In the world's wealthiest neighbourhoods, artificial intelligence (AI) systems are starting to steer self-driving cars down the streets, and homeowners are giving orders to their smart voice-controlled speakers. But the AI revolution has yet to offer much help to the 3 billion people globally who live in poverty. That discrepancy lies at the heart of a meeting in Geneva, Switzerland, on 7–9 June, grandly titled the AI for Good Global Summit. The meeting of United Nations agencies, AI experts, policymakers and industrialists will discuss how AI and robotics might be guided to address humanity's most enduring problems, such as poverty, malnutrition and inequality. Development agencies are buzzing with ideas, although only a few have reached the stage of pilot experiments. But scientists caution that the rise of AI will also bring societal disruption that will be hard to foresee or manage, and that could harm the world's most disadvantaged.

With a simple architecture and the ability to learn meaningful word embeddings efficiently from texts containing billions of words, word2vec remains one of the most popular neural language models used today. However, as only a single embedding is learned for every word in the vocabulary, the model fails to optimally represent words with multiple meanings. Additionally, it is not possible to create embeddings for new (out-of-vocabulary) words on the spot. Based on an intuitive interpretation of the continuous bag-of-words (CBOW) word2vec model's negative sampling training objective in terms of predicting context based similarities, we motivate an extension of the model we call context encoders (ConEc). By multiplying the matrix of trained word2vec embeddings with a word's average context vector, out-of-vocabulary (OOV) embeddings and representations for a word with multiple meanings can be created based on the word's local contexts. The benefits of this approach are illustrated by using these word embeddings as features in the CoNLL 2003 named entity recognition (NER) task.

Neural networks have been successfully applied in applications with a large amount of labeled data. However, the task of rapid generalization on new concepts with small training data while preserving performances on previously learned ones still presents a significant challenge to neural network models. In this work, we introduce a novel meta learning method, Meta Networks (MetaNet), that learns a meta-level knowledge across tasks and shifts its inductive biases via fast parameterization for rapid generalization. When evaluated on Omniglot and Mini-ImageNet benchmarks, our MetaNet models achieve a near human-level performance and outperform the baseline approaches by up to 6% accuracy. We demonstrate several appealing properties of MetaNet relating to generalization and continual learning.

Exploratory data analysis just got more fun. Feel free to use the Gitter community gitter.im/scattertext If you cannot (or don't want to) install spaCy, substitute nlp spacy.en.English() lines with nlp scattertext.WhitespaceNLP.whitespace_nlp. Note, this is not compatible with word_similarity_explorer, and the tokenization and sentence boundary detection capabilities will be low-performance regular expressions. Python 2.7 support is experimental.

After receiving more than 18,300 applications, NASA has finally announced its new class of astronauts – some of whom could move on to deep-space missions aboard the Orion spacecraft. The space agency introduced 12 men and women today on stage at the Johnson Space Center in Houston, during an event that was attended by Vice President Mike Pence. Vice President Pence wished'Godspeed' to the new class, and revealed the Trump administration will be reopening the National Space Council, with Pence as a chair, in efforts to'ensure that America will never again lose our lead in space exploration and space innovation technology.' The lineup includes: Kayla Barron, Zena Cardon, Raja Chari, Matthew Dominick, Bob Hines, Dr Warren'Woody' Hoburg, Jonny Kim, Robb Kulin, Jasmin Moghbeli, Loral O'Hara, Dr Frank Rubio, Jessica Watkins The chosen few will undergo two years of training, after which they will be assigned to various missions, including research on the International Space Station, launches aboard commercial spacecraft, and even deep-space exploration. After brief introductions from Johnson Center Director Ellen Ochoa and the showing of a video from current astronauts welcoming the newcomers, Flight Operations Director Brian Kelly introduced the new candidates one by one, in alphabetical order. The lineup includes: Kayla Barron, Zena Cardon, Raja Chari, Matthew Dominick, Bob Hines, Dr Warren'Woody' Hoburg, Jonny Kim, Robb Kulin, Jasmin Moghbeli, Loral O'Hara, Dr Frank Rubio, Jessica Watkins.

If you've been scratching your head at the FAA's extensive efforts to regulate your personal (or company) drone use, consider the chaos when they start filling the skies. That's why the agency partnered with NASA for a series of nationwide tests to explore potential systems that could track and manage a wide range of drones simultaneously. Google parent company Alphabet's Project Wing tried out its own UAV air traffic control platform yesterday, a system that might one day guide a massive volume of airborne drones to keep them from crashing into buildings, people or each other. Unsurprisingly, Project Wing's UTM (UAS Air Traffic Management) leans heavily on other Google products like Maps, Earth and Street View to navigate drones around obstacles and plan routes. During yesterday's tests, UTM managed flight paths for multiple UAVs simultaneously, according to the group's blog post.

The laws governing the use of drones in the United States are changing so fast it can be hard to keep up. But I'd like to explore here some proposed drone rules that never went into effect because the legislation that described them, Senate bill 2658 (the Federal Aviation Administration Reauthorization Act of 2016), was never passed. Why care about rules that didn't become law? It's my theory that although the legislation died in Congress last year, the people championing various parts of it are still around and may yet influence future laws. So an examination of the ill-fated legislation could provide a window on what the future holds for drone operators.

South Wales Police carried out the U.K.'s first arrest using facial recognition, Ars Technica UK reported. The arrest using automatic facial recognition (AFR) was made on May 31 and was not related to the Champions League final. It's unclear whether the apprehension was due to authorities testing the technology prior to the match. International Business Times has reached out to South Wales Police regarding the arrest. The department previously said there was "a very low number of arrests over the period of the festival including match day," in a statement on Tuesday.

The text analysis part of the AMiCA project (http://www.amicaproject.be), a cooperation between the University of Antwerp and the University of Ghent, developed methods and software to help moderators detect occurrences of unwanted or dangerous situations in their social networks. More specifically, the project developed prototype systems for the detection of cyberbullying, suicide announcements, and sexually transgressive behavior. In this talk I will focus on the text analysis methods that were used for normalization of social media text, for profiling users, and for detecting dangerous content. I will describe the architectures and results of the three resulting applications.

For militaries, tech like this provides an excellent platform for surveillance missions against military and terrorist targets. It can utilize its high flight ceiling to maintain line-of-sight contact with over 400,000 square miles of ground and water. For both militaries and tech firms, covering so much territory makes it an excellent data relay and communications node. This will allow the drone to replace or back up satellite communications, maintain coverage between distant aircraft and ships, or even provide broadband to rural Chinese households. While conversations around drone usage are often limited to their roles as potential missile-toting killers and parcel-delivering quadcopters, some of the most important drones of the future may be those like the Caihong X and Helios Prototype, unseen and high up, gathering data day in and day out.