Thanks to the likes of Google, Amazon, and Facebook, the terms artificial intelligence (AI) and machine learning have become much more widespread than ever before. They are often used interchangeably and promise all sorts from smarter home appliances to robots taking our jobs. But while AI and machine learning are very much related, they are not quite the same thing. AI is a branch of computer science attempting to build machines capable of intelligent behaviour, while Stanford University defines machine learning as "the science of getting computers to act without being explicitly programmed". You need AI researchers to build the smart machines, but you need machine learning experts to make them truly intelligent.

The growing popularity of artificial intelligence technology will likely lead to millions of lost jobs, especially among less-educated workers, and could exacerbate the economic divide between socioeconomic classes in the United States, according to a newly released White House report. But that same technology is also essential to improving the country's productivity growth, a key measure of how efficiently the economy produces goods. That could ultimately lead to higher average wages and fewer work hours. For that reason, the report concludes, our economy actually needs more artificial intelligence, not less. To reconcile the benefits of the technology with its expected toll, the report states that the federal government should expand access to education in technical fields and increase the scope of unemployment benefits.

Anyone who buys a drone in the UK may have to register it and take a safety test for the first time under new measures to prevent potential collisions with passenger jets. Measures proposed by ministers also include criminal liability for anyone who flies a drone in "no-fly zones" surrounding airports and prisons, and an increase in fines, which currently cannot exceed a maximum £2,500. Ministers also want to make drones electronically identifiable on the ground, in order to make it easier for police to track devices to their owners. The government has estimated that the drone industry will be worth £127bn by 2025, but ministers believe it will only be a success if it is done safely and with the consent of the public. While there are already strict rules for drone users, the unmanned aerial vehicles, which can operate either under remote control by a person or autonomously by onboard computers, have become increasingly widespread, and are cheaply available in high street shops and on Amazon.

In this Amsterdam Data Science Deep Dive session we will highlight cutting-edge research with a focus on Machine Learning with speakers from academia and industry. There will be an opportunity for audience participation and interaction to discuss the key challenges within this domain.

Jan Tian stood in nervous silence in the departure hall of Beijing Capital International Airport. Beside him, his sister held an envelope containing a thousand yuan, close to her entire year's wages. It was May 1993 and China's capital was humid, its parks ablaze with tulips, crab apples and red azaleas. But Tian, who had graduated from Beijing University a decade earlier and now worked in Vancouver for the video game company Electronic Arts, had not come to sightsee. The previous week, he had received a phone call to say that his father had suffered a stroke and Tian's bosses had booked him an emergency flight to China. After a week, the doctors had given their prognosis: Tian's father would be paralysed down his left side, but would recover. As concern yielded to relief, Tian's thoughts returned to the work he had left behind in Canada. The release date for EA Soccer, his current project, had recently been brought forward, after an executive walked past an office and heard staff, who were playing an early version of the game, whooping with excitement. For the game to be on shelves by Christmas, it would need to be finished by October. They had less than five months. While Tian and his dozen-or-so colleagues believed fervently in the project, EA's other executives were less enamoured.

A French robotics startup has raised €3.3 million ($3.5 million) to build and grow a fleet of mobile robots that help warehouses prepare orders for delivery. The company was founded in 2015 by former GE Healthcare software architect Renaud Heitz and BA Systèmes technical director Romain Moulin, and Exotec Solutions ("Exotec") robots have already been tested across a number of industries. With $3.5 million more in its coffers, the company expects to launch its first robot -- called Exo -- into the wild in early 2017. The most recent round was raised from 360 Capital Partners, Breega Capital, and a handful of its existing investors. The miniature robots are being targeted at any logistics operator that relies on humans to traverse large warehouses picking items off shelves, and it promises to cut employees' average daily distance covered from 15km to 4km per day and to "[double] the productivity" of each worker.

Today, in order to ready the United States for a future in which artificial intelligence (AI) plays a growing role, the White House released a report on Artificial Intelligence, Automation, and the Economy. This report follows up on the Administration's previous report, Preparing for the Future of Artificial Intelligence, which was released in October 2016, and which recommended that the White House publish a report on the economic impacts of artificial intelligence by the end of 2016. Accelerating AI capabilities will enable automation of some tasks that have long required human labor. These transformations will open up new opportunities for individuals, the economy, and society, but they will also disrupt the current livelihoods of millions of Americans. The new report examines the expected impact of AI-driven automation on the economy, and describes broad strategies that could increase the benefits of AI and mitigate its costs.

Artificial neural networks are most commonly trained with the back-propagation algorithm, where the gradient for learning is provided by back-propagating the error, layer by layer, from the output layer to the hidden layers. A recently discovered method called feedback-alignment shows that the weights used for propagating the error backward don't have to be symmetric with the weights used for propagation the activation forward. In fact, random feedback weights work evenly well, because the network learns how to make the feedback useful. In this work, the feedback alignment principle is used for training hidden layers more independently from the rest of the network, and from a zero initial condition. The error is propagated through fixed random feedback connections directly from the output layer to each hidden layer. This simple method is able to achieve zero training error even in convolutional networks and very deep networks, completely without error back-propagation. The method is a step towards biologically plausible machine learning because the error signal is almost local, and no symmetric or reciprocal weights are required. Experiments show that the test performance on MNIST and CIFAR is almost as good as those obtained with back-propagation for fully connected networks. If combined with dropout, the method achieves 1.45% error on the permutation invariant MNIST task.

In this paper, a Wide Learning architecture is proposed that attempts to automate the feature engineering portion of the machine learning (ML) pipeline. Feature engineering is widely considered as the most time consuming and expert knowledge demanding portion of any ML task. The proposed feature recommendation approach is tested on 3 healthcare datasets: a) PhysioNet Challenge 2016 dataset of phonocardiogram (PCG) signals, b) MIMIC II blood pressure classification dataset of photoplethysmogram (PPG) signals and c) an emotion classification dataset of PPG signals. While the proposed method beats the state of the art techniques for 2nd and 3rd dataset, it reaches 94.38% of the accuracy level of the winner of PhysioNet Challenge 2016. In all cases, the effort to reach a satisfactory performance was drastically less (a few days) than manual feature engineering.

Several methods exist to infer causal networks from massive volumes of observational data. However, almost all existing methods require a considerable length of time series data to capture cause and effect relationships. In contrast, memory-less transition networks or Markov Chain data, which refers to one-step transitions to and from an event, have not been explored for causality inference even though such data is widely available. We find that causal network can be inferred from characteristics of four unique distribution zones around each event. We call this Composition of Transitions and show that cause, effect, and random events exhibit different behavior in their compositions. We applied machine learning models to learn these different behaviors and to infer causality. We name this new method Causality Inference using Composition of Transitions (CICT). To evaluate CICT, we used an administrative inpatient healthcare dataset to set up a network of patients transitions between different diagnoses. We show that CICT is highly accurate in inferring whether the transition between a pair of events is causal or random and performs well in identifying the direction of causality in a bi-directional association.