Transfer learning is the process of using knowledge gained while solving one problem to solve a new, previously unencountered problem. Current research has concentrated on analogical transfer - a mechanic is able to fix a type of car he has never seen before by comparing it to cars he has fixed before. This approach is typical of case-based reasoning systems and has been successful on a wide variety of problems [Watson, 1997]. When a new problem is encountered, a database of previously solved problems is searched for a problem with similar features. The solution to the most similar problem is selected, adapted and then applied to the new problem. Similar methods exist for adapting reinforcement learning policies [Taylor and Stone, 2009]. We refer to the above approaches as solution adaptation algorithms - a pair of problems are matched on similarity and the solution to the first problem, after some adaptation, is applied to the second problem. The solution adaptation approach requires three things.

This paper examines an extended double auction model where market clearing is restricted by temporal constraints. It is found that the allocation problem in this model can be effectively transformed into a weighted bipartite matching in graph theory. By using the augmentation technique, we propose a Vickrey-Clarke-Groves (VCG) mechanism in this model and demonstrate the advantages of the payment compared with the classical VCG payment (the Clarke pivot payment). We also show that the algorithms for both allocation and payment calculation run in polynomial time. It is expected that the method and results provided in this paper can be applied to the design and analysis of dynamic double auctions and futures markets.

In this work we propose an approach for generalization in continuous domain Reinforcement Learning that, instead of using a single function approximator, tries many different function approximators in parallel, each one defined in a different region of the domain. Associated with each approximator is a relevance function that locally quantifies the quality of its approximation, so that, at each input point, the approximator with highest relevance can be selected. The relevance function is defined using parametric estimations of the variance of the q-values and the density of samples in the input space, which are used to quantify the accuracy and the confidence in the approximation, respectively. These parametric estimations are obtained from a probability density distribution represented as a Gaussian Mixture Model embedded in the input-output space of each approximator. In our experiments, the proposed approach required a lesser number of experiences for learning and produced more stable convergence profiles than when using a single function approximator.

Query containment has been studied extensively in KR and databases, for different kinds of query languages and domain constraints. We address the longstanding open problem of containment under expressive description logic (DL) constraints for two-way regular path queries (2RPQs) and their conjunctions, which generalize conjunctive queries with the ability to express regular navigation. We show that, surprisingly, functionality constraints alone make containment of 2RPQs already ExpTime-hard. By employing automata-theoretic techniques, we also provide a matching upper bound that extends to very expressive DL constraints. For conjunctive 2RPQs we prove a further exponential jump in complexity, and provide again a matching upper bound for expressive DLs. Our techniques provide also a solution to the problem of query entailment over DL knowledge bases in which individuals in the ABox may be related through regular role-paths.

As tweets has become a comprehensive repository of fresh information, Semantic Role Labeling (SRL) for tweets has aroused great research interests because of its center role in a wide range of tweet related studies such as fine-grained information extraction, sentiment analysis and summarization. However, the fact that a tweet is often too short and informal to provide sufficient information poses a main challenge. To tackle this challenge, we propose a new method to collectively label similar tweets. The underlying idea is to exploit similar tweets to make up for the lack of information in a tweet. Specifically, similar tweets are first grouped together by clustering. Then for each cluster a two-stage labeling is conducted: One labeler conducts SRL to get statistical information, such as the predicate/argument/role triples that occur frequently, from its highly confidently labeled results; then in the second stage, another labeler performs SRL with such statistical information to refine the results. Experimental results on a human annotated dataset show that our approach remarkably improves SRL by 3.1% F1.

Field classification is an extension of the traditional classification framework, by breaking the i.i.d. assumption. In field classification, patterns occur as groups (fields) of homogeneous styles. By utilizing style consistency, classifying groups of patterns is often more accurate than classifying single patterns. In this paper, we extend the Bayes decision theory, and develop the Field Bayesian Model (FBM) to deal with field classification. Specifically, we propose to learn a Style Normalized Transformation (SNT) for each field. Via the SNTs, the data of different fields are transformed to a uniform style space (i.i.d. space). The proposed model is a general and systematic framework, under which many probabilistic models can be easily extended for field classification. To transfer the model to unseen styles, we propose a transductive model called Transfer Bayesian Rule (TBR) based on self-training. We conducted extensive experiments on face, speech and a large-scale handwriting dataset, and got significant error rate reduction compared to the state-of-the-art methods.

The problem of feature selection has aroused considerable research interests in the past few years. Traditional learning based feature selection methods separate embedding learning and feature ranking. In this paper, we introduce a novel unsupervised feature selection approach via Joint Embedding Learning and Sparse Regression (JELSR). Instead of simply employing the graph laplacian for embedding learning and then regression, we use the weight via locally linear approximation to construct graph and unify embedding learning and sparse regression to perform feature selection. By adding the ell {2,1} -norm regularization, we can learn a sparse matrix for feature ranking. We also provide an effective method to solve the proposed problem. Compared with traditional unsupervised feature selection methods, our approach could integrate the merits of embedding learning and sparse regression simultaneously. Plenty of experimental results are provided to show the validity.

Many unsupervised learning methods for recognising patterns in data streams are based on fixed length data sequences, which makes them unsuitable for applications where the data sequences are of variable length such as in speech recognition, behaviour recognition and text classification. In order to use these methods on variable length data sequences, a pre-processing step is required to manually segment the data and select the appropriate features, which is often not practical in real-world applications. In this paper we suggest an unsupervised learning method that handles variable length data sequences by identifying structure in the data stream using text compression and the edit distance between ‘words’. We demonstrate that using this method we can automatically cluster unlabelled data in a data stream and perform segmentation. We evaluate the effectiveness of our proposed method using both fixed length and variable length benchmark datasets, comparing it to the Self-Organising Map in the first case. The results show a promising improvement over baseline recognition systems.

In agent-oriented programming and planning, agents' actions are typically specified in terms of postconditions, and the model of execution assumes that the environment carries the actions out exactly as specified. That is, it is assumed that the state of the environment after an action has been executed will satisfy its postcondition. In reality, however, such environments are rare: the actual execution of an action may fail, and the envisaged outcome is not met. We provide a conceptual framework for reasoning about success and failure of agents' behaviours. In particular, we propose a measure that reflects how "good" an environment is with respect to agent's capabilities and a given goal it might pursue. We also discuss which types of goals are worth pursuing, depending on the type of environment the agent is acting in.

In contrast to the studied in single-period optimization [Miller and Wagner, standard approach that studies the expected performance, 1965; Prékopa, 1970]. However, little has been done in we consider the policy that maximizes the context of sequential decision problem including MDPs. the probability of achieving a predetermined target The standard approaches in risk-averse MDPs include maximization performance, a criterion we term probabilistic of expected utility function [Bertsekas, 1995], goal Markov decision processes. We show that and optimization of a coherent risk measure [Riedel, 2004; this problem is NPhard, but can be solved using a Le Tallec, 2007]. Both approaches lead to formulations that pseudo-polynomial algorithm. We further consider can not be solved in polynomial time, except for special a variant dubbed "chance-constraint Markov decision cases including exponential utility function [Chung and Sobel, problems," that treats the probability of achieving 1987], piecewise linear utility function with a single target performance as a constraint instead of the break down point [Liu and Koenig, 2005], and risk measures maximizing objective. This variant is NPhard, but that can be reduced to robust MDPs satisfying the socalled can be solved in pseudo-polynomial time.