Design and development of automatic visual inspection system for PCB manufacturing

Inspection of solder joints has been a critical process in the electronic manufacturing industry to reduce manufacturing cost, improve yield, and ensure product quality and reliability. This paper proposes two inspection modules for an automatic solder joint classification system. The “front-end” inspection system includes illumination normalisation, localisation and segmentation. The “back-end” inspection involves the classification of solder joints using the Log Gabor filter and classifier fusion. Five different levels of solder quality with respect to the amount of solder paste have been defined. The Log Gabor filter has been demonstrated to achieve high recognition rates and is resistant to misalignment. This proposed system does not need any special illumination system, and the images are acquired by an ordinary digital camera. This system could contribute to the development of automated non-contact, non-destructive and low cost solder joint quality inspection systems.

Performance analysis of business process models with advanced constructs

The importance of actively managing and analysing business processes is acknowledged more than ever in organisations nowadays. Business processes form an essential part of an organisation and their application areas are manifold. Most organisations keep records of various activities that have been carried out for auditing purposes, but they are rarely used for analysis purposes. This paper describes the design and implementation of a process analysis tool that replays, analyses and visualises a variety of performance metrics using a process definition and its corresponding execution logs. The replayer uses a YAWL process model example to demonstrate its capacity to support advanced language constructs.

The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network

Sample complexity results from computational learning theory, when applied to neural network learning for pattern classification problems, suggest that for good generalization performance the number of training examples should grow at least linearly with the number of adjustable parameters in the network. Results in this paper show that if a large neural network is used for a pattern classification problem and the learning algorithm finds a network with small weights that has small squared error on the training patterns, then the generalization performance depends on the size of the weights rather than the number of weights. For example, consider a two-layer feedforward network of sigmoid units, in which the sum of the magnitudes of the weights associated with each unit is bounded by A and the input dimension is n. We show that the misclassification probability is no more than a certain error estimate (that is related to squared error on the training set) plus A3 √((log n)/m) (ignoring log A and log m factors), where m is the number of training patterns. This may explain the generalization performance of neural networks, particularly when the number of training examples is considerably smaller than the number of weights. It also supports heuristics (such as weight decay and early stopping) that attempt to keep the weights small during training. The proof techniques appear to be useful for the analysis of other pattern classifiers: when the input domain is a totally bounded metric space, we use the same approach to give upper bounds on misclassification probability for classifiers with decision boundaries that are far from the training examples.

Exponentiated gradient algorithms for conditional random fields and max-margin Markov Networks

Log-linear and maximum-margin models are two commonly-used methods in supervised machine learning, and are frequently used in structured prediction problems. Efficient learning of parameters in these models is therefore an important problem, and becomes a key factor when learning from very large data sets. This paper describes exponentiated gradient (EG) algorithms for training such models, where EG updates are applied to the convex dual of either the log-linear or max-margin objective function; the dual in both the log-linear and max-margin cases corresponds to minimizing a convex function with simplex constraints. We study both batch and online variants of the algorithm, and provide rates of convergence for both cases. In the max-margin case, O(1/ε) EG updates are required to reach a given accuracy ε in the dual; in contrast, for log-linear models only O(log(1/ε)) updates are required. For both the max-margin and log-linear cases, our bounds suggest that the online EG algorithm requires a factor of n less computation to reach a desired accuracy than the batch EG algorithm, where n is the number of training examples. Our experiments confirm that the online algorithms are much faster than the batch algorithms in practice. We describe how the EG updates factor in a convenient way for structured prediction problems, allowing the algorithms to be efficiently applied to problems such as sequence learning or natural language parsing. We perform extensive evaluation of the algorithms, comparing them to L-BFGS and stochastic gradient descent for log-linear models, and to SVM-Struct for max-margin models. The algorithms are applied to a multi-class problem as well as to a more complex large-scale parsing task. In all these settings, the EG algorithms presented here outperform the other methods.

Shifting : one-inclusion mistake bounds and sample compression

We present new expected risk bounds for binary and multiclass prediction, and resolve several recent conjectures on sample compressibility due to Kuzmin and Warmuth. By exploiting the combinatorial structure of concept class F, Haussler et al. achieved a VC(F)/n bound for the natural one-inclusion prediction strategy. The key step in their proof is a d=VC(F) bound on the graph density of a subgraph of the hypercube—one-inclusion graph. The first main result of this report is a density bound of n∙choose(n-1,≤d-1)/choose(n,≤d) < d, which positively resolves a conjecture of Kuzmin and Warmuth relating to their unlabeled Peeling compression scheme and also leads to an improved one-inclusion mistake bound. The proof uses a new form of VC-invariant shifting and a group-theoretic symmetrization. Our second main result is an algebraic topological property of maximum classes of VC-dimension d as being d-contractible simplicial complexes, extending the well-known characterization that d=1 maximum classes are trees. We negatively resolve a minimum degree conjecture of Kuzmin and Warmuth—the second part to a conjectured proof of correctness for Peeling—that every class has one-inclusion minimum degree at most its VC-dimension. Our final main result is a k-class analogue of the d/n mistake bound, replacing the VC-dimension by the Pollard pseudo-dimension and the one-inclusion strategy by its natural hypergraph generalization. This result improves on known PAC-based expected risk bounds by a factor of O(log n) and is shown to be optimal up to a O(log k) factor. The combinatorial technique of shifting takes a central role in understanding the one-inclusion (hyper)graph and is a running theme throughout

Adaptive online gradient descent

We study the rates of growth of the regret in online convex optimization. First, we show that a simple extension of the algorithm of Hazan et al eliminates the need for a priori knowledge of the lower bound on the second derivatives of the observed functions. We then provide an algorithm, Adaptive Online Gradient Descent, which interpolates between the results of Zinkevich for linear functions and of Hazan et al for strongly convex functions, achieving intermediate rates between [square root T] and [log T]. Furthermore, we show strong optimality of the algorithm. Finally, we provide an extension of our results to general norms.

Optimistic linear programming gives logarithmic regret for irreducible MDPs

We present an algorithm called Optimistic Linear Programming (OLP) for learning to optimize average reward in an irreducible but otherwise unknown Markov decision process (MDP). OLP uses its experience so far to estimate the MDP. It chooses actions by optimistically maximizing estimated future rewards over a set of next-state transition probabilities that are close to the estimates, a computation that corresponds to solving linear programs. We show that the total expected reward obtained by OLP up to time T is within C(P) log T of the reward obtained by the optimal policy, where C(P) is an explicit, MDP-dependent constant. OLP is closely related to an algorithm proposed by Burnetas and Katehakis with four key differences: OLP is simpler, it does not require knowledge of the supports of transition probabilities, the proof of the regret bound is simpler, but our regret bound is a constant factor larger than the regret of their algorithm. OLP is also similar in flavor to an algorithm recently proposed by Auer and Ortner. But OLP is simpler and its regret bound has a better dependence on the size of the MDP.

The role of information avoidance in everyday-life information behaviors

This paper presents the results from a study of information behaviors in the context of people's everyday lives as part of a larger study of information behaviors (IB). 34 participants from across 6 countries maintained a daily information journal or diary – mainly through a secure web log – for two weeks, to an aggregate of 468 participant days over five months. The text-rich diary data was analyzed using Grounded Theory analysis. The findings indicate that information avoidance is a common phenomenon in everyday life and consisted of both passive avoidance and active avoidance. This has implications for several aspects of peoples' lives including health, finance, and personal relationships.

Esprit Futur : Le Corbusier and José Oubrerie at Saint-Pierre De Firminy, Firminy-Vert, France, 1968–2006

What is the secret mesmerism that death possesses and under the operation of which a modern architect – strident, confident, resolute – becomes rueful, pessimistic, or melancholic?1 Five years before Le Corbusier’s death at sea in 1965, the architect reluctantly agreed to adopt the project for L’Église Saint-Pierre de Firminy in Firminy-Vert (1960–2006), following the death of its original architect, André Sive, from leukemia in 1958.2 Le Corbusier had already developed, in 1956, the plan for an enclave in the new “green” Firminy town, which included his youth and culture center and a stadium and swimming pool; the church and a “boîte à miracles” near the youth center were inserted into the plan in the ’60s. (Le Corbusier was also invited, in 1962, to produce another plan for three Unités d’Habitation outside Firminy-Vert.) The Saint-Pierre church should have been the zenith of the quartet (the largest urban concentration of works by Le Corbusier in Europe, and what the architect Henri Ciriani termed Le Corbusier’s “acropolis”3) but in the early course of the project, Le Corbusier would suffer the diocese’s serial objections to his vision for the church – not unlike the difficulties he experienced with Notre Dame du Haut at Ronchamp (1950–1954) and the resistance to his proposed monastery of Sainte-Marie de la Tourette (1957–1960). In 1964, the bishop of Saint-Étienne requested that Le Corbusier relocate the church to a new site, but Le Corbusier refused and the diocese subsequently withdrew from the project. (With neither the approval, funds, nor the participation of the bishop, by then the cardinal archbishop of Lyon, the first stone of the church was finally laid on the site in 1970.) Le Corbusier’s ambivalence toward the project, even prior to his quarrels with the bishop, reveals...

Fixing the city one photo at a time : mobile logging of maintenance requests

We have designed a mobile application that takes advantage of the built-in features of smart phones such as camera and GPS that allow users to take geo-tagged photos while on the move. Urban residents can take pictures of broken street furniture and public property requiring repair, attach a brief description, and submit the information as a maintenance request to the local government organisation of their city. This paper discusses the design approach that led to the application, highlights a built-in mechanism to elicit user feedback, and evaluates the progress to date with user feedback and log statistics. It concludes with an outlook highlighting user requested features and our own design aspirations for moving from a reporting tool to a civic engagement tool.

Private supervisors of learner drivers and Queensland's graduated driver licensing system : how important are they?

Background: A key element of graduated driver licensing systems is the level of support provided by parents. In mid-2007 changes were made to Queensland’s graduated driver licensing system, including amendments to the learner licence with one of the more significant changes requiring learners to record 100 hours of supervised driving practice in a logbook. Prior to mid-2007, there was no minimum supervision requirement. Aims: The aim of this study was to document the experiences of the supervisors of Queensland learner drivers after the changes made to the graduated driver licensing system in mid-2007. Methods: The sample of 552 supervisors of learner drivers was recruited using a combination of convenience and snowball sampling techniques. The internet survey was open for completion between July 2009 and May 2010 and took approximately 15 to 20 minutes for participants to complete. Results: For 59.7 per cent of the participants, this was the first time that they had supervised a learner driver. For 63.2 per cent, they classified themselves as the main supervisor for the learner driver. Participants provided an average of 79.62 hours of supervision (sd = 92.38), while other private supervisors provided 34.89 hours of supervision (sd = 41.74) to the same learner and professional driving instructors 18.55 hours of supervision (sd = 27.54). The vast majority of supervisors recorded all or most of the practice that they provided their learner driver in their log book with most supervisors indicating that they believed that the hours recorded in the learner’s logbook were either accurate or very accurate. While many supervisors stated that they did not receive any advice regarding the supervision of learner drivers, some had received advice from others such as friends or through discussions with a professional driving instructor. Discussion and conclusions: While graduated driver licensing systems implicitly encourage the involvement of parents and other private supervisors, these people tend not to be systematically involved. As demonstrated in this study, private supervisors provide a significant amount of supervised practice and seek to record this practice accurately and honestly in the learner’s logbook. However, even though a significant number of participants reported that this was the first time that they had supervised a learner driver, they accessed little support or guidance for their role. This suggests a need to more overtly encourage and support the role of private supervisors for learner drivers.

Effective hybrid recommendation combining users-searches correlations using tensors

Most recommendation methods employ item-item similarity measures or use ratings data to generate recommendations. These methods use traditional two dimensional models to find inter relationships between alike users and products. This paper proposes a novel recommendation method using the multi-dimensional model, tensor, to group similar users based on common search behaviour, and then finding associations within such groups for making effective inter group recommendations. Web log data is multi-dimensional data. Unlike vector based methods, tensors have the ability to highly correlate and find latent relationships between such similar instances, consisting of users and searches. Non redundant rules from such associations of user-searches are then used for making recommendations to the users.

Identifying interests of web users for effective recommendations

Search log data is multi dimensional data consisting of number of searches of multiple users with many searched parameters. This data can be used to identify a user’s interest in an item or object being searched. Identifying highest interests of a Web user from his search log data is a complex process. Based on a user’s previous searches, most recommendation methods employ two-dimensional models to find relevant items. Such items are then recommended to a user. Two-dimensional data models, when used to mine knowledge from such multi dimensional data may not be able to give good mappings of user and his searches. The major problem with such models is that they are unable to find the latent relationships that exist between different searched dimensions. In this research work, we utilize tensors to model the various searches made by a user. Such high dimensional data model is then used to extract the relationship between various dimensions, and find the prominent searched components. To achieve this, we have used popular tensor decomposition methods like PARAFAC, Tucker and HOSVD. All experiments and evaluation is done on real datasets, which clearly show the effectiveness of tensor models in finding prominent searched components in comparison to other widely used two-dimensional data models. Such top rated searched components are then given as recommendation to users.

Clustering of web users using the tensor decomposed models

We propose to use the Tensor Space Modeling (TSM) to represent and analyze the user’s web log data that consists of multiple interests and spans across multiple dimensions. Further we propose to use the decomposition factors of the Tensors for clustering the users based on similarity of search behaviour. Preliminary results show that the proposed method outperforms the traditional Vector Space Model (VSM) based clustering.