Starting today, Google will rely more heavily on artificial intelligence when it translates language. The new method, called Google Machine Neural Translation, cuts down errors by 80% compared to its current algorithm, and is nearly indistinguishable from human translation on standardized tests, the company said. It's a radical change from how Google translates now, which is called phrase-based translation, and is standard for the industry. Under this method, an algorithm cuts up a sentence, like one entered by a high-schooler trying to game their homework, and attempts to match words or phrases to a large dictionary. The new method takes that same large dictionary and uses it to train two neural networks.

Last March, a computer built by a team of Google engineers beat one of the world's top players at the ancient game of Go. The match between AlphaGo and Korean grandmaster Lee Sedol was so exhilarating, so upsetting, and so unexpectedly powerful, we turned it into a cover story for the magazine. On a Friday in late April, we were about an hour away from sending this story to the printer when I got an email. According to the email, Lee had won all five matches--and all against top competition--since his loss to AlphaGo. Even as it surpasses human talents, AI can also pull humans to new heights--a theme that ran through our magazine story.

As a native speaker (and reader and writer) of both Mandarin Chinese (both complex and traditional alphabets) and English, I've often cringed at Google Translate's output. But looking at the examples provided by Google on its blog post, I am impressed. The new system uses what the company calls Google Neural Machine Translation (GNMT), which looks at entire sentences as it decodes instead of breaking them up into words and phrases to be considered independently. The latter method often resulted in disjointed results that sometimes didn't make sense. According to the company, this new technique is better, because "it requires fewer engineering design choices than previous Phrase-Based translation systems." It still breaks up sentences into individual characters, but now considers each one in relation to those before and after it.

Mandarin Chinese is a notoriously difficult language to translate to English, and for those who rely on Google Translate to decipher important information, machine-based tools simply aren't good enough. All that is about to change, as Google today announced it has implemented a new learning system in its web and mobile translation apps that will bring significantly better results. As a native speaker (and reader and writer) of both Mandarin Chinese (both complex and traditional alphabets) and English, I've often cringed at Google Translate's output. But looking at the examples provided by Google on its blog post, I am impressed. The new system uses what the company calls Google Neural Machine Translation (GNMT), which looks at entire sentences as it decodes instead of breaking them up into words and phrases to be considered independently. The latter method often resulted in disjointed results that sometimes didn't make sense.

Google is one of the leading providers of artificial intelligence-assisted language translation, and the company now says a new technique for doing so is vastly improving the results. The company's AI team calls it the Google Neural Machine Translation system, or GNMT, and it initially provided a less resource-intensive way to ingest a sentence in one language and produce that same sentence in another language. Instead of digesting each word or phrase as a standalone unit, as prior methods do, GNMT takes in the entire sentence as a whole. "The advantage of this approach is that it requires fewer engineering design choices than previous Phrase-Based translation systems," writes Quoc V. Le and Mike Schuster, researchers on the Google Brain team. When the technique was first employed, it was able to match the accuracy of those existing translation systems.

Last March, a computer built by a team of Google engineers beat one of the world's top players at the ancient game of Go. The match between AlphaGo and Korean grandmaster Lee Sedol was so exhilarating, so upsetting, and so unexpectedly powerful, we turned it into a cover story for the magazine. On a Friday in late April, we were about an hour away from sending this story to the printer when I got an email. According to the email, Lee had won all five matches--and all against top competition--since his loss to AlphaGo. Even as it surpasses human talents, AI can also pull humans to new heights--a theme that ran through our magazine story.

Embedding words in a vector space has gained a lot of attention in recent years. While state-of-the-art methods provide efficient computation of word similarities via a low-dimensional matrix embedding, their motivation is often left unclear. In this paper, we argue that word embedding can be naturally viewed as a ranking problem due to the ranking nature of the evaluation metrics. Then, based on this insight, we propose a novel framework WordRank that efficiently estimates word representations via robust ranking, in which the attention mechanism and robustness to noise are readily achieved via the DCG-like ranking losses. The performance of WordRank is measured in word similarity and word analogy benchmarks, and the results are compared to the state-of-the-art word embedding techniques. Our algorithm is very competitive to the state-of-the- arts on large corpora, while outperforms them by a significant margin when the training set is limited (i.e., sparse and noisy). With 17 million tokens, WordRank performs almost as well as existing methods using 7.2 billion tokens on a popular word similarity benchmark. Our multi-node distributed implementation of WordRank is publicly available for general usage.

Which authors of this paper are endorsers? Disable MathJax (What is MathJax?)

APPLE released the iPhone 7 last week along with iOS 10, a major upgrade for its popular mobile operating system. What better time to download some apps to take the new hardware and software for a spin? In one update, Apple's messaging system, iMessage, got a turbo boost. It now lets people embellish conversations with stickers, interactive drawings and animations. The messaging system even gets its own App Store for downloading third-party stickers and games.

This tutorial will explain some practical tips about how to train a neural machine translation system. It is partly based around examples using the lamtram toolkit. Note that this will not cover the theory behind NMT in detail, nor is it a survey meant to cover all the work on neural MT, but it will show you how to use lamtram, and also demonstrate some things that you have to do in order to make a system that actually works well (focusing on ones that are implemented in my toolkit). This tutorial will assume that you have already installed lamtram (and the cnn backend library that it depends on) on Linux or Mac. Then, use git to pull this tutorial and the corresponding data. The data in the data/ directory is Japanese-English data that I have prepared doing some language-specific preprocessing (tokenization, lowercasing, etc.). Machine translation is a method for translating from a source sequence F with words f_1, ..., f_J to a target sequence E with words e_1, ..., e_I. This usually means that we translate between a sentence in a source language (e.g.