Exploration in sparse-reward reinforcement learning (RL) is difficult due to the need for long, coordinated sequences of actions in order to achieve any reward. Skill learning, from demonstrations or interaction, is a promising approach to address this, but skill extraction and inference are expensive for current methods. We present a novel method to extract skills from demonstrations for use in sparse-reward RL, inspired by the popular Byte-Pair Encoding (BPE) algorithm in natural language processing. With these skills, we show strong performance in a variety of tasks, 1000 \times acceleration for skill-extraction and 100 \times acceleration for policy inference. Given the simplicity of our method, skills extracted from 1\% of the demonstrations in one task can be transferred to a new loosely related task.

Maintenance can be a complex undertaking, requiring thorough planning and an astute understanding of a facility's risk profile. This is particularly true of high-risk facilities. Maintenance does not occur in a'vacuum' and can result in costly downtimes if it is unexpected or unplanned. Even planned maintenance can lead to lengthy downtimes resulting in huge losses. For example, oil refineries in the US alone lose an estimated $47 billion from 213,000 hours of downtime each year.

Looking back, things probably started to unravel when my wife, Katie, found me tucked up in bed, talking to another woman. 'Goodnight,' she heard me say. 'Sweet dreams,' came the woman's obliging response. The next morning, Katie walked past the sitting room door to catch me muttering something about the weather. 'I wasn't talking to you,' I replied.

Only one in 20 companies has extensively incorporated AI in offerings or processes. Less than 39% of all companies have an AI strategy in place. According to MIT Sloan Review, the largest companies -- those with at least 100,000 employees -- are the most likely to have an AI strategy, but only half have one. Despite claims that AI is already being subsumed into an array of applications, we're not there yet and won't be in 2018. It is still the early days of adoption, and those companies that are implementing AI now will see the biggest competitive value.

Artificial intelligence (AI) is changing the face of business. No longer a futuristic concept, its impact is real. From tech giants like Google, Apple and Amazon to user-centric behemoths like Uber and Starbucks, everyone seems to be using AI technology to transform the customer experience (CX). But, it's not just corporate giants that are deploying AI. Smaller organizations are following suit.

In general, modeling is a complex and creative task, and building qualitative models is no exception. One way of automating this task is by means of machine learning. Observed behaviors of a modeled system are used as examples for a learning algorithm that constructs a model that is consistent with the data. In this article, we review approaches to learning qualitative models, either from numeric data or qualitative observations. However, an important practical question is how do we construct qualitative models in the first place.

Individual agent skills, such as kicking and dribbling (running with the ball), are important prerequisites for team collaboration. For each of these skills, many parameters affect the details of the skill execution. For example, in the ball skill of dribbling, there are parameters that affect how quickly the agent runs, how far ahead it kicks the ball, and on which side of its body the agent keeps the ball while it dribbles. The settings for these parameters usually involve a tradeoff, such as speed versus safety or power versus accuracy. It is important to gain an understanding of what exactly these tradeoffs are before "correct" parameter settings can be made. We created a trainer client that connects to the server as an omniscient offline coach client.

CHIP's highest-level goals were programmed C and runs on board. The RAP system is designed to deal with achieving goals in a dynamic environment. Each RAP task description encodes a set of methods for carrying out the task in different situations, a success check to tell when the task has accomplished its purpose, and notations that describe when things are not going as expected. At run time, a RAP task examines its methods and selects one that is appropriate in the current situation. By doing method selection at run time, RAPs are more likely to select the best method, even if the world is changing or contains details that cannot be predicted in advance.

The RoboCupJunior division of RoboCup is now entering its third year of international participation and is growing rapidly in size and popularity. This article first outlines the history of the junior league since it was first demonstrated in Paris at RoboCup-1998 and describes how it has evolved into the international sensation it is today. Although the popularity of the event is self-evident, we are working to identify and quantify the educational benefits of the initiative. The remainder of the article focuses on describing our efforts to encapsulate these qualities, highlighting results from a pilot study conducted at RoboCupJunior-2000 and presenting new data from a subsequent study of RoboCupJunior-2001. In 1998, Lund and Pagliarini demonstrated the idea of a children's league for RoboCup, using robots constructed and programmed with the Lego Mindstorms kit that could play soccer (Land and Pagliarini 1998).