We propose a new regularization method based on virtual adversarial loss: a new measure of local smoothness of the conditional label distribution given input. Virtual adversarial loss is defined as the robustness of the conditional label distribution around each input data point against local perturbation. Unlike adversarial training, our method defines the adversarial direction without label information and is hence applicable to semi-supervised learning. Because the directions in which we smooth the model are only "virtually" adversarial, we call our method virtual adversarial training (VAT). The computational cost of VAT is relatively low. For neural networks, the approximated gradient of virtual adversarial loss can be computed with no more than two pairs of forward- and back-propagations. In our experiments, we applied VAT to supervised and semi-supervised learning tasks on multiple benchmark datasets. With a simple enhancement of the algorithm based on the entropy minimization principle, our VAT achieves state-of-the-art performance for semi-supervised learning tasks on SVHN and CIFAR-10.

Abstract:The unclear developmentdirection of human society is a deep reason for that it is difficult to form a uniform ethical standard for human society and artificial intelligence. Through the study of the Internet brain model, AI's IQ evaluation, and the evolution of the brain, this paper proposes that the evolution of population knowledge base is the key for judging the development direction of human society,thereby discussing the standards and norms for the construction of artificial intelligence ethics. Keywords: Direction of Evolution; Population Knowledge Base; Artificial Intelligence Ethics 1. Dilemma for Formation of Human and Artificial Intelligence Ethics Like the cases of intelligence, consciousness, life and universe, it is difficult to give ethics a uniform definition. Generally,ethics refers to the principles and guidelines that should be followed when one deals with the people-to-people and people-to-society relationships, and it is a philosophical reflection on moral phenomena from a conceptual perspective. The ethical impact from the artificial intelligence has become more prominent because of the incompleteness and controversy of ethical issues.

Online symptom checkers have significant potential to improve patient care, however their reliability and accuracy remain variable. We hypothesised that an artificial intelligence (AI) powered triage and diagnostic system would compare favourably with human doctors with respect to triage and diagnostic accuracy. We performed a prospective validation study of the accuracy and safety of an AI powered triage and diagnostic system. Identical cases were evaluated by both an AI system and human doctors. Differential diagnoses and triage outcomes were evaluated by an independent judge, who was blinded from knowing the source (AI system or human doctor) of the outcomes. Independently of these cases, vignettes from publicly available resources were also assessed to provide a benchmark to previous studies and the diagnostic component of the MRCGP exam. Overall we found that the Babylon AI powered Triage and Diagnostic System was able to identify the condition modelled by a clinical vignette with accuracy comparable to human doctors (in terms of precision and recall). In addition, we found that the triage advice recommended by the AI System was, on average, safer than that of human doctors, when compared to the ranges of acceptable triage provided by independent expert judges, with only a minimal reduction in appropriateness.

China's government and businesses are trying to attract top-caliber engineers, scientists and other skilled technical workers, particularly U.S.-based Chinese professionals, according to officials, recruiting agents and those being approached. Silicon Valley, with its nexus of big tech firms, research labs and venture investors, is a prime target. Early this year, a leading Chinese government adviser on technology policy addressed a packed convention center hall here to promote the prospects created by China's demand for cutting-edge technologies and its deep pockets. "We've gathered you all here to explore new opportunities for cooperation," said Ye Tianchun, speaking in Mandarin. Over 300 people, the majority of them Chinese and Chinese Americans, showed up to hear Mr. Ye, who guides China's research into semiconductors.

A panel of AI experts have been grilled on the impact and importance of artificial general intelligence by the US House of Representatives on Tuesday. The hearing was ominously named "Artificial Intelligence – With Great Power Comes Great Responsibility." Narrow AI for specific tasks has been rapidly advancing and the committee wanted to know how far off artificial general intelligence (AGI), where a system can learn multiple actions and do them better than humans, would be. Greg Brockman, co-founder and CTO at OpenAI, defined AGI as "highly autonomous systems that outperform humans at most economically valuable work". Progress is spearheaded by three areas: data, computation, and algorithms. A recent OpenAI study estimated there had been a 300,000-times increase in the amount of compute used to train AI systems since 2012.

In October 2017, China's State Council laid out a plan to develop a domestic artificial intelligence (AI) industry worth $150 billion by 2030, with the aim of becoming a global leader in AI technologies. President Xi Jinping called upon Chinese companies to "step up, and own the core technologies." It is no surprise, then, that the Chinese SenseTime was the highest-valued AI startup, though it held this title for a little less than a month. It was overtaken, unsurprisingly, by another Chinese startup: UBTECH Robotics, valued at an estimated $5 billion. This AI race has spilled over to healthcare, and to medical image analysis, which is the bellwether application area for healthcare AI.

The technology is used in China to make kids pay attention and in California to order burgers. You can of course use your face to unlock your iPhone, but the tech also has the potential to screen passengers at US airports and recognize criminals. These last two uses are problematic, as the tech isn't ready to handle darker skin tones and genders. Microsoft hopes to help fix this problem with improved facial recognition technology the company claims has reduced error rates for men and women with darker skin by 20 percent. According to Microsoft, commercially-available software performs best on males with lighter skin and the worst on females with darker skin.

A former top U.S. Defense Department official said Tuesday he was "alarmed" by a recent decision by Alphabet Inc.'s Google to withdraw from work on a department initiative applying artificial intelligence tools to analyzing drone footage. "I fully agree that it might wind up with us taking a shot, but it could easily save lives," said Deputy Defense Secretary Robert O. Work, who started the initiative known as "Project Maven." "I believe the Google employees created an enormous moral hazard for themselves." More than 4,000 Google employees signed a petition calling for the cancellation of the Project Maven contract, citing Google's history of avoiding military work and worries about autonomous weapons. The company has since pledged not to use its AI for weapons, illegal surveillance and technologies that cause "overall harm," while saying it will continue to pursue government work in other areas.

Google employees have spoken out against the project and their opposition ultimately led to the company deciding not to renew the contract when it expires next year. Now, former Deputy Defense Secretary Robert Work, who started the Project Maven initiative, is saying he's "alarmed" by Google's decision to walk away from the program. Bloomberg reports that Work shared his concerns about the move during an event on tech in the military held today in Washington. "I fully agree that it might wind up with us taking a shot, but it could easily save lives. I believe the Google employees created an enormous moral hazard for themselves."

When a group of about 40 players first tested out a live game called the Machine Learning President at a private event in San Francisco this February, they were unaware that the game would end up memorialized in the pages of The New Yorker. But during a ski vacation in March, the Republican mega-donor Rebekah Mercer gathered her friends together to play several rounds of the game, which pits special interest groups, political candidates, and activist organizations against each other in a simulated presidential election, aided by cash and artificial intelligence. A lawyer for Mercer told The New Yorker that she owned a copy of the Machine Learning President but had not created it and that it did not reflect her family's views. It's not hard to draw comparisons between the rules of the game, with its reliance on big cash and tech capabilities, and the actions of the Mercer-backed Cambridge Analytica during the 2016 U.S. presidential election. But, as Mercer's lawyer stated, she had nothing to do with creating the game--in fact, it was conceptualized by one of her vocal critics.