Evaluating retraining rules for semi-supervised learning in neural network based cursive word recognition

Training a system to recognize handwritten words is a task that requires a large amount of data with their correct transcription. However, the creation of such a training set, including the generation of the ground truth, is tedious and costly. One way of reducing the high cost of labeled training data acquisition is to exploit unlabeled data, which can be gathered easily. Making use of both labeled and unlabeled data is known as semi-supervised learning. One of the most general versions of semi-supervised learning is self-training, where a recognizer iteratively retrains itself on its own output on new, unlabeled data. In this paper we propose to apply semi-supervised learning, and in particular self-training, to the problem of cursive, handwritten word recognition. The special focus of the paper is on retraining rules that define what data are actually being used in the retraining phase. In a series of experiments it is shown that the performance of a neural network based recognizer can be significantly improved through the use of unlabeled data and self-training if appropriate retraining rules are applied.

Improving Collaborative Filtering Based Recommender Systems Using Pareto Dominance

Los sistemas de recomendaci��n son un tipo de soluci��n al problema de sobrecarga de informaci��n que sufren los usuarios de los sitios web en los que se pueden votar ciertos art��culos. El sistema de recomendaci��n de filtrado colaborativo es considerado como el m��todo con m��s ��xito debido a que sus recomendaciones se hacen bas��ndose en los votos de usuarios similares a un usuario activo. Sin embargo, el m��todo de filtrado de colaboraci��n tradicional selecciona usuarios insuficientemente representativos como vecinos de cada usuario activo. Esto significa que las recomendaciones hechas a posteriori no son lo suficientemente precisas. El m��todo propuesto en esta tesis realiza un pre-filtrado del proceso, mediante el uso de dominancia de Pareto, que elimina los usuarios menos representativos del proceso de selecci��n k-vecino y mantiene los m��s prometedores. Los resultados de los experimentos realizados en MovieLens y Netflix muestran una mejora significativa en todas las medidas de calidad estudiadas en la aplicaci��n del m��todo propuesto. ABSTRACTRecommender systems are a type of solution to the information overload problem suffered by users of websites on which they can rate certain items. The Collaborative Filtering Recommender System is considered to be the most successful approach as it make its recommendations based on votes of users similar to an active user. Nevertheless, the traditional collaborative filtering method selects insufficiently representative users as neighbors of each active user. This means that the recommendations made a posteriori are not precise enough. The method proposed in this thesis performs a pre-filtering process, by using Pareto dominance, which eliminates the less representative users from the k-neighbor selection process and keeps the most promising ones. The results from the experiments performed on Movielens and Netflix show a significant improvement in all the quality measures studied on applying the proposed method.

Artificial pancreas using a personalized rule-based controller achieves overnight normoglycemia in patients with type 1 diabetes

Objective: This study assessed the ef���cacy of a closed-loop (CL) system consisting of a predictive rule-based algorithm (pRBA) on achieving nocturnal and postprandial normoglycemia in patients with type 1 diabetes mellitus (T1DM). The algorithm is personalized for each patient���s data using two different strategies to control nocturnal and postprandial periods. Research Design and Methods: We performed a randomized crossover clinical study in which 10 T1DM patients treated with continuous subcutaneous insulin infusion (CSII) spent two nonconsecutive nights in the research facility: one with their usual CSII pattern (open-loop [OL]) and one controlled by the pRBA (CL). The CL period lasted from 10 p.m. to 10 a.m., including overnight control, and control of breakfast. Venous samples for blood glucose (BG) measurement were collected every 20 min. Results: Time spent in normoglycemia (BG, 3.9���8.0 mmol/L) during the nocturnal period (12 a.m.���8 a.m.), expressed as median (interquartile range), increased from 66.6% (8.3���75%) with OL to 95.8% (73���100%) using the CL algorithm (P<0.05). Median time in hypoglycemia (BG, <3.9 mmol/L) was reduced from 4.2% (0���21%) in the OL night to 0.0% (0.0���0.0%) in the CL night (P<0.05). Nine hypoglycemic events (<3.9 mmol/L) were recorded with OL compared with one using CL. The postprandial glycemic excursion was not lower when the CL system was used in comparison with conventional preprandial bolus: time in target (3.9���10.0 mmol/L) 58.3% (29.1���87.5%) versus 50.0% (50���100%). Conclusions: A highly precise personalized pRBA obtains nocturnal normoglycemia, without signi���cant hypoglycemia, in T1DM patients. There appears to be no clear bene���t of CL over prandial bolus on the postprandial glycemia

Personalized rule-based closed-loop control algorithm for type 1 diabetes

Type 1 diabetes-mellitus implies a life-threatening absolute insulin deficiency. Artificial pancreas (CGM sensor, insulin pump and control algorithm) is promising to outperform current open-loop therapies.

RESSIST: a research object-based recommender system

The scientific method is a methodological approach to the process of inquiry { in which empirically grounded theory of nature is constructed and verified [14]. It is a hard, exhaustive and dedicated multi-stage procedure that a researcher must perform to achieve valuable knowledge. Trying to help researchers during this process, a recommender system, intended as a researcher assistant, is designed to provide them useful tools and information for each stage of the procedure. A new similarity measure between research objects and a representational model, based on domain spaces, to handle them in dif ferent levels are created as well as a system to build them from OAI-PMH (and RSS) resources. It tries to represents a sound balance between scientific insight into individual scientific creative processes and technical implementation using innovative technologies in information extraction, document summarization and semantic analysis at a large scale.

Formal considerations of rule-based messaging for Business Process Integration

In the last decade, with the expansion of organizational scope and the tendency for outsourcing, there has been an increasing need for Business Process Integration (BPI), understood as the sharing of data and applications among business processes. The research efforts and development paths in BPI pursued by many academic groups and system vendors, targeting heterogeneous system integration, continue to face several conceptual and technological challenges. This article begins with a brief review of major approaches and emerging standards to address BPI. Further, we introduce a rule-driven messaging approach to BPI, which is based on the harmonization of messages in order to compose a new, often cross-organizational process. We will then introduce the design of a temporal first order language (Harmonized Messaging Calculus) that provides the formal foundation for general rules governing the business process execution. Definitions of the language terms, formulae, safety, and expressiveness are introduced and considered in detail.

A rule based model for the objective determination of the e-business requirements of a manufacturing organisation

The starting point of this research was the belief that manufacturing and similar industries need help with the concept of e-business, especially in assessing the relevance of possible e-business initiatives. The research hypotheses was that it should be possible to produce a systematic model that defines, at a useful level of detail, the probable e-business requirements of an organisation based on objective criteria with an accuracy of 85%-90%. This thesis describes the development and validation of such a model. A preliminary model was developed from a variety of sources, including a survey of current and planned e-business activity and representative examples of e-business material produced by e-business solution providers. The model was subject to a process of testing and refinement based on recursive case studies, with controls over the improving accuracy and stability of the model. Useful conclusions were also possible as to the relevance of e-business functions to the case study participants themselves. Techniques were evolved to synthesise the e-business requirements of an organisation and present them at a management summary level of detail. The results of applying these techniques to all the case studies used in this research were discussed. The conclusion of the research was that the case study methodology employed was successful. A model was achieved suitable for practical application in a manufacturing organisation requiring help with a requirements definition process.

Limit Behaviour of Dynamic Rule-based Systems

The paper suggests a classification of dynamic rule-based systems. For each class of systems, limit behavior is studied. Systems with stabilizing limit states or stabilizing limit trajectories are identified, and such states and trajectories are found. The structure of the set of limit states and trajectories is investigated.