The classic example of "crowd wisdom" dates back to 1906, when Sir Francis Galton observed a contest in which attendees were asked to guess the weight of an ox at a country fair in England. In what many consider to be the first experiment on crowd wisdom, the average of the 800 guesses was within one pound of being correct. Consider that these kinds of experiments can now be done digitally – across cultures and time zones and fairly instantaneously. The classic experiment was reenacted recently with a digital crowd when a photo of a cow was posted online and viewers were invited to guess her weight. More than 17,000 votes were cast and the average guess was within 5 percent of being accurate.

As part of its projects with Britain's National Health Service, Google's artificial intelligence unit DeepMind announced last week it's developing a new way to protect confidential health data--from itself. Its problem: How to assure hospitals, and the public at large, that patient confidentiality isn't compromised as it processes the sensitive medical health records entrusted to it. DeepMind's proposed solution is to create an indelible data log that can't be tampered with. It would show when a piece of data was used, and for what purpose. Importantly, DeepMind itself wouldn't be able to modify logs to use the data nefariously.

The company is looking to develop partnerships and opportunities with other organisations around the world to, create the capability, competencies and jobs to supply the technology and components required. On March 8, 2017, the latest part of that programme has been confirmed with the announcement, in Finland, of a significant research grant by Tekes - the Finnish Funding Agency for Innovation. The funding will enable Rolls-Royce to invest further in a research and development centre in Turku, Finland. The company plans to carry out further development projects there focused on the future development of land-based control centres, and the use of artificial intelligence in future remote and autonomous shipping operations. Mikael Mäkinen, Rolls-Royce, President – Marine said: "Digitalisation will transform the shipping industry in the years ahead, and the time is now right to set out how we are going to make this happen. Over the coming years we need to invest globally to develop the required capabilities and to establish a range of market-ready products and systems to take advantage of what is a significant global market opportunity."

Tata Consultancy Services (BSE: 532540, NSE: TCS), a leading global IT services, consulting and business solutions organization, today unveiled its Global Trend Study titled, "Getting Smarter by the Day: How AI is Elevating the Performance of Global Companies." Focused on the current and future impact of Artificial Intelligence (AI), the study polled 835 executives across 13 global industry sectors in four regions of the world, finding that 84% of companies see the use of AI as "essential" to competitiveness, with a further 50% seeing the technology as "transformative." Widespread AI adoption expected across job functions Exploring the views and actions of decision makers from global companies with average revenues of $20 billion, the study revealed AI is spreading across almost all areas of a company. The biggest adopters of AI today are, not surprisingly, IT departments, with two-thirds (67%) of survey respondents using AI to detect security intrusions, user issues and deliver automation. However, by 2020, almost a third (32%) of companies believe AI's greatest impact will be in sales, marketing or customer service, while one in five (20%) see AI's impact being largest in non-customer facing corporate functions, including finance, strategic planning, corporate development, and HR.

Certain kinds of autonomous vehicles may not be safe, especially in an emergency situation, according to a new study published by the Lords Science and Technology Committee on Wednesday. With driverless technology, drivers may become over-reliant and complacent. However, with the development in the automotive technology over time, accidents by human error may be significantly reduced -- but they just might increase before they do. The committee also reported people may use driverless cars for shorter distances, as well, causing laziness and may prevent them from "getting exercise by walking." The UK Economic Opportunity split vehicles into levels from 0 to 5. Zero was fully controlled by an individual, and five was completely automated. According to peers on the committee, there was a "very dangerous" problem with vehicles reaching the middle of the scale, BBC News reported.

This work addresses various open questions in the theory of active learning for nonparametric classification. Our contributions are both statistical and algorithmic: -We establish new minimax-rates for active learning under common \textit{noise conditions}. These rates display interesting transitions -- due to the interaction between noise \textit{smoothness and margin} -- not present in the passive setting. Some such transitions were previously conjectured, but remained unconfirmed. -We present a generic algorithmic strategy for adaptivity to unknown noise smoothness and margin; our strategy achieves optimal rates in many general situations; furthermore, unlike in previous work, we avoid the need for \textit{adaptive confidence sets}, resulting in strictly milder distributional requirements.

In many data exploration tasks it is meaningful to identify groups of attribute interactions that are specific to a variable of interest. For instance, in a dataset where the attributes are medical markers and the variable of interest (class variable) is binary indicating presence/absence of disease, we would like to know which medical markers interact with respect to the binary class label. These interactions are useful in several practical applications, for example, to gain insight into the structure of the data, in feature selection, and in data anonymisation. We present a novel method, based on statistical significance testing, that can be used to verify if the data set has been created by a given factorised class-conditional joint distribution, where the distribution is parametrised by a partition of its attributes. Furthermore, we provide a method, named ASTRID, for automatically finding a partition of attributes describing the distribution that has generated the data. State-of-the-art classifiers are utilised to capture the interactions present in the data by systematically breaking attribute interactions and observing the effect of this breaking on classifier performance. We empirically demonstrate the utility of the proposed method with examples using real and synthetic data.

This report provides an introduction to some Machine Learning tools within the most common development environments. It mainly focuses on practical problems, skipping any theoretical introduction. It is oriented to both students trying to approach Machine Learning and experts looking for new frameworks.

We introduce an architecture based on deep hierarchical decompositions to learn effective representations of large graphs. Our framework extends classic R-decompositions used in kernel methods, enabling nested "part-of-part" relations. Unlike recursive neural networks, which unroll a template on input graphs directly, we unroll a neural network template over the decomposition hierarchy, allowing us to deal with the high degree variability that typically characterize social network graphs. Deep hierarchical decompositions are also amenable to domain compression, a technique that reduces both space and time complexity by exploiting symmetries. We show empirically that our approach is competitive with current state-of-the-art graph classification methods, particularly when dealing with social network datasets.

The aim of this article is to design a moment transformation for Student- t distributed random variables, which is able to account for the error in the numerically computed mean. We employ Student-t process quadrature, an instance of Bayesian quadrature, which allows us to treat the integral itself as a random variable whose variance provides information about the incurred integration error. Advantage of the Student- t process quadrature over the traditional Gaussian process quadrature, is that the integral variance depends also on the function values, allowing for a more robust modelling of the integration error. The moment transform is applied in nonlinear sigma-point filtering and evaluated on two numerical examples, where it is shown to outperform the state-of-the-art moment transforms.