AI has evolved dramatically in the last two decades. Technologies like image recognition and machine translation are now a part of everyday life for millions. AI has transformed industries all over the world, and created entirely new ones. And in the process, it promises an increase in quality of life and work never before imagined. But there's still much more we can do--after all, AI is still a nascent field of many opportunities and challenges.

Summer Lab Diversity Coordinator Jihane Lamouri believes that having a variety of perspectives, as the program encourages, is crucial to the development of AI: If a society is biased, so, too, are its machines. At the Lab's closing event, Lamouri referenced the alleged sexism that machine translation services like Google Translate or Microsoft's Bing Translator exhibit. When translating phrases from gender-neutral languages like Finnish or Turkish, machine translators may assign gender pronouns illustrative of a gender bias to the English translation. Users have complained that in the hands of a machine translator the phrase "they are engineer" becomes "he is an engineer," whereas the phrase "they are a nurse" becomes "she is a nurse." Lamouri hopes that having more women in the industry will lead to the identification of gender bias in AI.

The ability to generate natural language sequences from source code snippets can be used for code summarization, documentation, and retrieval. Sequence-to-sequence (seq2seq) models, adopted from neural machine translation (NMT), have achieved state-of-the-art performance on these tasks by treating source code as a sequence of tokens. We present ${\rm {\scriptsize CODE2SEQ}}$: an alternative approach that leverages the syntactic structure of programming languages to better encode source code. Our model represents a code snippet as the set of paths in its abstract syntax tree (AST) and uses attention to select the relevant paths during decoding, much like contemporary NMT models. We demonstrate the effectiveness of our approach for two tasks, two programming languages, and four datasets of up to 16M examples. Our model significantly outperforms previous models that were specifically designed for programming languages, as well as general state-of-the-art NMT models.

I sometimes also carry an Iridium satellite phone, and I always travel with a power strip in case there aren't enough outlets where I am. I try to keep my backpack light in Gaza. Other items -- armor-plated flak jacket, Kevlar helmet, gas mask and spare filters, and a trauma kit -- add lots of weight to my load. Probably the most vital tech tool I carry is the lightest: a paper clip to switch SIM cards. I've been here less than a year and finally got a local number from the Palestinian provider Jawwal, which has good coverage across Gaza and can be quickly replenished at countless retail shops.

You probably don't want to make grammar errors in your emails (or blog posts), but every now and then, they do slip in. Your standard spell-checking tool won't catch them unless you use an extension like Grammarly. Well, Grammarly is getting some competition today in the form of a new machine learning-based grammar checker from Google that's soon going live in Google Docs. These new grammar suggestions in Docs, which are now available through Google's Early Adopter Program, are powered by what is essentially a machine translation algorithm that can recognize errors and suggest corrections as you type. Google says it can catch anything from wrongly used articles ("an" instead of "a") to more complicated issues like incorrectly used subordinate clauses.

A new wave of decision-support systems are being built today using AI services that draw insights from data (like text and video) and incorporate them in human-in-the-loop assistance. However, just as we expect humans to be ethical, the same expectation needs to be met by automated systems that increasingly get delegated to act on their behalf. A very important aspect of an ethical behavior is to avoid (intended, perceived, or accidental) bias. Bias occurs when the data distribution is not representative enough of the natural phenomenon one wants to model and reason about. The possibly biased behavior of a service is hard to detect and handle if the AI service is merely being used and not developed from scratch, since the training data set is not available. In this situation, we envisage a 3rd party rating agency that is independent of the API producer or consumer and has its own set of biased and unbiased data, with customizable distributions. We propose a 2-step rating approach that generates bias ratings signifying whether the AI service is unbiased compensating, data-sensitive biased, or biased. The approach also works on composite services. We implement it in the context of text translation and report interesting results.

For different language pairs, word-level neural machine translation (NMT) models with a fixed-size vocabulary suffer from the same problem of representing out-of-vocabulary (OOV) words. The common practice usually replaces all these rare or unknown words with a 〈UNK〉 token, which limits the translation performance to some extent. Most of recent work handled such a problem by splitting words into characters or other specially extracted subword units to enable open-vocabulary translation. Byte pair encoding (BPE) is one of the successful attempts that has been shown extremely competitive by providing effective subword segmentation for NMT systems. In this paper, we extend the BPE style segmentation to a general unsupervised framework with three statistical measures: frequency (FRQ), accessor variety (AV) and description length gain (DLG). We test our approach on two translation tasks: German to English and Chinese to English. The experimental results show that AV and DLG enhanced systems outperform the FRQ baseline in the frequency weighted schemes at different significant levels.

Google Docs is, at long last, getting a grammar-checking feature, which'll be able to identify mixed up words (like "affect" and "effect"), incorrect tenses, improper uses of commas and clauses, and more. To do all of that, Google says it'll be relying on machine translation -- the same technology it uses to translate between multiple languages. Except, instead of translating a sentence from, say, French to German, it sounds as though it'll be translating your imperfect writing into a grammatically correct passage. Details on what the grammar checking feature is capable of and exactly how its AI will work are limited right now. All we really know is that Google is already quite capable when it comes to machine translation -- two years ago, the company said its tech was approaching human levels of accuracy. So it makes sense that Google would lean on its already established tech when developing this feature.

Of course, not all languages can be easily translated. What you can do, however, is practice face-to-face communication skills. A study by the Harvard Business School and the University of Chicago finds that hand-shaking and other social interactions in a business setting promote "cooperative strategies and influences negotiation outcomes." The verbal and nonverbal communication that takes place during an in-person meeting is essential to the business relationship. Nothing would be possible without your employees: They are the core of your company and you rely on them to build partnerships and foster overall growth.

A newly-discovered glitch in Google Translate is causing the online tool to transform gibberish suggestions into doomsday warnings and prophesies about Jesus. The AI that powers Google Translate starts to produce the nonsensical warnings about the end of the world when asked to translate the phrase'dog dog dog dog dog dog dog dog dog' from Hawaiian to English. The nonsense sentence, when translated, throws up references to the doomsday clock and the second coming of Jesus Christ. Once the glitch was discovered, Google Translate fans quickly flooded social media with variations on the phrase, mocking the bizarre results thrown-up by the AI. Google Translate has been malfunctioning recently, spouting prophetic verses from gibberish.