agreement errors to more complex phenomena such as the correct translation of aspectual values and tenses. Our experiments confirm, along with other studies (Bentivogli et al., 2016), that neural MT has surpassed statistical MT in many aspects. However, some problems remain and others have emerged. We cover a series of problems related to the integration of specific linguistic features into statistical and neural MT, aiming to analyse and provide a solution to some of them. Our work focuses on addressing three main research questions that revolve around the complex relationship between linguistics and MT in general. We identify linguistic information that is lacking in order for automatic translation systems to produce more accurate translations and integrate additional features into the existing pipelines. We identify overgeneralization or 'algorithmic bias' as a potential drawback of neural MT and link it to many of the remaining linguistic issues.

Artificial Intelligence (AI) is the theory and development of computer systems that can perform tasks that normally require human intelligence. These tasks include visual perception, speech recognition, decision making, and language translation. Systems capable of performing such tasks are steadily transitioning from research laboratories into industry usage. AI technology is unique in that it is flexible in application. It can be used to improve processes, enhance interactions, and solve problems that, until recently, could only be performed by humans.