Birg{\'e} and Massart proposed in 2001 the slope heuristics as a way to choose optimally from data an unknown multiplicative constant in front of a penalty. It is built upon the notion of minimal penalty, and it has been generalized since to some 'minimal-penalty algorithms'. This paper reviews the theoretical results obtained for such algorithms, with a self-contained proof in the simplest framework, precise proof ideas for further generalizations, and a few new results. Explicit connections are made with residual-variance estimators-with an original contribution on this topic, showing that for this task the slope heuristics performs almost as well as a residual-based estimator with the best model choice-and some classical algorithms such as L-curve or elbow heuristics, Mallows' C p , and Akaike's FPE. Practical issues are also addressed, including two new practical definitions of minimal-penalty algorithms that are compared on synthetic data to previously-proposed definitions. Finally, several conjectures and open problems are suggested as future research directions.

When temperatures drop, the expansion and contraction of water that seeps into cracks in asphalt can create giant tire-chompers. But these problem spots are too often the final result of damage that has been brewing for a while. "There are things you can do 5 or even 10 years before that happens to push the lifespan of a road," says Benjamin Schmidt, CTO of RoadBotics in Pittsburgh. RoadBotics is using state-of-the-art computer-vision techniques to help local governments better manage roads. The company's machine-learning algorithms process images of the road collected via smartphone.

Machine learning is currently one of the biggest buzzwords in cybersecurity and the tech industry in general, but the phrase is often overused and misapplied, leaving many with their own, incorrect definition. So, how do you cut through all the noise to separate fact from fiction? And how can this tool be best applied to security operations? Machine learning (ML) is an algorithm that gives the software applications it is applied to the ability to autonomously learn from its own environment, then improve operations based on the data collected. It does this without much human supervision or being specifically programmed to do so.

It's part of a field of work that is building machines that can provide real-time help using only limited data as input. Standard machine-learning algorithms often need to process thousands of possibilities before deciding on a solution, which may be impractical in pressurised scenarios where fast adaptation is critical. After Japan's Fukushima nuclear disaster in 2011, for example, robots were sent into the power plant to clear up radioactive debris in conditions far too dangerous for humans. The problem, says robotics researcher Professor Jean-Baptiste Mouret is that the robots kept breaking down or came across hazards that stopped them in their tracks. As part of the ResiBots initiative, he is designing a lower-cost robot that can last long periods without needing constant human maintenance for breakages and are better at overcoming unexpected obstacles.

A reinforcement learning agent tries to maximize its cumulative payoff by interacting in an unknown environment. It is important for the agent to explore suboptimal actions as well as to pick actions with highest known rewards. Yet, in sensitive domains, collecting more data with exploration is not always possible, but it is important to find a policy with a certain performance guaranty. In this paper, we present a brief survey of methods available in the literature for balancing exploration-exploitation trade off and computing robust solutions from fixed samples in reinforcement learning.

Abstract--As cellular networks become denser, a scalable and dynamic tuning of wireless base station parameters can only be achieved through automated optimization. Although the contextual banditframework arises as a natural candidate for such a task, its extension to a parallel setting is not straightforward: one needs to carefully adapt existing methods to fully leverage the multi-agent structure of this problem. We propose two approaches: one derived from a deterministic UCB-like method and the other relying on Thompson sampling. Thanks to its bayesian nature, the latter is intuited to better preserve the exploration-exploitation balance in the bandit batch. This is verified on toy experiments, where Thompson sampling shows robustness to the variability of the contexts. Finally, we apply both methods on a real base station network dataset and evidence that Thompson sampling outperforms both manual tuning and contextual UCB. I. INTRODUCTION The land area covered by a cellular wireless network, such as a mobile phone network, is divided into small areas called cells, each cell being covered by the antenna of a fixed base station (see Figure 1).

The availability of high-fidelity energy networks brings significant value to academic and commercial research. However, such releases also raise fundamental concerns related to privacy and security as they can reveal sensitive commercial information and expose system vulnerabilities. This paper investigates how to release power networks where the parameters of transmission lines and transformers are obfuscated. It does so by using the framework of Differential Privacy (DP), that provides strong privacy guarantees and has attracted significant attention in recent years. Unfortunately, simple DP mechanisms often result in AC-infeasible networks. To address these concerns, this paper presents a novel differential privacy mechanism that guarantees AC-feasibility and largely preserves the fidelity of the obfuscated network. Experimental results also show that the obfuscation significantly reduces the potential damage of an attacker exploiting the release of the dataset.

Nowadays two new technologies are roaring in the tech world – Artificial Intelligence and Blockchain. Both of these technologies have the potential to revolutionize the world. But the most discussed topic so far is that, whether these two can really be beneficial for each other. We already know that blockchain has the capability to offer a decentralized ledger system and many are already adopting the tech. On the other hand, Artificial Intelligence also started to streamlining processes for our benefit. But can blockchain based AI be the next technological milestone? Well, let's see if blockchain is really capable of powering AI or not. Artificial Intelligent is the simulation of human-like intelligence through computer systems. Usually, these computer systems are programmed in a way to mimic human-like actions. Apparently, the process is utterly complex as human activities are complicated to simulate. But the primary ability of artificial intelligence would be too rationalize like humans and take actions based on intellectual thinking.

He goes by the name "Marty." Tall, slow-moving and gray, he has big cartoonish eyes that disguise something unique about the newest employee at Giant Food Stores: Marty is deliberate and relentless, and -- unlike his fellow employees -- he has the ability to work a seemingly endless number of hours without pay. Though he doesn't say much, a small message is always plastered to his slender trunk: "This store is monitored by Marty for your safety," it reads. "Marty is an autonomous robot that uses image capturing technology to report spills, debris and other potential hazards to store employees to improve your shopping experience." After a pilot program that kicked off in several Pennsylvania stores this past fall, Giant Food Stores announced Monday that it will place Martys in each of the supermarket chain's 172 stores across Pennsylvania, Maryland, Virginia and West Virginia.

A new machine-learning model developed by a West Virginia University student has potential applications in the energy, environmental and health-care fields. The model, which can be used to predict adsorption energies -- i.e., adhesive capabilities in gold nanoparticles -- was developed by Gihan Panapitiya, a doctoral physics student from Sri Lanka. Gold nanoparticles have historically been used by artists to bring out vibrant colors via their interaction with light. Now they are increasingly used in high-technology applications such as electronic conductors and others. "Machine learning recently came into the spotlight, and we wanted to do something linking machine learning with gold nanoparticles as catalysts," Panapitiya said.