HONG KONG – Alphabet Inc.'s Google is preparing to launch a censored version of its search engine for China that will block results Beijing considers sensitive, The Intercept reported. Google's been working on a project code-named Dragonfly since the spring of 2017 and demonstrated a sanitized version of its search app to Chinese officials, the news outlet reported, citing company documents and unidentified people familiar with the matter. A final version of the app could be launched within six to nine months, it said. "We provide a number of mobile apps in China, such as Google Translate and Files Go, help Chinese developers, and have made significant investments in Chinese companies like JD.com. But we don't comment on speculation about future plans," Google said in an emailed statement.

In 2010, Google made a moral calculus. The company had been censoring search results in China at the behest of the Communist government since launching there in 2006. But after a sophisticated phishing attack to gain access to the Gmail accounts of Chinese human rights activists, Google decided to stop censoring results, even though it cost the company access to the lucrative Chinese market. Across nearly a decade, Google's decision to weigh social good over financial profit became part of Silicon Valley folklore, a handy anecdote that cast the tech industry as a democratizing force in the world. But to tech giants with an insatiable appetite for growth, China's allure is just as legendary.

For many years, a combination of principal component analysis (PCA) and independent component analysis (ICA) has been used as a blind source separation (BSS) technique to separate hidden sources of natural data. However, it is unclear why these linear methods work well because most real-world data involve nonlinear mixtures of sources. We show that a cascade of PCA and ICA can solve this nonlinear BSS problem accurately as the variety of input signals increases. Specifically, we present two theorems that guarantee asymptotically zero-error BSS when sources are mixed by a feedforward network with two processing layers. Our first theorem analytically quantifies the performance of an optimal linear encoder that reconstructs independent sources. Zero-error is asymptotically reached when the number of sources is large and the numbers of inputs and nonlinear bases are large relative to the number of sources. The next question involves finding an optimal linear encoder without observing the underlying sources. Our second theorem guarantees that PCA can reliably extract all the subspace represented by the optimal linear encoder, so that a subsequent application of ICA can separate all sources. Thereby, for almost all nonlinear generative processes with sufficient variety, the cascade of PCA and ICA performs asymptotically zero-error BSS in an unsupervised manner. We analytically and numerically validate the theorems. These results highlight the utility of linear BSS techniques for accurately recovering nonlinearly mixed sources when observations are sufficiently diverse. We also discuss a possible biological BSS implementation.

Singapore's Info-communications Media Development Authority (IMDA) recently announced the creation of an Advisory Council on Ethical Use of AI and Data as part of an effort to bring together a range of key stakeholders to inform the government on possible approaches to ensure consumer trust in AI-powered products and services. ITU News recently caught up with IMDA's Assistant Chief Executive of Data Innovation and Protection, Yeong Zee Kin, to learn more about Singapore's approach to this important and timely issue. With the recent launch of the Digital Economy Framework for Action, Singapore has entered a new phase of its digitalisation journey. The ability to use and share data innovatively and responsibly can become a competitive advantage for businesses. Infusing AI into business operations can accelerate digital transformation through new features and functionalities.

The deal is with Paltac Corporation, the biggest supplier to drugstore-style shops in Japan. It begins modestly, with a single pilot store in the city of Sendai, about four hours north of Tokyo, set to open in early 2019. Then it ramps up fast: The plan is to outfit over 3,000 stores in time for the Tokyo Olympics in July 2020. "The government is pushing its stores and its companies to put their best digital foot forward for the Olympics," says Michael Suswal, Standard Cognition's COO and one of the Bay Area startup's seven cofounders. Partnering with Paltac, which supplies most of Japan's small retail industry, allows Standard Cognition to reach a diverse market.

Airborne data collecting platforms like the RQ-4 Global Hawk have a problem: the usefulness of the data they collect is limited by how fast and how well it can be analyzed. US military intelligence gathers a lot of data, but in order to make the data useful for a decision making process, the Air Force needs a "sensing grid that fuses together data," C4ISRNET reported Wednesday. AI will help the force interpret that fused data. The AI will harvest information from airborne systems in development such as Gremlin drones, which the US military portrays as a swarm of small drones that take off from an aircraft mid-flight and are recovered by the same aircraft. "How do I get the data so I can fuse it, look at it and then ask the right questions from the data to reveal what trends are out there?" Lt. Gen. VeraLinn Jamieson said in a July 31 interview with the news outlet.

On this week's If Then, Will Oremus and April Glaser talk about a new study that suggests the internet might not have played the crucial role in Trump's election victory that we previously assumed. The hosts are joined by Justin Erlich, the new VP of strategy, policy, and legal at Voyage, a self-driving vehicle company in Silicon Valley. Before that, he was head of policy for autonomous vehicles and urban aviation at Uber. The hosts discuss when these "cars" will hit the skies, what this means for investment in public transit, and how we'll know they'll be safe.

'The U.S. Department of Defense will for the first time be using large-scale artificial intelligence systems that could automate mundane tasks and augment the work of military members as a result of an $885 million five-year contract, said Josh Sullivan, senior vice president at government consulting firm Booz Allen Hamilton. The technology will allow the Defense Department to better compete with nations including China and Russia, said Mr. Sullivan, who leads the analytics business for Booz Allen. "Part of this is (about) making sure our government has the access to the best technology and using it responsibly in service of our citizens and warfighters," he said. The use of AI systems such as neural networks that mimic the human brain could help the Defense Department sift through the "overwhelming" amount of data related to such areas as national security and health care. In turn, soldiers and military members can be freed up to identify threats on the battlefield sooner, spend more time with military patients and work on problems that require higher-level contextual reasoning, Mr. Sullivan said.

A poll of British consumers conducted by software firm OpenText found that one in four Brits think robots would do a better job than humans as politicians. Years ago, The Muppet Show ran a segment mocking politicians for their stereotypical robotic behaviour. Last April a robot was nominated to run to be Tokyo's mayor, promising fair and balanced representation. In a world where reality is sometimes more bizarre than an episode of Black Mirror, what if we replaced our current politicians with algorithms? In a period where trust in politicians is low and government efficiency is questionable, might we be better off? Upgrade your inbox and get our Daily Dispatch and Editor's Picks.