We show that for a general class of convex online learning problems, Mirror Descent can always achieve a (nearly) optimal regret guarantee.

Multi-task learning (MTL) learns multiple related tasks simultaneously to improve generalization performance. Alternating structure optimization (ASO) is a popular MTL method that learns a shared low-dimensional predictive structure on hypothesis spaces from multiple related tasks. It has been applied successfully in many real world applications. As an alternative MTL approach, clustered multi-task learning (CMTL) assumes that multiple tasks follow a clustered structure, i.e., tasks are partitioned into a set of groups where tasks in the same group are similar to each other, and that such a clustered structure is unknown a priori. The objectives in ASO and CMTL differ in how multiple tasks are related. Interestingly, we show in this paper the equivalence relationship between ASO and CMTL, providing significant new insights into ASO and CMTL as well as their inherent relationship. The CMTL formulation is non-convex, and we adopt a convex relaxation to the CMTL formulation. We further establish the equivalence relationship between the proposed convex relaxation of CMTL and an existing convex relaxation of ASO, and show that the proposed convex CMTL formulation is significantly more efficient especially for high-dimensional data. In addition, we present three algorithms for solving the convex CMTL formulation. We report experimental results on benchmark datasets to demonstrate the efficiency of the proposed algorithms.

Diversified retrieval and online learning are two core research areas in the design of modern information retrieval systems. In this paper, we propose the linear submodular banditsproblem, which is an online learning setting for optimizing a general class of feature-rich submodular utility models for diversified retrieval. We present an algorithm, called LSBGREEDY, and prove that it efficiently converges to a near-optimal model. As a case study, we applied our approach to the setting of personalized news recommendation, where the system must recommend small sets of news articles selected from tens of thousands of available articles each day. In a live user study, we found that LSBGREEDY significantly outperforms existing online learning approaches.

Unlike existing nonparametric Bayesian models, which rely solely on specially conceived priors to incorporate domain knowledge for discovering improved latent representations, we study nonparametric Bayesian inference with regularization on the desired posterior distributions. While priors can indirectly affect posterior distributions through Bayes' theorem, imposing posterior regularization is arguably more direct and in some cases can be much easier. We particularly focus on developing infinite latent support vector machines (iLSVM) and multi-task infinite latent support vector machines (MT-iLSVM), which explore the large-margin idea in combination with a nonparametric Bayesian model for discovering predictive latent features for classification and multi-task learning, respectively. We present efficient inference methods and report empirical studies on several benchmark datasets. Our results appear to demonstrate the merits inherited from both large-margin learning and Bayesian nonparametrics.

A majority of approximate dynamic programming approaches to the reinforcement learning problem can be categorized into greedy value function methods and value-based policy gradient methods. The former approach, although fast, is well known to be susceptible to the policy oscillation phenomenon. We take a fresh view to this phenomenon by casting a considerable subset of the former approach as a limiting special case of the latter. We explain the phenomenon in terms of this view and illustrate the underlying mechanism with artificial examples. We also use it to derive the constrained natural actor-critic algorithm that can interpolate between the aforementioned approaches. In addition, it has been suggested in the literature that the oscillation phenomenon might be subtly connected to the grossly suboptimal performance in the Tetris benchmark problem of all attempted approximate dynamic programming methods. We report empirical evidence against such a connection and in favor of an alternative explanation. Finally, we report scores in the Tetris problem that improve on existing dynamic programming based results.

The word intrinsic motivation was first used in I. APPROACHESFOR ADAPTIVEPERSONALROBOTS psychology to describe the capability of humans to be attracted toward different activities for the pleasure that they experience The promise of personal robots operating in human environments intrinsically. These mechanisms have been shown crucial for to interact with people on a daily basis points out the humans to autonomously learn and discover new capabilities importance of adaptivity of the machine to its environment and [14]-[16]. This inspired the creation of fully autonomous users. The robot can no longer simply be all-programmed in robots [17]-[22] with meta-exploration mechanisms monitoring advance by engineers, and reproduce only actions predesigned the evolution of learning performances of the robot, in in factories. It needs to match its behaviour and learn new order to maximise informational gain, and with heuristics skills as the environment and users' needs change.

When faced with the problem of learning a model of a high-dimensional environment, a common approach is to limit the model to make only a restricted set of predictions, thereby simplifying the learning problem. These partial models may be directly useful for making decisions or may be combined together to form a more complete, structured model. However, in partially observable (non-Markov) environments, standard model-learning methods learn generative models, i.e. models that provide a probability distribution over all possible futures (such as POMDPs). It is not straightforward to restrict such models to make only certain predictions, and doing so does not always simplify the learning problem. In this paper we present prediction profile models: non-generative partial models for partially observable systems that make only a given set of predictions, and are therefore far simpler than generative models in some cases. We formalize the problem of learning a prediction profile model as a transformation of the original model-learning problem, and show empirically that one can learn prediction profile models that make a small set of important predictions even in systems that are too complex for standard generative models.

The current study investigates the relationship between children’s curiosity and question asking ability. Generation of two types of questions was assessed: identification (yes/no questions asked to identify a target from an array) and understanding questions, asked to learn more about a topic. The latter was related to children’s curiosity, as was the ability to recognize the effectiveness of questions in solving a mystery. Training on asking identification questions was effective in improving children’s ability to ask that type of question, but did not transfer to the other task. Training on asking understanding questions was not successful. Children’s curiosity did not influence the effectiveness of the training.

This paper describes a prototype system that generates mathematical word problems from ontologies in unrestricted domains. It builds on an existing ontology verbaliser that renders logical statements written in Web Ontology Language (OWL) as English sentences. This kind of question is more complex than those normally attempted by question generation systems, since mathematical word problems consist of a number of sentences that communicate a short narrative (in addition to providing the relevant numerical information required to solve the underlying mathematical problem). Thus, they embody many research issues that do not crop up with single-sentence questions. As well as describing the prototype system, I discuss five ways in which the difficulty of the generated questions may be controlled automatically during generation.

People acquire new knowledge in various ways and this helps them to adapt to changing environment properly. In this paper, we investigatethe interoperation of multiple learning mechanisms within a single system. We extend a cognitive architecture, ICARUS, to have three different modes of learning. Through experiments in a modified Blocks World and a route generation domain, we test and demonstrate the system's ability to get synergistic effects from these learning mechanisms.