We study a two-level multiview learning with more than two views under the PAC-Bayesian framework. This approach, sometimes referred as late fusion, consists in learning sequentially multiple view-specific classifiers at the first level, and then combining these view-specific classifiers at the second level. Our main theoretical result is a generalization bound on the risk of the majority vote which exhibits a term of diversity in the predictions of the view-specific classifiers. From this result it comes out that controlling the trade-off between diversity and accuracy is a key element for multiview learning, which complements other results in multiview learning. Finally, we experiment our principle on multiview datasets extracted from the Reuters RCV1/RCV2 collection.

Auto-Encoders are unsupervised models that aim to learn patterns from observed data by minimizing a reconstruction cost. The useful representations learned are often found to be sparse and distributed. On the other hand, compressed sensing and sparse coding assume a data generating process, where the observed data is generated from some true latent signal source, and try to recover the corresponding signal from measurements. Looking at auto-encoders from this \textit{signal recovery perspective} enables us to have a more coherent view of these techniques. In this paper, in particular, we show that the \textit{true} hidden representation can be approximately recovered if the weight matrices are highly incoherent with unit $ \ell^{2} $ row length and the bias vectors takes the value (approximately) equal to the negative of the data mean. The recovery also becomes more and more accurate as the sparsity in hidden signals increases. Additionally, we empirically demonstrate that auto-encoders are capable of recovering the data generating dictionary when only data samples are given.

For sophisticated reinforcement learning (RL) systems to interact usefully with real-world environments, we need to communicate complex goals to these systems. In this work, we explore goals defined in terms of (non-expert) human preferences between pairs of trajectory segments. We show that this approach can effectively solve complex RL tasks without access to the reward function, including Atari games and simulated robot locomotion, while providing feedback on less than one percent of our agent's interactions with the environment. This reduces the cost of human oversight far enough that it can be practically applied to state-of-the-art RL systems. To demonstrate the flexibility of our approach, we show that we can successfully train complex novel behaviors with about an hour of human time. These behaviors and environments are considerably more complex than any that have been previously learned from human feedback.

Many real-world tasks involve multiple agents with partial observability and limited communication. Learning is challenging in these settings due to local viewpoints of agents, which perceive the world as non-stationary due to concurrently-exploring teammates. Approaches that learn specialized policies for individual tasks face problems when applied to the real world: not only do agents have to learn and store distinct policies for each task, but in practice identities of tasks are often non-observable, making these approaches inapplicable. This paper formalizes and addresses the problem of multi-task multi-agent reinforcement learning under partial observability. We introduce a decentralized single-task learning approach that is robust to concurrent interactions of teammates, and present an approach for distilling single-task policies into a unified policy that performs well across multiple related tasks, without explicit provision of task identity.

Prediction in a small-sized sample with a large number of covariates, the "small n, large p" problem, is challenging. This setting is encountered in multiple applications, such as precision medicine, where obtaining additional samples can be extremely costly or even impossible, and extensive research effort has recently been dedicated to finding principled solutions for accurate prediction. However, a valuable source of additional information, domain experts, has not yet been efficiently exploited. We formulate knowledge elicitation generally as a probabilistic inference process, where expert knowledge is sequentially queried to improve predictions. In the specific case of sparse linear regression, where we assume the expert has knowledge about the values of the regression coefficients or about the relevance of the features, we propose an algorithm and computational approximation for fast and efficient interaction, which sequentially identifies the most informative features on which to query expert knowledge. Evaluations of our method in experiments with simulated and real users show improved prediction accuracy already with a small effort from the expert.

Answer Set Programming (ASP) is a well-established declarative paradigm. One of the successes of ASP is the availability of efficient systems. State-of-the-art systems are based on the ground+solve approach. In some applications this approach is infeasible because the grounding of one or few constraints is expensive. In this paper, we systematically compare alternative strategies to avoid the instantiation of problematic constraints, that are based on custom extensions of the solver. Results on real and synthetic benchmarks highlight some strengths and weaknesses of the different strategies. (Under consideration for acceptance in TPLP, ICLP 2017 Special Issue.)

While NASA stays diligently focused on its marathon to Mars, it seems everyone else in space exploration wants to sprint to the moon. Between Vice President Mike Pence taking aim at the moon in a speech last week, SpaceX's plans to carry two rich people into lunar orbit in 2018, and the Google Lunar XPrize--a literal moon race for private companies--our lonely satellite is becoming the center of attention lately. Moon Express is one of the companies hoping to take home the $20 million Lunar XPrize for being the first commercial company to land on the moon, travel 500 meters, and send images back to Earth. Today, it released plans for building a robotic outpost on the moon and returning a sample of moon dust to Earth by 2020. "Outpost" is a bit of a strong word here.

The first colony on the moon might be full of robots instead of people, if a plan from the private space company Moon Express works out. It is aiming to build an unmanned lunar outpost, supported by a fleet of spacecraft explorers that would launch from low Earth orbit and deliver supplies to and land on the moon, according to The Verge. The MX explorer fleet would rely on the Electron rocket that the space company Rocket Lab has designed to launch out of New Zealand in order to get into space before heading to the moon. The lunar station, at the moon's south pole, could come as soon as 2020. The plan is to have robots staying around the clock on the moon to mine it for resources like minerals and then sell them, The Verge explains.

Spaceflight company Moon Express has released its plans to mine the moon with robots and it aims to get started by 2020. The company was founded in 2010 with the aim of winning the Google Lunar Xprize -- a competition to get privately funded spacecraft on the moon. And while it still has its sights on that prize, Moon Express has planned beyond that and has laid out a strategy for establishing its lunar outpost in just a few years. The company's first moon mission, dubbed Lunar Scout, will use Rocket Lab's Electron rocket to send its MX-1E robotic explorer to land on and deliver several payloads -- including the International Lunar Observatory -- to the moon. If completed by the end of this year and before the four other finalists for the Lunar Xprize, this mission could win Moon Express the competition's $20 million prize.

Chinese artificial intelligence start-up SenseTime has completed a US$410 million series B round, in what the company calls the largest private financing rounds ever closed by an AI start-up globally. The company said it is now valued at over RMB10 billion (US$1.47 billion) in an announcement, revealing its valuation for the first time. It means another company has joined China Money Network's China Unicorn Ranking, which currently counts 108 unicorns in China with aggregate valuation of US$445.66 billion. The round includes two portions: a series B1 round, which was led by Chinese private equity firm CDH Investments, and a B2 round led by Chinese investment firm Sailing Capital. China Money Network previously reported that SenseTime raised US$120 million last December and US$60 million in April, both were part of the series B round.