Learning with Privileged Information for Improved Target Classification

2014 ◽  
Vol 2 (3) ◽  
pp. 50-66 ◽  
Author(s):  
Roman Ilin ◽  
Simon Streltsov ◽  
Rauf Izmailov
Keyword(s):  
Classification Accuracy ◽  
Wide Area ◽  
Future Research ◽  
Target Classification ◽  
Privileged Information ◽  
Training Time ◽  
Additional Information ◽  
Motion Imagery ◽  
Learning Using Privileged Information

This work considers “Learning Using Privileged Information” (LUPI) paradigm. LUPI improves classification accuracy by incorporating additional information available at training time and not available during testing. In this contribution, the LUPI paradigm is tested on a Wide Area Motion Imagery (WAMI) dataset and on images from the Caltech 101 dataset. In both cases a consistent improvement in classification accuracy is observed. The results are discussed and the directions of future research are outlined.

scholarly journals Federated Quantum Machine Learning

Entropy ◽  
2021 ◽  
Vol 23 (4) ◽  
pp. 460
Author(s):  
Samuel Yen-Chi Chen ◽  
Shinjae Yoo
Keyword(s):  
Machine Learning ◽  
Data Privacy ◽  
Research Direction ◽  
Future Research ◽  
Quantum Computers ◽  
Training Time ◽  
Quantum Neural Network ◽  
Distributed Training ◽  
Machine Learning Model ◽  
Quantum Machine Learning

Distributed training across several quantum computers could significantly improve the training time and if we could share the learned model, not the data, it could potentially improve the data privacy as the training would happen where the data is located. One of the potential schemes to achieve this property is the federated learning (FL), which consists of several clients or local nodes learning on their own data and a central node to aggregate the models collected from those local nodes. However, to the best of our knowledge, no work has been done in quantum machine learning (QML) in federation setting yet. In this work, we present the federated training on hybrid quantum-classical machine learning models although our framework could be generalized to pure quantum machine learning model. Specifically, we consider the quantum neural network (QNN) coupled with classical pre-trained convolutional model. Our distributed federated learning scheme demonstrated almost the same level of trained model accuracies and yet significantly faster distributed training. It demonstrates a promising future research direction for scaling and privacy aspects.

scholarly journals Textual Analysis and Sentiment Analysis in Accounting

2021 ◽  
Vol 24 (2) ◽  
pp. 168-183
Author(s):  
Juan L. Gandía ◽  
David Huguet
Keyword(s):  
Sentiment Analysis ◽  
Textual Analysis ◽  
Financial Information ◽  
Digital Tools ◽  
Future Research ◽  
Analysis Techniques ◽  
Additional Information ◽  
New Perspective ◽  
Medios De Comunicacion

A pesar del relativamente escaso uso de técnicas de análisis textual y de análisis del sentimiento en finanzas y contabilidad, éstas tienen un gran potencial en contabilidad, tanto por el elevado volumen de documentos utilizados para la comunicación de información financiera como por el crecimiento en el uso de herramientas digitales y medios de comunicación social. En este sentido, estas técnicas de análisis pueden ayudar a los investigadores a analizar pistas ocultas o buscar información adicional a la observada a través de los estados financieros, incrementando la cantidad y calidad de la información tradicionalmente utilizada, y proporcionando una nueva perspectiva de análisis. Por ello, el objetivo de este estudio es realizar una revisión del uso del análisis textual y del análisis del sentimiento en contabilidad. Tras presentar los conceptos de análisis textual y análisis del sentimiento y justificar teóricamente su papel en la investigación en contabilidad, llevamos a cabo una revisión de la literatura previa en el uso de estas técnicas en finanzas y contabilidad y describimos las principales técnicas de análisis del sentimiento, así como el procedimiento a seguir para el uso de esta metodología. Finalmente, sugerimos tres líneas de investigación futura que pueden beneficiarse del uso del análisis textual y del análisis del sentimiento. In spite of the relatively scarce use of textual analysis and sentiment analysis techniques in finance and accounting, they have great potential in accounting, both because of the volume of documents used for the communication of information and due to the growth in the use of digital tools and social media. In that regard, these techniques of analysis may help researchers to analyse hidden clues or look for additional information to that one observed through financial information, increasing the quantity and quality of the information traditionally used, and providing a new perspective of analysis. The aim of this study is to review the use of textual analysis and sentiment analysis in accounting. After presenting the concepts of textual analysis and sentiment analysis and expose their interest in accounting, we perform a review of the previous literature on the use of these techniques in finance and accounting and describe the main techniques of sentiment analysis, as well as the procedure to be followed for the use of this methodology. Finally, we suggest three lines of future research that may benefit from the use of textual and sentiment analysis.

scholarly journals A New Subject-Specific Discriminative and Multi-Scale Filter Bank Tangent Space Mapping Method for Recognition of Multiclass Motor Imagery

2021 ◽  
Vol 15 ◽  
Author(s):  
Fan Wu ◽  
Anmin Gong ◽  
Hongyun Li ◽  
Lei Zhao ◽  
Wei Zhang ◽  
...  
Keyword(s):  
Classification Accuracy ◽  
Tangent Space ◽  
Filter Bank ◽  
Space Mapping ◽  
Training Time ◽  
Frequency Bands ◽  
Multi Scale ◽  
Average Accuracy ◽  
Subject Specific ◽  
The Subject

Objective: Tangent Space Mapping (TSM) using the geometric structure of the covariance matrices is an effective method to recognize multiclass motor imagery (MI). Compared with the traditional CSP method, the Riemann geometric method based on TSM takes into account the nonlinear information contained in the covariance matrix, and can extract more abundant and effective features. Moreover, the method is an unsupervised operation, which can reduce the time of feature extraction. However, EEG features induced by MI mental activities of different subjects are not the same, so selection of subject-specific discriminative EEG frequency components play a vital role in the recognition of multiclass MI. In order to solve the problem, a discriminative and multi-scale filter bank tangent space mapping (DMFBTSM) algorithm is proposed in this article to design the subject-specific Filter Bank (FB) so as to effectively recognize multiclass MI tasks.Methods: On the 4-class BCI competition IV-2a dataset, first, a non-parametric method of multivariate analysis of variance (MANOVA) based on the sum of squared distances is used to select discriminative frequency bands for a subject; next, a multi-scale FB is generated according to the range of these frequency bands, and then decompose multi-channel EEG of the subject into multiple sub-bands combined with several time windows. Then TSM algorithm is used to estimate Riemannian tangent space features in each sub-band and finally a liner Support Vector Machines (SVM) is used for classification.Main Results: The analysis results show that the proposed discriminative FB enhances the multi-scale TSM algorithm, improves the classification accuracy and reduces the execution time during training and testing. On the 4-class BCI competition IV-2a dataset, the average session to session classification accuracy of nine subjects reached 77.33 ± 12.3%. When the training time and the test time are similar, the average classification accuracy is 2.56% higher than the latest TSM method based on multi-scale filter bank analysis technology. When the classification accuracy is similar, the training speed is increased by more than three times, and the test speed is increased two times more. Compared with Supervised Fisher Geodesic Minimum Distance to the Mean (Supervised FGMDRM), another new variant based on Riemann geometry classifier, the average accuracy is 3.36% higher, we also compared with the latest Deep Learning method, and the average accuracy of 10-fold cross validation improved by 2.58%.Conclusion: Research shows that the proposed DMFBTSM algorithm can improve the classification accuracy of MI tasks.Significance: Compared with the MFBTSM algorithm, the algorithm proposed in this article is expected to select frequency bands with good separability for specific subject to improve the classification accuracy of multiclass MI tasks and reduce the feature dimension to reduce training time and testing time.

Categorizing Blogs as Information Sources for Libraries and Information Science

2015 ◽  
pp. 4833-4845
Author(s):  
Mark-Shane Scale ◽  
Anabel Quan-Haase
Keyword(s):  
Social Media ◽  
Information Needs ◽  
Information Sources ◽  
Information Science ◽  
Information Source ◽  
Future Research ◽  
Collection Development ◽  
Sources Of Information ◽  
Additional Information ◽  
Source Category

Blogs are important sources of information currently used in the work of professionals, institutions and academics. Nevertheless, traditional information needs and uses research has not yet discussed where blogs fit in the existing typologies of information sources. Blogs and other types of social media have several characteristics that blur the lines of distinction existent between traditional information source categories. This chapter brings this research problem to the fore. Not only do we examine why blogs do not neatly fit into existing information source categories, but we also deliberate the implications for libraries in terms of the need to consider blogs as an information source to be included in collection development. We discuss the opportunities and possibilities for blogs to be integrated into the collection development efforts of academic and public libraries to better serve patrons. In order to accommodate for blogs and other types of social media as information sources, we propose the introduction of an additional information source category. We suggest new avenues of future research that investigate how blogs are being used to meet information needs in various social settings, such as corporations, health care and educational settings (e.g., higher education, and schools). In this chapter, we develop a framework of how blogs may function as information sources to provide libraries with a better understanding of how blogs are integrated into the context of everyday information seeking. By grouping the ways in which people employ blogs to acquire information, we propose that blogs provide information sources along a continuum ranging from non-fiction to fictional information.

Central Language Hypothesis in the Decision-Making Process

2016 ◽  
pp. 66-83 ◽  
Author(s):  
Duygu Buğa
Keyword(s):  
Decision Making ◽  
Theoretical Framework ◽  
Future Research ◽  
Limbic Cortex ◽  
Decision Making Process ◽  
Discussion Section ◽  
Research Directions ◽  
Additional Information ◽  
Future Research Directions ◽  
The Brain

The purpose of this chapter is to explore the potential connection between neuroeconomics and the Central Language Hypothesis (CLH) which refers to the language placed within the subconscious mind of an individual. The CLH forwards that in the brains of bilingual and multilingual people, one language is more suppressive as it dominates reflexes, emotions, and senses. This central language (CL) is located at the centre of the limbic cortex of the brain. Therefore, when there is a stimulus on the limbic cortex (e.g., fear, anxiety, sadness), the brain produces the central language. The chapter begins with an Introduction followed by a Theoretical Framework. The next section discusses the neurolinguistic projection of the central language and includes the survey and the results used in this study. The Discussion section provides additional information regarding the questionnaire and the CLH, followed by Future Research Directions, Implications, and finally the Conclusion.

An Improved Cross-Domain Sentiment Analysis Based on a Semi-Supervised Convolutional Neural Network

2022 ◽  
pp. 155-170
Author(s):  
Lap-Kei Lee ◽  
Kwok Tai Chui ◽  
Jingjing Wang ◽  
Yin-Chun Fung ◽  
Zhanhui Tan
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Sentiment Analysis ◽  
Language Processing ◽  
Research Work ◽  
Latent Semantic Indexing ◽  
Future Research ◽  
Training Time ◽  
Cross Domain ◽  
Document Frequency

The dependence on Internet in our daily life is ever-growing, which provides opportunity to discover valuable and subjective information using advanced techniques such as natural language processing and artificial intelligence. In this chapter, the research focus is a convolutional neural network for three-class (positive, neutral, and negative) cross-domain sentiment analysis. The model is enhanced in two-fold. First, a similarity label method facilitates the management between the source and target domains to generate more labelled data. Second, term frequency-inverse document frequency (TF-IDF) and latent semantic indexing (LSI) are employed to compute the similarity between source and target domains. Performance evaluation is conducted using three datasets, beauty reviews, toys reviews, and phone reviews. The proposed method enhances the accuracy by 4.3-7.6% and reduces the training time by 50%. The limitations of the research work have been discussed, which serve as the rationales of future research directions.

Modeling, Experiment, and Fabrication of Micro-Grooved Heat Pipes: An Update

2007 ◽  
Vol 60 (3) ◽  
pp. 107-119 ◽  
Author(s):  
Balram Suman
Keyword(s):  
Heat Pipe ◽  
Heat Pipes ◽  
Sensitivity Analyses ◽  
Experimental Investigations ◽  
Future Research ◽  
Detailed Knowledge ◽  
Additional Information ◽  
Modeling Experiment ◽  
Transient Models ◽  
Work Done

The review presents an update of the work done in the micro heat pipe research and development, with an aim to give updated detailed knowledge to individuals new to the field, as well as to those already working in this area. Presented here is a summary of the recent advances in these devices occurring since the early 1990s. The following review describes the historical development of these devices, along with a review of the steady state and the transient models, sensitivity analyses, recent experimental investigations and fabrication techniques. The critical heat input, dryout length, fill charge, various heat pipe limitations and design have also been discussed in brief. Finally, future research and areas in which additional information is required are identified and delineated. This article has 204 references.

scholarly journals Information encountering re-encountered

2020 ◽  
Vol 76 (3) ◽  
pp. 731-751 ◽  
Author(s):  
Sanda Erdelez ◽  
Stephann Makri
Keyword(s):  
Information Acquisition ◽  
Theoretical Description ◽  
Theoretical Development ◽  
Future Research ◽  
Information Discovery ◽  
Research Attention ◽  
Content Type ◽  
Additional Information ◽  
Human Information Behavior ◽  
Information Encountering

PurposeIn order to understand the totality, diversity and richness of human information behavior, increasing research attention has been paid to examining serendipity in the context of information acquisition. However, several issues have arisen as this research subfield has tried to find its feet; we have used different, inconsistent terminology to define this phenomenon (e.g. information encountering, accidental information discovery, incidental information acquisition), the scope of the phenomenon has not been clearly defined and its nature was not fully understood or fleshed-out.Design/methodology/approachIn this paper, information encountering (IE) was proposed as the preferred term for serendipity in the context of information acquisition.FindingsA reconceptualized definition and scope of IE was presented, a temporal model of IE and a refined model of IE that integrates the IE process with contextual factors and extends previous models of IE to include additional information acquisition activities pre- and postencounter.Originality/valueBy providing a more precise definition, clearer scope and richer theoretical description of the nature of IE, there was hope to make the phenomenon of serendipity in the context of information acquisition more accessible, encouraging future research consistency and thereby promoting deeper, more unified theoretical development.

Sign in / Sign up

Export Citation Format

Share Document