Neural language models are a critical component of state-of-the-art systems for machine translation, summarization, audio transcription, and other tasks. These language models are almost universally autoregressive in nature, generating sentences one token at a time from left to right. This paper studies the influence of token generation order on model quality via a novel two-pass language model that produces partially-filled sentence "templates" and then fills in missing tokens. We compare various strategies for structuring these two passes and observe a surprisingly large variation in model quality. We find the most effective strategy generates function words in the first pass followed by content words in the second. We believe these experimental results justify a more extensive investigation of generation order for neural language models.

Although the recent flurry of blogs (I count myself in this), emails and media coverage on the topic could leave you thinking otherwise, data security is much more than just GDPR. It's about placing customer and data privacy at the centre of everything you do. It may sound simple, but in a world that also includes legislation such as the Electronic Communications Privacy Act (ECPA) and the China Data Protection Regulation (CDPR), and more specifically in financial services MiFID II and PSD2, complying with different regulations while keeping data privacy front of mind is a complex environment for any digital business to navigate. Especially for brands that engage with content hungry customers across multiple countries and languages. While my primary focus is our financial services industry, since May the number of retailers and IT organizations I speak to has grown.

There have been several attempts to define a plausible motivation for a chit-chat dialogue agent that can lead to engaging conversations. In this work, we explore a new direction where the agent specifically focuses on discovering information about its interlocutor. We formalize this approach by defining a quantitative metric. We propose an algorithm for the agent to maximize it. We validate the idea with human evaluation where our system outperforms various baselines. We demonstrate that the metric indeed correlates with the human judgments of engagingness.

Evolution shows that social mimicry is a major component of our survival mechanism, pushing us to belong to groups. We're all subject to the chameleon effect, this tendency to unconsciously mirror what "others" do, and one of its most apparent manifestations is language and accents. Members of the same social group tend to mimic the speech patterns of others, leading to the rise of different regional accents within the same language. In the Southern United States, for example, English has developed in contact with Spanish, leading speakers on the two sides of the borders to pick up dialectal elements Like in Puerto Rico, the mix was so deep that "Spanglish" appeared. On both sides of the pond, in the United States and Britain, people speak English, yet with very distinctive accents that, in some cases, could be mutually unintelligible.

Natural Language Processing; it's Artificial Intelligence that learns words and patterns of words so that it can respond to human searches and questions. Siri and Alexa are examples of this technology. And this technology is continually improving. As more and more conversations are held with these machines, they continue to learn and respond more accurately. Machines are also in use for translations.

AI hype is at fever pitch. It's slated to fundamentally change everything about our world, from our economies to the way we get around cities. But how much of the hype is credible, and how much will AI change the nature of business in the near future? Will AI completely take over the realm of human expertise? Algorithms are changing the world.

This is a guest blog post by Andrea Tabacchi, the Solution Architects team lead at Memsource. Memsource is always looking out for exciting new integrations that enhance its cutting-edge translation solutions. With machine translation (MT) continuing to be a hot topic in the localization industry, Memsource is focusing on integrating with innovative MT engines that meet customers' growing MT needs. In particular, Memsource strives to offer neural machine translation (NMT) engines, such as Amazon Translate. NMT is proving to be a highly influential technology. The quality of NMT output continues to improve, making it a powerful productivity tool and therefore more in demand.

As a promising area in artificial intelligence, a new learning paradigm, called Small Sample Learning (SSL), has been attracting prominent research attention in the recent years. In this paper, we aim to present a survey to comprehensively introduce the current techniques proposed on this topic. Specifically, current SSL techniques can be mainly divided into two categories. The first category of SSL approaches can be called "concept learning", which emphasizes learning new concepts from only few related observations. The purpose is mainly to simulate human learning behaviors like recognition, generation, imagination, synthesis and analysis. The second category is called "experience learning", which usually co-exists with the large sample learning manner of conventional machine learning. This category mainly focuses on learning with insufficient samples, and can also be called small data learning in some literatures. More extensive surveys on both categories of SSL techniques are introduced and some neuroscience evidences are provided to clarify the rationality of the entire SSL regime, and the relationship with human learning process. Some discussions on the main challenges and possible future research directions along this line are also presented.

A number of recent machine learning papers work with an automated style transfer for texts and, counter to intuition, demonstrate that there is no consensus formulation of this NLP task. Different researchers propose different algorithms, datasets and target metrics to address it. This short opinion paper aims to discuss possible formalization of this NLP task in anticipation of a further growing interest to it.

Google services such as its image search and translation tools use sophisticated machine learning which allow computers to see, listen and speak in much the same way as human do. Machine learning is the term for the current cutting-edge applications in artificial intelligence. Basically, the idea is that by teaching machines to "learn" by processing huge amounts of data they will become increasingly better at carrying out tasks that traditionally can only be completed by human brains. These techniques include "computer vision" – training computers to recognize images in a similar way we do. For example, an object with four legs and a tail has a high probability of being an animal.