Despite the many ominous connotations trumpeted in works of fiction, the adoption and growth of AI can is simply another phase of the technological advance that has marked the development of human society. Yet, because we associate intelligence with living creatures, particularly our own species, the idea of machines that possess that faculty excites some trepidation. AI agents may turn out to be as unpredictable and perverse as any intelligent human. No such worry is evident in Silicon Valley. Sundar Pichai, Google's chief, speaking at the World Economic Forum in Davos, Switzerland, enthused about the technology: "AI is probably the most important thing humanity has ever worked on. I think of it as something more profound than electricity or fire," he said. Google is a major participant in an AI market that is clipping along at a five-year compound annual growth rate (CAGR) of 17.5%. Globally, the industry is projected to swell to $554.3 billion by 2024. Other players of note are IBM, Intuit, Microsoft, OpenText, Palantir, SAS, and Slack.

Context-aware machine translation models are designed to leverage contextual information, but often fail to do so. As a result, they inaccurately disambiguate pronouns and polysemous words that require context for resolution. In this paper, we ask several questions: What contexts do human translators use to resolve ambiguous words? Are models paying large amounts of attention to the same context? What if we explicitly train them to do so? To answer these questions, we introduce SCAT (Supporting Context for Ambiguous Translations), a new English-French dataset comprising supporting context words for 14K translations that professional translators found useful for pronoun disambiguation. Using SCAT, we perform an in-depth analysis of the context used to disambiguate, examining positional and lexical characteristics of the supporting words. Furthermore, we measure the degree of alignment between the model's attention scores and the supporting context from SCAT, and apply a guided attention strategy to encourage agreement between the two.

Neural Machine Translation (NMT) is an end-to-end learning approach for automated translation [1]. Its strength comes from the fact that it learns the mapping directly from input text to associated output text. It has been proven to be more effective than traditional phrase-based machine translation, which requires much more effort to design the model. On the other hand, NMT models are costly to train, especially on large-scale translation datasets. They are also significantly slower at inference time due to the large number of parameters used.

Back in 2017, the media was frenzying over Facebook's decision to scrap one of its artificial intelligence engines, which was said to have created its own language that could not be understood by humans. AI can sometimes be quirky and suspicious. In the current world technology is not a luxury, but a necessity. AI is one of the most popular of them, which has successfully aided several innovations across all industries. There have been many speculations around this disruptive technology and it has often been looked upon with fear.

Fujitsu Ltd. on Thursday released a multilingual speech translation service that does not require users to operate devices by hand. The service is designed for settings in which multilingual communication is needed amid a rise in the domestic population of non-Japanese speakers, such as medical facilities. It automatically translates speech after identifying the voices and locations of users on the basis of sound picked up by directional microphones connected to tablet devices. Fujitsu said that the voice recognition is highly accurate thanks to technology limiting the effects of background noise. In addition to medical settings, the service is expected to be used at tourist sites.

Ensembles of machine learning models yield improved system performance as well as robust and interpretable uncertainty estimates; however, their inference costs may often be prohibitively high. Ensemble Distribution Distillation is an approach that allows a single model to efficiently capture both the predictive performance and uncertainty estimates of an ensemble. For classification, this is achieved by training a Dirichlet distribution over the ensemble members' output distributions via the maximum likelihood criterion. Although theoretically principled, this criterion exhibits poor convergence when applied to large-scale tasks where the number of classes is very high. In our work, we analyze this effect and show that for the Dirichlet log-likelihood criterion classes with low probability induce larger gradients than high-probability classes. This forces the model to focus on the distribution of the ensemble tail-class probabilities. We propose a new training objective which minimizes the reverse KL-divergence to a Proxy-Dirichlet target derived from the ensemble. This loss resolves the gradient issues of Ensemble Distribution Distillation, as we demonstrate both theoretically and empirically on the ImageNet and WMT17 En-De datasets containing 1000 and 40,000 classes, respectively.

In the past 10 years, hundreds of projects have applied artificial intelligence (AI) to creating social good. The right tool applied to an appropriate problem has the potential to drastically improve millions of lives through better service delivery and better-informed policy design. But what kind of investments do AI solutions need to be successful, and which applications have the most potential for social impact? AI excels at helping humans harness large-scale or complex data to predict, categorize, or optimize at a scale and speed beyond human ability. We believe that more targeted, sustained investments in AI for social impact (sometimes called "AI for good")--rather than multiple, short-term grants across a variety of areas--are important for two reasons.

Questions of fairness, robustness, and transparency are paramount to address before deploying NLP systems. Central to these concerns is the question of reliability: Can NLP systems reliably treat different demographics fairly and function correctly in diverse and noisy environments? To address this, we argue for the need for reliability testing and contextualize it among existing work on improving accountability. We show how adversarial attacks can be reframed for this goal, via a framework for developing reliability tests. We argue that Figure 1: How DOCTOR can integrate with existing reliability testing -- with an emphasis on interdisciplinary system development workflows. Test (left) and system collaboration -- will enable rigorous development (right) take place in parallel, separate and targeted testing, and aid in the enactment teams. Reliability tests can thus be constructed independent and enforcement of industry standards. of the system development team, either by an internal "red team" or by independent auditors.

Generating accurate terminology is a crucial component for the practicality and reliability of neural machine translation (NMT) systems. To address this, lexically constrained NMT explores various methods to ensure pre-specified words and phrases to appear in the translations. In many cases, however, those methods are evaluated on general domain corpora, where the terms are mostly uni- and bi-grams (>98%). In this paper, we instead tackle a more challenging setup consisting of domain-specific corpora with much longer n-gram and highly specialized terms. To encourage span-level representations in generation, we additionally impose a source-sentence conditioned masked span prediction loss in the decoder and observe improvements on both terminology translation as well as BLEU scores. Experimental results on three domain-specific corpora in two language pairs demonstrate that the proposed training scheme can improve the performance of existing lexically constrained methods that can operate both with or without a term dictionary at test time.

Signed languages are the primary means of communication for many deaf and hard of hearing individuals. Since signed languages exhibit all the fundamental linguistic properties of natural language, we believe that tools and theories of Natural Language Processing (NLP) are crucial towards its modeling. However, existing research in Sign Language Processing (SLP) seldom attempt to explore and leverage the linguistic organization of signed languages. This position paper calls on the NLP community to include signed languages as a research area with high social and scientific impact. We first discuss the linguistic properties of signed languages to consider during their modeling. Then, we review the limitations of current SLP models and identify the open challenges to extend NLP to signed languages. Finally, we urge (1) the adoption of an efficient tokenization method; (2) the development of linguistically-informed models; (3) the collection of real-world signed language data; (4) the inclusion of local signed language communities as an active and leading voice in the direction of research.