Most empirical work in human categorization has studied learning in either fully supervised or fully unsupervised scenarios. Most real-world learning scenarios, however, are semi-supervised: Learners receive a great deal of unlabeled information from the world, coupled with occasional experiences in which items are directly labeled by a knowledgeable source. A large body of work in machine learning has investigated how learning can exploit both labeled and unlabeled data provided to a learner. Using equivalences (...) between models found in human categorization and machine learning research, we explain how these semi-supervised techniques can be applied to human learning. A series of experiments are described which show that semi-supervised learning models prove useful for explaining human behavior when exposed to both labeled and unlabeled data. We then discuss some machine learning models that do not have familiar human categorization counterparts. Finally, we discuss some challenges yet to be addressed in the use of semi-supervised models for modeling human categorization. (shrink)

Three experiments with 88 college-aged participants explored how unlabeled experiences—learning episodes in which people encounter objects without information about their category membership—influence beliefs about category structure. Participants performed a simple one-dimensional categorization task in a brief supervised learning phase, then made a large number of unsupervised categorization decisions about new items. In all three experiments, the unsupervised experience altered participants’ implicit and explicit mental category boundaries, their explicit beliefs about the most representative members of each category, and (...) even their memory for the items encountered during the supervised learning phase. These changes were influenced by both the range and frequency distribution of the unlabeled stimuli: mental category boundaries shifted toward the middle of the range and toward the trough of the bimodal distribution of unlabeled items, whereas beliefs about the most representative category members shifted toward the modes of the unlabeled distribution. One consequence of this shift in representations is a false-consensus effect (Experiment 3) where participants, despite receiving very disparate training experiences, show strong agreement in judgments about representativeness and boundary location following unsupervised category judgments. (shrink)

The purpose of cognitive diagnostic modeling is to classify students' latent attribute profiles using their responses to the diagnostic assessment. In recent years, each diagnostic classification model makes different assumptions about the relationship between a student's response pattern and attribute profile. The previous research studies showed that the inappropriate DCMs and inaccurate Q-matrix impact diagnostic classification accuracy. Artificial Neural Networks have been proposed as a promising approach to convert a pattern of item responses into a diagnostic classification in some research (...) studies. However, the ANNs methods produced very unstable and unappreciated estimation unless a great deal of care was taken. In this research, we combined ANNs with two typical DCMs, the deterministic-input, noisy, “and” gate model and the deterministic-inputs, noisy, “or” gate model, within a semi-supervised learning framework to achieve a robust and accurate classification. In both simulated study and real data study, the experimental results showed that the proposed method could achieve appreciated performance across different test conditions, especially when the diagnostic quality of assessment was not high and the Q-matrix contained misspecified elements. This research study is the first time of applying the thinking of semi-supervised learning into CDM. Also, we used the validating test to choose the appropriate parameters for the ANNs instead of using typical statistical criteria. (shrink)

Computational drug repositioning has been proven as a promising and efficient strategy for discovering new uses from existing drugs. To achieve this goal, a number of computational methods have been proposed, which are based on different data sources of drugs and diseases. These methods approach the problem using either machine learning- or network-based models with an assumption that similar drugs can be used for similar diseases to identify new indications of drugs. Therefore, similarities between drugs and between diseases are (...) usually used as inputs. In addition, known drug-disease associations are also needed for the methods as prior information. It should be noted that those associations are still not well established due to the fact that many of marketed drugs have been withdrawn and this could affect the outcome of the methods. In this study, we propose a novel method named RLSDR to find new uses of drugs. More specifically, it relies on a semi-supervised learning model, Regularized Least Square, thus it does not require definition of non-drug-disease associations as previously proposed machine learning-based methods. In addition, the similarity between drugs measured by chemical structures of drug compounds and the similarity between diseases which share phenotypes can be represented in a form of either similarity network or similarity matrix as inputs of the method. Moreover, instead of using a gold-standard set of known drug-disease associations, we construct an artificial set of the associations based on known disease-gene and drug-target associations. Experiment results demonstrate that RLSDR achieves better prediction performance on the artificial set of drug-disease associations than that on the gold-standard ones in terms of area under the Receiver Operating Characteristic curve. In addition, it outperforms two representative network-based methods irrespective of the prior information of drug-disease associations. Novel indications for a number of drugs are also identified and validated by evidences from a different data resource. (shrink)

Predicting the defect-prone modules when the previous defect labels of modules are limited is a challenging problem encountered in the software industry. Supervised classification approaches cannot build high-performance prediction models with few defect data, leading to the need for new methods, techniques, and tools. One solution is to combine labeled data points with unlabeled data points during learning phase. Semi-supervised classification methods use not only labeled data points but also unlabeled ones to improve the generalization capability. (...) In this study, we evaluated four semi-supervised classification methods for semi-supervised defect prediction. Low-density separation, support vector machine, expectation-maximization, and class mass normalization methods have been investigated on NASA data sets, which are CM1, KC1, KC2, and PC1. Experimental results showed that SVM and LDS algorithms outperform CMN and EM-SEMI algorithms. In addition, LDS algorithm performs much better than SVM when the data set is large. In this study, the LDS-based prediction approach is suggested for software defect prediction when there are limited fault data. (shrink)

This paper reconsiders the diphone-based word segmentation model of Cairns, Shillcock, Chater, and Levy (1997) and Hockema (2006), previously thought to be unlearnable. A statistically principled learning model is developed using Bayes’ theorem and reasonable assumptions about infants’ implicit knowledge. The ability to recover phrase-medial word boundaries is tested using phonetic corpora derived from spontaneous interactions with children and adults. The (unsupervised and semi-supervised) learning models are shown to exhibit several crucial properties. First, only a small (...) amount of language exposure is required to achieve the model’s ceiling performance, equivalent to between 1 day and 1 month of caregiver input. Second, the models are robust to variation, both in the free parameter and the input representation. Finally, both the learning and baseline models exhibit undersegmentation, argued to have significant ramifications for speech processing as a whole. (shrink)

A beautiful and thoughtful collection of essays on reading, writing and learning, _Writing and Psychoanalysis_ grows out of a colloquium. The results are wondrous and impact on the reader at many different levels. In the act of writing, we all discover something about what we know previously unknown to us, and we learn more about our inner world that we knew before we set pen to paper. Patrick Mahony goes so far as to argue that Freud's self-analysis was essentially (...) a "writing cure." _Writing in Psychoanalysis_ is the first volume in the projected Monograph Series, Psychoanalytic Issues, the _Rivista di Psicoanalisi_ is undertaking in conjunction with Karnac Books. This series constitutes a major effort to bring about a dialogue among psychoanalysts who while ultimately bound together by a common psychoanalytic heritage nonetheless are separated in their thinking by different idioms, whether linguistic or theoretical. While featuring writers of very different idioms, this series will also present a venue to make some important Italian voices known to English speaking analysts. (shrink)

Connected vehicle fleets have formed significant component of industrial internet of things scenarios as part of Industry 4.0 worldwide. The number of vehicles in these fleets has grown at a steady pace. The vehicles monitoring with machine learning algorithms has significantly improved maintenance activities. Predictive maintenance potential has increased where machines are controlled through networked smart devices. Here, benefits are accrued considering uptimes optimization. This has resulted in reduction of associated time and labor costs. It has also provided significant (...) increase in cost benefit ratios. Considering vehicle fault trends in this research predictive maintenance problem is addressed through hybrid deep learning-based ensemble method (HDLEM). The ensemble framework which acts as predictive analytics engine comprises of three deep learning algorithms viz modified cox proportional hazard deep learning (MCoxPHDL), modified deep learning embedded semi supervised learning (MDLeSSL) and merged LSTM (MLSTM) networks. Both sensor as well as historical maintenance data are collected and prepared using benchmarking methods for HDLEM training and testing. Here, times between failures (TBF) modeling and prediction on multi-source data are successfully achieved. The results obtained are compared with stated deep learning models. This ensemble framework offers great potential towards achieving more profitable, efficient and sustainable vehicle fleet management solutions. This helps better telematics data implementation which ensures preventative management towards desired solution. The ensemble method's superiority is highlighted through several experimental results. (shrink)

This article reviews a number of different areas in the foundations of formal learning theory. After outlining the general framework for formal models of learning, the Bayesian approach to learning is summarized. This leads to a discussion of Solomonoff's Universal Prior Distribution for Bayesian learning. Gold's model of identification in the limit is also outlined. We next discuss a number of aspects of learning theory raised in contributed papers, related to both computational and representational complexity. (...) The article concludes with a description of how semi-supervised learning can be applied to the study of cognitive learning models. Throughout this overview, the specific points raised by our contributing authors are connected to the models and methods under review. (shrink)

Most current online distributed machine learning algorithms have been studied in a data-parallel architecture among agents in networks. We study online distributed machine learning from a different perspective, where the features about the same samples are observed by multiple agents that wish to collaborate but do not exchange the raw data with each other. We propose a distributed feature online gradient descent algorithm and prove that local solution converges to the global minimizer with a sublinear rate O 2 (...) T. Our algorithm does not require exchange of the primal data or even the model parameters between agents. Firstly, we design an auxiliary variable, which implies the information of the global features, and estimate at each agent by dynamic consensus method. Then, local parameters are updated by online gradient descent method based on local data stream. Simulations illustrate the performance of the proposed algorithm. (shrink)

Formal learning theory formalizes the process of inferring a general result from examples, as in the case of inferring grammars from sentences when learning a language. In this work, we develop a general framework—the supervised learning game—to investigate the interaction between Teacher and Learner. In particular, our proposal highlights several interesting features of the agents: on the one hand, Learner may make mistakes in the learning process, and she may also ignore the potential relation between (...) different hypotheses; on the other hand, Teacher is able to correct Learner’s mistakes, eliminate potential mistakes and point out the facts ignored by Learner. To reason about strategies in this game, we develop a modal logic of supervised learning and study its properties. Broadly, this work takes a small step towards studying the interaction between graph games, logics and formal learning theory. (shrink)

Scenarios for the emergence or bootstrap of a lexicon involve the repeated interaction between at least two agents who must reach a consensus on how to name N objects using H words. Here we consider minimal models of two types of learning algorithms: cross-situational learning, in which the individuals determine the meaning of a word by looking for something in common across all observed uses of that word, and supervised operant conditioning learning, in which there is (...) strong feedback between individuals about the intended meaning of the words. Despite the stark differences between these learning schemes, we show that they yield the same communication accuracy in the limits of large N and H , which coincides with the result of the classical occupancy problem of randomly assigning N objects to H words. Key words: lexicon bootstrapping; cross-situational learning; supervised learning; random occupancy problems. (shrink)

Scenarios for the emergence or bootstrap of a lexicon involve the repeated interaction between at least two agents who must reach a consensus on how to name N objects using H words. Here we consider minimal models of two types of learning algorithms: cross-situational learning, in which the individuals determine the meaning of a word by looking for something in common across all observed uses of that word, and supervised operant conditioning learning, in which there is (...) strong feedback between individuals about the intended meaning of the words. Despite the stark differences between these learning schemes, we show that they yield the same communication accuracy in the limits of large N and H, which coincides with the result of the classical occupancy problem of randomly assigning N objects to H words. Key words: lexicon bootstrapping; cross-situational learning; supervised learning; random occupancy problems. (shrink)

Malware is a sophisticated, malicious, and sometimes unidentifiable application on the network. The classifying network traffic method using machine learning shows to perform well in detecting malware. In the literature, it is reported that this good performance can depend on a reduced set of network features. This study presents an empirical evaluation of two statistical methods of reduction and selection of features in an Android network traffic dataset using six supervised algorithms: Naïve Bayes, support vector machine, multilayer perceptron (...) neural network, decision tree, random forest, and K-nearest neighbors. The principal component analysis and logistic regression methods with p value were applied to select the most representative features related to the time properties of flows and features of bidirectional packets. The selected features were used to train the algorithms using binary and multiclass classification. For performance evaluation and comparison metrics, precision, recall, F-measure, accuracy, and area under the curve were used. The empirical results show that random forest obtains an average accuracy of 96% and an AUC-ROC of 0.98 in binary classification. For the case of multiclass classification, again random forest achieves an average accuracy of 87% and an AUC-ROC over 95%, exhibiting better performance than the other machine learning algorithms. In both experiments, the 13 most representative features of a mixed set of flow time properties and bidirectional network packets selected by LR were used. In the case of the other five classifiers, their results in terms of precision, recall, and accuracy, are competitive with those obtained in related works, which used a greater number of input features. Therefore, it is empirically evidenced that the proposed method for the selection of features, based on statistical techniques of reduction and extraction of attributes, allows improving the identification performance of malware traffic, discriminating it from the benign traffic of Android applications. (shrink)

This article puts forward a novel class of Bézier smooth semi-supervised support vector machines(BS4VMs) for classification. As is well known, semi-supervised support vector machine is introduced for dealing with quantities of unlabeled data in the real world. Labeled data is utilized to train the algorithm and then adapting it to classify the unlabeled data. However, the objective semi-supervised function is not differentiable globally. It is required to endure heavy burden in solving two quadratic programming (...) problems with inversion matrix operation. To solve this problem, the interpolation Bézier functions for replacing the non-smooth term have been investigated. Firstly, the development of the novel BS4VMs classifier is introduced. Then, the non-linear model and convergence analysis are presented. To show how the BS4VMs can be practically implemented, experiments on synthetic and real data sets are offered. The theoretical analysis and experiments comparisons clearly illustrate the BS4VMs are superior to the proposed methods in both cross-validation classification accuracy and calculating time. (shrink)

The no-free-lunch theorems promote a skeptical conclusion that all possible machine learning algorithms equally lack justification. But how could this leave room for a learning theory, that shows that some algorithms are better than others? Drawing parallels to the philosophy of induction, we point out that the no-free-lunch results presuppose a conception of learning algorithms as purely data-driven. On this conception, every algorithm must have an inherent inductive bias, that wants justification. We argue that many standard (...) class='Hi'>learning algorithms should rather be understood as model-dependent: in each application they also require for input a model, representing a bias. Generic algorithms themselves, they can be given a model-relative justification. (shrink)

One of the most active areas of research in semi-supervised learning has been to study methods for constructing good ensembles of classifiers. Ensemble systems are techniques that create multiple models and then combine them to produce improved results. These systems usually produce more accurate solutions than a single model would. Specially, multi-view ensemble systems improve the accuracy of text classification because they optimize the functions to exploit different views of the same input data. However, despite being more (...) promising than the single-view approaches, document datasets often have no natural multiple views available. This study proposes an algorithm to generate a synthetic view from a standard text dataset. The model generates a new view from the standard bag-of-words approach using an algorithm based on hidden Markov models. To show the effectiveness of the proposed HMM-based synthetic view generation method, it has been integrated in a co-training ensemble system and tested with four text corpora: Reuters, 20 Newsgroup, TREC Genomics and OHSUMED. The results obtained are promising, showing a significant increase in the efficiency of the ensemble system compared to a single-view approach. (shrink)

Some of the features of animal and human categorical perception (CP) for color, pitch and speech are exhibited by neural net simulations of CP with one-dimensional inputs: When a backprop net is trained to discriminate and then categorize a set of stimuli, the second task is accomplished by "warping" the similarity space (compressing within-category distances and expanding between-category distances). This natural side-effect also occurs in humans and animals. Such CP categories, consisting of named, bounded regions of similarity space, may be (...) the ground level out of which higher-order categories are constructed; nets are one possible candidate for the mechanism that learns the sensorimotor invariants that connect arbitrary names (elementary symbols?) to the nonarbitrary shapes of objects. This paper examines how and why such compression/expansion effects occur in neural nets. (shrink)

Patent landscaping is the process of finding patents related to a particular topic. It is important for companies, investors, governments, and academics seeking to gauge innovation and assess risk. However, there is no broadly recognized best approach to landscaping. Frequently, patent landscaping is a bespoke human-driven process that relies heavily on complex queries over bibliographic patent databases. In this paper, we present Automated Patent Landscaping, an approach that jointly leverages human domain expertise, heuristics based on patent metadata, and machine (...) class='Hi'>learning to generate high-quality patent landscapes with minimal effort. In particular, this paper describes a flexible automated methodology to construct a patent landscape for a topic based on an initial seed set of patents. This approach takes human-selected seed patents that are representative of a topic, such as operating systems, and uses structure inherent in patent data such as references and class codes to “expand” the seed set to a set of “probably-related” patents and anti-seed “probably-unrelated” patents. The expanded set of patents is then pruned with a semi-supervised machine learning model trained on seed and anti-seed patents. This removes patents from the expanded set that are unrelated to the topic and ensures a comprehensive and accurate landscape. (shrink)

. Network data analysis helps in capturing node usage behavior. Existing algorithms use reduced feature set to manage high runtime complexity. Ignoring features may increase classification errors. This paper presents a model, allowing classification of network traffic, while considering all the relevant features. Learning phase partitions training sample on values of the respective features. This creates equivalence classes related to m features. During classification, each feature value of the test instance results in picking one set from equivalence class generated (...) during learning. Algorithm captures new behavior in semi-supervised incremental learning mode. For problems having m features and n training samples the model has incremental learning complexity of and average classification complexity is of the order. (shrink)

The invisibility of nursing work has been discussed in the international literature but not in relation to learning clinical skills. Evans and Guile's (Practice‐based education: Perspectives and strategies, Rotterdam: Sense, 2012) theory of recontextualisation is used to explore the ways in which invisible or unplanned and unrecognised learning takes place as newly qualified nurses learn to delegate to and supervise the work of the healthcare assistant. In the British context, delegation and supervision are thought of as skills which (...) are learnt “on the job.” We suggest that learning “on‐the‐job” is the invisible construction of knowledge in clinical practice and that delegation is a particularly telling area of nursing practice which illustrates invisible learning. Using an ethnographic case study approach in three hospital sites in England from 2011 to 2014, we undertook participant observation, interviews with newly qualified nurses, ward managers and healthcare assistants. We discuss the invisible ways newly qualified nurses learn in the practice environment and present the invisible steps to learning which encompass the embodied, affective and social, as much as the cognitive components to learning. We argue that there is a need for greater understanding of the “invisible learning” which occurs as newly qualified nurses learn to delegate and supervise. (shrink)

In the last decades, supervised machine learning has seen the widespread growth of highly complex, non-interpretable models, of which deep neural networks are the most typical representative. Due to their complexity, these models have showed an outstanding performance in a series of tasks, as in image recognition and machine translation. Recently, though, there has been an important discussion over whether those non-interpretable models are able to provide any sort of understanding whatsoever. For some scholars, only interpretable models can (...) provide understanding. More popular, however, is the idea that understanding can come from a careful analysis of the dataset or from the model’s theoretical basis. In this paper, I wish to examine the possible forms of obtaining understanding of such non-interpretable models. Two main strategies for providing understanding are analyzed. The first involves understanding without interpretability, either through external evidence for the model’s inner functioning or through analyzing the data. The second is based on the artificial production of interpretable structures, through three main forms: post hoc models, hybrid models, and quasi-interpretable structures. Finally, I consider some of the conceptual difficulties in the attempt to create explanations for these models, and their implications for understanding. (shrink)

The free energy principle says that any self-organising system that is at nonequilibrium steady-state with its environment must minimize its free energy. It is proposed as a grand unifying principle for cognitive science and biology. The principle can appear cryptic, esoteric, too ambitious, and unfalsifiable—suggesting it would be best to suspend any belief in the principle, and instead focus on individual, more concrete and falsifiable ‘process theories’ for particular biological processes and phenomena like perception, decision and action. Here, I explain (...) the free energy principle, and I argue that it is best understood as offering a conceptual and mathematical analysis of the concept of existence of self-organising systems. This analysis offers a new type of solution to long-standing problems in neurobiology, cognitive science, machine learning and philosophy concerning the possibility of normatively constrained, self-supervised learning and inference. The principle can therefore uniquely serve as a regulatory principle for process theories, to ensure that process theories conforming to it enable self-supervision. This is, at least for those who believe self-supervision is a foundational explanatory task, good reason to believe the free energy principle. (shrink)

The article describes the author's approach to the construction of general-level artificial cognitive agents based on the so-called "semantic supervised learning", within which, in accordance with the hybrid paradigm of artificial intelligence, both machine learning methods and methods of the symbolic ap­ proach and knowledge-based systems are used ("good old-fashioned artificial intelligence"). А descrip­ tion of current proЬlems with understanding of the general meaning and context of situations in which narrow AI agents are found is presented. The (...) definition of semantic supervised learning is given and its relationship with other machine learning methods is described. In addition, а thought experiment is presented, which shows the essence and meaning of supervised semantic learning. (shrink)

Doctoral students in clinical, counseling, and school psychology programs often collaborate with faculty on research projects in their training as scientist-practitioners. Yet, the determination of publications' credit and order of authorship on resulting manuscripts continues to be a major concern and challenging process for professional psychologists and student collaborators. This article describes the use of case-based learning and semi-structured interview approaches to instruct first-year clinical psychology doctoral students in publication ethics during a research seminar. The instructor models ethical (...) decision-making with 1) a discussion of four cases from his own professional experiences and 2) a description of the Authorship Eligibility Assessment form, which he developed for use with junior researchers. The authors advocate for more educational strategies to supplement the APA standards in teaching the ethics of the publication process. Implications in terms of graduate student development and research collaborations in the field of professional psychology were discussed. (shrink)