In RoboCup, several To assist automated learning of team behavior, we provide a large dataset generated using 10different leagues exist to emphasize specific research problems by using different kinds of the top participants in RoboCup 2016 or 2017. of robots and rules. There are different soccer While it is possible to use the simulator for robot leagues in the RoboCup with different types and learning, we also generate additional data that is sizes of hardware and software: small size, middle not normally available from playing other teams size, standard platform league, humanoid, 2D directly: We modified the simulator to record and 3D simulation (Kitano et al., 1997). In the data from each robots local perspective, that is, soccer simulation leagues (Akiyama et al., 2015), with the restricted views that depend on each the emphasis is on multi-robot team work with robots situation and actions, and also include partial and noisy information, in real-time.

Stochastic optimization of continuous objectives is at the heart of modern machine learning. However, many important problems are of discrete nature and often involve submodular objectives. We seek to unleash the power of stochastic continuous optimization, namely stochastic gradient descent and its variants, to such discrete problems. We first introduce the problem of stochastic submodular optimization, where one needs to optimize a submodular objective which is given as an expectation. Our model captures situations where the discrete objective arises as an empirical risk (e.g., in the case of exemplar-based clustering), or is given as an explicit stochastic model (e.g., in the case of influence maximization in social networks). By exploiting that common extensions act linearly on the class of submodular functions, we employ projected stochastic gradient ascent and its variants in the continuous domain, and perform rounding to obtain discrete solutions. We focus on the rich and widely used family of weighted coverage functions. We show that our approach yields solutions that are guaranteed to match the optimal approximation guarantees, while reducing the computational cost by several orders of magnitude, as we demonstrate empirically.

We study the fundamental problem of learning the parameters of a high-dimensional Gaussian in the presence of noise -- where an $\varepsilon$-fraction of our samples were chosen by an adversary. We give robust estimators that achieve estimation error $O(\varepsilon)$ in the total variation distance, which is optimal up to a universal constant that is independent of the dimension. In the case where just the mean is unknown, our robustness guarantee is optimal up to a factor of $\sqrt{2}$ and the running time is polynomial in $d$ and $1/\epsilon$. When both the mean and covariance are unknown, the running time is polynomial in $d$ and quasipolynomial in $1/\varepsilon$. Moreover all of our algorithms require only a polynomial number of samples. Our work shows that the same sorts of error guarantees that were established over fifty years ago in the one-dimensional setting can also be achieved by efficient algorithms in high-dimensional settings.

Kernel embeddings of distributions and the Maximum Mean Discrepancy (MMD), the resulting distance between distributions, are useful tools for fully nonparametric two-sample testing and learning on distributions. However, it is rarely that all possible differences between samples are of interest -- discovered differences can be due to different types of measurement noise, data collection artefacts or other irrelevant sources of variability. We propose distances between distributions which encode invariance to additive symmetric noise, aimed at testing whether the assumed true underlying processes differ. Moreover, we construct invariant features of distributions, leading to learning algorithms robust to the impairment of the input distributions with symmetric additive noise.

Recent work has shown that state-of-the-art classifiers are quite brittle, in the sense that a small adversarial change of an originally with high confidence correctly classified input leads to a wrong classification again with high confidence. This raises concerns that such classifiers are vulnerable to attacks and calls into question their usage in safety-critical systems. We show in this paper for the first time formal guarantees on the robustness of a classifier by giving instance-specific lower bounds on the norm of the input manipulation required to change the classifier decision. Based on this analysis we propose the Cross-Lipschitz regularization functional. We show that using this form of regularization in kernel methods resp.

We investigate properties of ABA+, a formalism that extends the well studied structured argumentation formalism Assumption-Based Argumentation (ABA) with a preference handling mechanism. In particular, we establish desirable properties that ABA+ semantics exhibit. These pave way to the satisfaction by ABA+ of some (arguably) desirable principles of preference handling in argumentation and nonmonotonic reasoning, as well as non-monotonic inference properties of ABA+ under various semantics.

With the rise of cloud computing and the ever-decreasing costs associated with computations, now and in the future this technology will be more and more widely available. One of the most process heavy steps in AI systems is model training and validation. Being able to pay per minute or even second for the use of computing power removes the need for large upfront investment and data centre maintenance costs. With Google Cloud, IBM Bluemix and Amazon Cloud the power to perform highly complex computations is readily available for everyone today [11]. The systems architecture for machine learning which underpins artificial intelligence is also seamlessly provided by cloud solutions.

The board game Go is older and more complex than chess. While it's been 20 years since IBM's Deep Blue beat world chess champion Garry Kasparov, computers only started beating Go experts a few years ago. An Oct. 18 report in the science journal Nature tells us that this particular man/machine contest is done. A system built by the DeepMind unit of Alphabet (ticker: GOOGL) beat Go's reigning world champ 100 games to none. The deposed champ, you should know, is a prior version of the same artificial intelligence system, which beat one of humankind's international champions in 2016.

Conventional reinforcement learning methods for Markov decision processes rely on weakly-guided, stochastic searches to drive the learning process. It can therefore be difficult to predict what agent behaviors might emerge. In this paper, we consider an information-theoretic cost function for performing constrained stochastic searches that promote the formation of risk-averse to risk-favoring behaviors. This cost function is the value of information, which provides the optimal trade-off between the expected return of a policy and the policy's complexity; policy complexity is measured by number of bits and controlled by a single hyperparameter on the cost function. As the policy complexity is reduced, the agents will increasingly eschew risky actions. This reduces the potential for high accrued rewards. As the policy complexity increases, the agents will take actions, regardless of the risk, that can raise the long-term rewards. The obtainable reward depends on a single, tunable hyperparameter that regulates the degree of policy complexity. We evaluate the performance of value-of-information-based policies on a stochastic version of Ms. Pac-Man. A major component of this paper is the demonstration that ranges of policy complexity values yield different game-play styles and explaining why this occurs. We also show that our reinforcement-learning search mechanism is more efficient than the others we utilize. This result implies that the value of information theory is appropriate for framing the exploitation-exploration trade-off in reinforcement learning.

Video Friday is your weekly selection of awesome robotics videos, collected by your Automaton bloggers. We'll also be posting a weekly calendar of upcoming robotics events for the next two months; here's what we have so far (send us your events!): Let us know if you have suggestions for next week, and enjoy today's videos. We already posted about the unveiling of Sony's new Aibo, but here's a bit of extra video from the event showing the little robotic dog in live action: In this video we show a compilation of our research for the last 4 years on autonomous navigation of bipedal robots. It is part of the DFG-founded project "Versatile and Robust Walking in Uneven Terrain" (German Research Foundation) and includes development in environment perception and modeling, motion planning and stability control.