There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

The concept of an agent has become important in both Artificial Intelligence (AI) and mainstream computer science. Our aim in this paper is to point the reader at what we perceive to be the most important theoretical and practical issues associated with the design and construction of intelligent agents. For convenience, we divide these issues into three areas (though as the reader will see, the divisions are at times somewhat arbitrary). Agent theory is concerned with the question of what an agent is, and the use of mathematical formalisms for representing and reasoning about the properties of agents. Agent architectures can be thought of as software engineering models of agents;researchers in this area are primarily concerned with the problem of designing software or hardware systems that will satisfy the properties specified by agent theorists. Finally, agent languages are software systems for programming and experimenting with agents; these languages may embody principles proposed by theorists. The paper is not intended to serve as a tutorial introduction to all the issues mentioned; we hope instead simply to identify the most important issues, and point to work that elaborates on them. The article includes a short review of current and potential applications of agent technology.

/pdf/intelligent-agents-theory-and-practice-5dchj7p8pc.pdf

Intelligent Agents: Theory and Practice

A 2001 IBM manifesto observed that a looming software complexity crisis -caused by applications and environments that number into the tens of millions of lines of code - threatened to halt progress in computing. The manifesto noted the almost impossible difficulty of managing current and planned computing systems, which require integrating several heterogeneous environments into corporate-wide computing systems that extend into the Internet. Autonomic computing, perhaps the most attractive approach to solving this problem, creates systems that can manage themselves when given high-level objectives from administrators. Systems manage themselves according to an administrator's goals. New components integrate as effortlessly as a new cell establishes itself in the human body. These ideas are not science fiction, but elements of the grand challenge to create self-managing computing systems.

The vision of autonomic computing

Sequential auctions are an important mechanism for buying/selling multiple objects. Existing work has studied sequential auctions for objects that are exclusively either common value or private value. However, in many real-world cases an object has both features. Also, in such cases, the common value component (which is the same for all bidders) depends on how much each bidder values the object. Moreover, an individual bidder generally does not know the true common value, since it may not know how much the other bidders value it. On the other hand, a bidder's private value is independent of the others' private values. Given this, we study settings that have both common and private value elements by treating each bidder's information about the common value as uncertain. We first determine equilibrium bidding strategies for each auction in a sequence using English auction rules. On the basis of this equilibrium, we analyse the efficiency of auctions. Specifically, we show that the inefficiency that arises as a result of uncertainty about the common values can be reduced if the auctioneer makes its information about the common value known to all bidders. Moreover, our analysis also shows that the efficiency of auctions in an agent-based setting is higher than that in an all-human setting.

/pdf/sequential-auctions-for-objects-with-common-and-private-5d99kpf56g.pdf

Sequential auctions for objects with common and private values

In many multiagent systems, agents can improve their performance by forming coalitions, i.e., pooling their efforts and resources so as to achieve the tasks at hand in a more efficient way. This holds both for cooperative agents, i.e., agents who share a common set of goals, and for selfish agents who only care about their own payoffs. For cooperative agents, to find the optimal collaboration pattern, it suffices to identify the best way of splitting agents into teams. In contrast, when the agents are selfish, we also have to specify how to distribute the gains from cooperation, since each agent needs to be incentivized to participate in the proposed solution. In this chapter, we discuss coalition formation in multiagent systems for both selfish and cooperative agents. To deal with selfish agents, we introduce classic solution concepts of coalitional game theory that capture the notions of stability and fairness in coalition formation settings. We then give an overview of existing representation formalisms for coalitional games. For each such formalism, we discuss the complexity of computing the solution concepts defined earlier in the chapter, focusing on algorithms whose running time is polynomial in the number of agents n. In the second half of the chapter, we focus on practical approaches for finding an optimal partition of agents into teams. We present the state-of-the-art algorithms for this problem, and compare their relative strengths and weaknesses.

Computational Coalition Formation

Current electricity tariffs do not reflect the real costs that a customer incurs to a supplier, as units are charged at the same rate, regardless of the consumption pattern. In this paper, we propose a prediction-of-use tariff that better reflects these costs, which asks customers to predict a baseline consumption, and charges them based both on their actual consumption, and the deviation from their prediction. We show how under this tariff no customer would have an incentive to consume in excess of their actual needs, and derive closed form expressions for their optimal prediction and expected payments.Second, using principles from cooperative game theory, we study how customers can collectively reduce their potential deviation by aggregating under a group-buying scheme. We prove that the associated cost game is concave, which means grouping reduces the total expected bill and that this payment can be fairly allocated among customers by their Shapley values. Third, considering a model where customers can join the group online, we propose marginal payment allocation schemes that incentivise them to commit early, thus preventing start-up inertia. Finally, we validate our model using real data from a set of 3000 consumers from the UK.

https://eprints.soton.ac.uk/360819/1/document.pdf

Prediction-of-use games: a cooperative game theoryapproach to sustainable energy tariffs

Many crowdsourcing applications require spatial modelling of data to make sense of location-based observations provided by multiple users. In this context, We propose a new spatial function modelling approach to address the problem of fusing multiple spatial observations reported by possibly untrustworthy users in the domains of participatory sensing and crowdsourcing applications. Specifically, we use a heteroskedastic Gaussian process model to incorporate user trust modelling into Bayesian spatial regression. In particular, by training the model with the reports gathered from the crowd, we are able to estimate the spatial function at any location of interest and also learn the level of trustworthiness of each user. We show that our method outperforms other standard homoskedastic and heteroskedastic Gaussian processes by up to 23% on a crowdsourced radiation dataset collected during the 2011 Fukushima earthquake in Japan. We also show that our method is able to improve the quality of spatial predictions on synthetic data by up to 70% and is robust in settings of up to 30% presence of untrustworthy users within the crowd.

/pdf/crowdsourcing-spatial-phenomena-using-trust-based-22vj7l0hmb.pdf

Crowdsourcing Spatial Phenomena Using Trust-Based Heteroskedastic Gaussian Processes

Agent interaction in realistic applications is subject to many forms of uncertainty -- including information and network uncertainty, trust of and conflicts with other participants, lack of stability in a deal and risks about agreements and commitments. However, one of the most common forms of uncertainty occurs when a group has divergent beliefs about the interaction they are engaged in -- some agents believe an agreement has been reached, while others believe it has been rejected or that they are still bargaining. Such misunderstandings can arise because of loss of network performance, spurious connections, message loss or delays. Against this background, this paper develops synchronisation protocols for a group of agents to attain the same beliefs about an interaction, independent of the reliability of the underlying communication layer. This paper includes and proves theorems about a group's mutual beliefs, on which the safety of an interaction relies. Specifically, protocols for message exchange and belief revision and the reasoning for reachability of states during interactions are presented. Each protocol is proved to show that an increasing level of mutual and consistent belief is reached, thereby guaranteeing an interaction's integrity.

/pdf/ensuring-consistency-in-the-joint-beliefs-of-interacting-58f6wmpxwx.pdf

Nicholas R. Jennings

Papers

Sequential auctions for objects with common and private values

Computational Coalition Formation

Prediction-of-use games: a cooperative game theoryapproach to sustainable energy tariffs

Crowdsourcing Spatial Phenomena Using Trust-Based Heteroskedastic Gaussian Processes

Ensuring consistency in the joint beliefs of interacting agents