It is noted that some unusual moves against a strong chess program greatly weaken its ability to see the serious targets of the game, and its whole level of play... It is suggested to create programs with different weaknesses in order to analyze similar human behavior. Finally, a new version of chess, "Chess Corrida" is suggested.

Existing approaches to analyzing the asymptotics of graph Laplacians typically assume a well-behaved kernel function with smoothness assumptions. We remove the smoothness assumption and generalize the analysis of graph Laplacians to include previously unstudied graphs including kNN graphs. We also introduce a kernel-free framework to analyze graph constructions with shrinking neighborhoods in general and apply it to analyze locally linear embedding (LLE). We also describe how for a given limiting Laplacian operator desirable properties such as a convergent spectrum and sparseness can be achieved choosing the appropriate graph construction.

A typical approach in estimating the learning rate of a regularized learning scheme is to bound the approximation error by the sum of the sampling error, the hypothesis error and the regularization error. Using a reproducing kernel space that satisfies the linear representer theorem brings the advantage of discarding the hypothesis error from the sum automatically. Following this direction, we illustrate how reproducing kernel Banach spaces with the l1 norm can be applied to improve the learning rate estimate of l1-regularization in machine learning.

Hybrid learning methods use theoretical knowledge of a domain and a set of classified examples to develop a method for classification. Methods that use domain knowledge have been shown to perform better than inductive learners. However, there is no general method to include domain knowledge into all inductive learning algorithms as all hybrid methods are highly specialized for a particular algorithm. We present an algorithm that will take domain knowledge in the form of propositional rules, generate artificial examples from the rules and also remove instances likely to be flawed. This enriched dataset then can be used by any learning algorithm. Experimental results of different scenarios are shown that demonstrate this method to be more effective than simple inductive learning.

Different features have different relevance to a particular learning problem. Some features are less relevant; while some very important. Instead of selecting the most relevant features using feature selection, an algorithm can be given this knowledge of feature importance based on expert opinion or prior learning. Learning can be faster and more accurate if learners take feature importance into account. Correlation aided Neural Networks (CANN) is presented which is such an algorithm. CANN treats feature importance as the correlation coefficient between the target attribute and the features. CANN modifies normal feed-forward Neural Network to fit both correlation values and training data. Empirical evaluation shows that CANN is faster and more accurate than applying the two step approach of feature selection and then using normal learning algorithms.

External information propagates in the cell mainly through signaling cascades and transcriptional activation, allowing it to react to a wide spectrum of environmental changes. High throughput experiments identify numerous molecular components of such cascades that may, however, interact through unknown partners. Some of them may be detected using data coming from the integration of a protein-protein interaction network and mRNA expression profiles. This inference problem can be mapped onto the problem of finding appropriate optimal connected subgraphs of a network defined by these datasets. The optimization procedure turns out to be computationally intractable in general. Here we present a new distributed algorithm for this task, inspired from statistical physics, and apply this scheme to alpha factor and drug perturbations data in yeast. We identify the role of the COS8 protein, a member of a gene family of previously unknown function, and validate the results by genetic experiments. The algorithm we present is specially suited for very large datasets, can run in parallel, and can be adapted to other problems in systems biology. On renowned benchmarks it outperforms other algorithms in the field.

We consider the problem of analyzing the heterogeneity of clustering distributions for multiple groups of observed data, each of which is indexed by a covariate value, and inferring global clusters arising from observations aggregated over the covariate domain. We propose a novel Bayesian nonparametric method reposing on the formalism of spatial modeling and a nested hierarchy of Dirichlet processes. We provide an analysis of the model properties, relating and contrasting the notions of local and global clusters. We also provide an efficient inference algorithm, and demonstrate the utility of our method in several data examples, including the problem of object tracking and a global clustering analysis of functional data where the functional identity information is not available.

Engineered systems are designed to deftly operate under predetermined conditions yet are notoriously fragile when unexpected perturbations arise. In contrast, biological systems operate in a highly flexible manner; learn quickly adequate responses to novel conditions, and evolve new routines/traits to remain competitive under persistent environmental change. A recent theory on the origins of biological flexibility has proposed that degeneracy - the existence of multi-functional components with partially overlapping functions - is a primary determinant of the robustness and adaptability found in evolved systems. While degeneracy's contribution to biological flexibility is well documented, there has been little investigation of degeneracy design principles for achieving flexibility in systems engineering. Actually, the conditions that can lead to degeneracy are routinely eliminated in engineering design. With the planning of transportation vehicle fleets taken as a case study, this paper reports evidence that degeneracy improves robustness and adaptability of a simulated fleet without incurring costs to efficiency. We find degeneracy dramatically increases robustness of a fleet to unpredicted changes in the environment while it also facilitates robustness to anticipated variations. When we allow a fleet's architecture to be adapted in response to environmental change, we find degeneracy can be selectively acquired, leading to faster rates of design adaptation and ultimately to better designs. Given the range of conditions where favorable short-term and long-term performance outcomes are observed, we propose that degeneracy design principles fundamentally alter the propensity for adaptation and may be useful within several engineering and planning contexts.

Although researchers often comment on the rising popularity of nature-inspired meta-heuristics (NIM), there has been a paucity of data to directly support the claim that NIM are growing in prominence compared to other optimization techniques. This study presents evidence that the use of NIM is not only growing, but indeed appears to have surpassed mathematical optimization techniques (MOT) in several important metrics related to academic research activity (publication frequency) and commercial activity (patenting frequency). Motivated by these findings, this article discusses some of the possible origins of this growing popularity. I review different explanations for NIM popularity and discuss why some of these arguments remain unsatisfying. I argue that a compelling and comprehensive explanation should directly account for the manner in which most NIM success has actually been achieved, e.g. through hybridization and customization to different problem environments. By taking a problem lifecycle perspective, this paper offers a fresh look at the hypothesis that nature-inspired meta-heuristics derive much of their utility from being flexible. I discuss global trends within the business environments where optimization algorithms are applied and I speculate that highly flexible algorithm frameworks could become increasingly popular within our diverse and rapidly changing world.

The context of a software developer is something hard to define and capture, as it represents a complex network of elements across different dimensions that are not limited to the work developed on an IDE. We propose the definition of a software developer context model that takes into account all the dimensions that characterize the work environment of the developer. We are especially focused on what the software developer context encompasses at the project level and how it can be captured. The experimental work done so far show that useful context information can be extracted from project management tools. The extraction, analysis and availability of this context information can be used to enrich the work environment of the developer with additional knowledge to support her/his work.