Filters








280,310 Hits in 8.6 sec

How to define a rejection class based on model learning?

Sarah Laroui, Xavier Descombes, Aurelia Vernay, Florent Villiers, Francois Villalba, Eric Debreuve
2021 2020 25th International Conference on Pattern Recognition (ICPR)  
It is then crucial to define a classifier which has the ability to reject a sample, i.e., to classify it into a rejection class that has not been yet defined.  ...  For that, we rely on modeling each class as the combination of a probability density function (PDF) and a threshold that is computed with respect to the other classes.  ...  Conclusion and perspectives We proposed a method of rejection class based on model learning in a supervised context.  ... 
doi:10.1109/icpr48806.2021.9412381 fatcat:sd5bjtt2afhclh6o2sm2a4dr3q

On discriminative learning of prediction uncertainty

Vojtech Franc, Daniel Prusa
2019 International Conference on Machine Learning  
The classical cost based model of an optimal classifier with a reject option requires the cost of rejection to be defined explicitly.  ...  An optimal rejection strategy for both models is based on thresholding the conditional risk defined by posterior probabilities which are usually unavailable.  ...  We show that an optimal rejection strategy can be constructed based on thresholding the class conditional risk, which corresponds to the optimal strategy of the cost-based model.  ... 
dblp:conf/icml/FrancP19 fatcat:fu3noypacnaodkicx6q65xd2ay

Learning Phonological Mappings by Learning Strictly Local Functions

Jane Chandlee, Adam Jardine
2014 Proceedings of the Annual Meetings on Phonology  
; Pullum 2011; Rogers et al. 2013), and show how they can model a range of phonological processes.  ...  We provide a proof that the SLFLA learns the class of SL functions and discuss these results alongside previous studies on using OSTIA to learn phonological mappings (Gildea and Jurafsky 1996).</p>  ...  We suggest that additional functional classes can be defined that correspond to these sub-regular languages and model the non-local processes.  ... 
doi:10.3765/amp.v1i1.13 fatcat:4jgdqksv7vbstabilzmzn5rxti

Action-Reinforcement Learning Versus Rule Learning

Dale O. Stahl
2003 Social Science Research Network  
How well do various learning models predict the dynamics of the population distribution of play in a variety of games?  ...  We formulate a population Rule Learning model that nests the LBRIAE model and find that it is statistically superior.  ...  By "action-reinforcement learning," we refer to the class of dynamic models in which the objects of reinforcement are the actions available to the players in a one-shot game.  ... 
doi:10.2139/ssrn.410928 fatcat:untarwvou5brnmby4fw4laqn3a

Learning without Seeing nor Knowing: Towards Open Zero-Shot Learning [article]

Federico Marmoreo, Julio Ivan Davila Carrazco, Vittorio Murino, Jacopo Cavazza
2021 arXiv   pre-print
We achieve this by optimizing a generative process to sample unknown class embeddings as complementary to the seen and the unseen.  ...  In Generalized Zero-Shot Learning (GZSL), unseen categories (for which no visual data are available at training time) can be predicted by leveraging their class embeddings (e.g., a list of attributes describing  ...  After the initial formalization of [36] on how to learn in the open world paradigm, many approaches have proposed for letting traditional machine learning models to deal with the unknown [37, 3, 21,  ... 
arXiv:2103.12437v2 fatcat:ojwkcycngba4feqiz3l6mrixfm

THE EFFECT OF LEARNING MODEL LEARNING LEARNING WITH YOUTUBE VERSUS MEDIA FLASH CARD MEDIA AND LEARNING MOTIVATION ON LEARNING OUTCOMES IN ENGLISH

Rahesa Nalendra, Iskandar Wiyokusumo, Ibut Priono Leksono
2020 JEES: Journal of English Educational Study  
The study was conducted with a quantitative approach with a factorial design experimental model.  ...  This study aims to determine the effect of active learning model learning with Youtube media versus flash card media and students' learning motivation towards student learning outcomes in SMP N 1 Sidoarjo  ...  Learning based on active learning (active learning) allows can have a positive impact on the ability of students to solve problems in the classroom when teaching and learning activities.  ... 
doi:10.31932/jees.v3i2.717 fatcat:qfmyjbjom5hk3cazikpcjewzh4

THE EFFECT OF IMPLEMENTING PAIKEM USING STUDENT-CENTERED LEARNING, CASE-BASED LEARNING, AND COOPERATIVE LEARNING ON LEARNING EFFICIENCY AND EFFECTIVENESS

I Indrayati
2019 Erudio  
The results revealed that the project-based learning, cooperative learning, and fun learning variables had a significant effect on efficiency while the teacher-centered learning, student-centered learning  ...  Furthermore, student-centered learning, cooperative learning, effective learning, and fun learning variables had a significant effect on learning effectiveness; while the teacher-centered learning, case-based  ...  Acknowledgment The researcher would like to express gratitude to the Director of the Malang State Polytechnic and the Directorate of Higher Education (DIKTI) for funding this research.  ... 
doi:10.18551/erudio.6-1.9 fatcat:b3wdsflgtbgsxagcqvtyy7bypi

Synthetic Unknown Class Learning for Learning Unknowns [article]

Jaeyeon Jang
2021 arXiv   pre-print
This paper addresses the open set recognition (OSR) problem, where the goal is to correctly classify samples of known classes while detecting unknown samples to reject.  ...  Thus, this paper proposes a novel synthetic unknown class learning method that generates unknown-like samples while maintaining diversity between the generated samples and learns these samples.  ...  Then, the student learns only the unknown samples that are worth learning based on the loss defined in (5) .  ... 
arXiv:2111.08062v1 fatcat:zlfmfntazbb35adjwzo2vxkbtm

Learning How to Self-Learn: Enhancing Self-Training Using Neural Reinforcement Learning [article]

Chenhua Chen, Yue Zhang
2018 arXiv   pre-print
Based on neural network representation of sentences, our model automatically learns an optimal policy for instance selection.  ...  To address these challenges, we propose a deep reinforcement learning method to learn the self-training strategy automatically.  ...  Learning how to self-train In order to learn a self-training function automatically, we design a deep reinforcement learning neural network to optimize the function parameters based on feedback from a  ... 
arXiv:1804.05734v1 fatcat:xfik64r4i5fcbgn5uw4yhcnqnq

Unsupervised Learning via Meta-Learning [article]

Kyle Hsu and Sergey Levine and Chelsea Finn
2019 arXiv   pre-print
Many prior unsupervised learning works aim to do so by developing proxy objectives based on reconstruction, disentanglement, prediction, and other metrics.  ...  Surprisingly, we find that, when integrated with meta-learning, relatively simple task construction mechanisms, such as clustering embeddings, lead to good performance on a variety of downstream, human-specified  ...  ACKNOWLEDGMENTS We thank Kelvin Xu, Richard Zhang, Brian Cheung, Ben Poole, Aäron van den Oord, Luke Metz, Siddharth Reddy, and the anonymous reviewers for feedback on an early draft of this paper.  ... 
arXiv:1810.02334v6 fatcat:b5xxbul44zegjac3mjvspqnlka

Designing Learning Analytics for Effective Learning Through Problem-Based Learning

2020 Journal of Education and Practice  
This is an educational pedagogy and practice paper to design and develop Learning Analytics using Problem-Based Learning (PBL).  ...  The PBL is adopted as it demonstrates a learning design that integrates industry experience with practice-based learning within the classroom.  ...  Identify the data required to develop a predictive model to admit or reject students' application, which kind of predictive models would be developed and finally, how do you measure the performance of  ... 
doi:10.7176/jep/11-18-03 fatcat:m5dhedffvbgv3mmoxiwzjhpiwm

Robust Classification with Convolutional Prototype Learning [article]

Hong-Ming Yang, Xu-Yao Zhang, Fei Yin, Cheng-Lin Liu
2018 arXiv   pre-print
Moreover, a prototype loss (PL) is proposed as a regularization to improve the intra-class compactness of the feature representation, which can be viewed as a generative model based on the Gaussian assumption  ...  In this paper, we argue that the lack of robustness for CNN is caused by the softmax layer, which is a totally discriminative model and based on the assumption of closed world (i.e., with a fixed number  ...  class-incremental learning process, we did not re-train any part of the model, and due to the advantage of prototype-based decision making, we can directly add a new prototype to represent the new class  ... 
arXiv:1805.03438v1 fatcat:vkmt2c4vnrgdrmjmcb4vxhqycu

The influence of problem based learning towards social science learning outcomes viewed from learning interest

Berti Dyah Permatasari, Gunarhadi Gunarhadi, Riyadi Riyadi
2019 International Journal of Evaluation and Research in Education (IJERE)  
The aim of this study is to determine the influences of Problem Based Learning and learning interest at improving the cognitive learning outcomes in social science of fourth-grade elementary school students  ...  The students from experiment group are given the application of Problem Based Learning, while the students in control group are given the application of Direct Instruction.  ...  This is reinforced by the findings that classes that are subject to PBL models have a higher problem-solving ability than classes subjected to conventional models [13] .  ... 
doi:10.11591/ijere.v8i1.15594 fatcat:4wz33l4cybhdlap6i5usutc3s4

Robust Classification with Convolutional Prototype Learning

Hong-Ming Yang, Xu-Yao Zhang, Fei Yin, Cheng-Lin Liu
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
Moreover, a prototype loss (PL) is proposed as a regularization to improve the intra-class compactness of the feature representation, which can be viewed as a generative model based on the Gaussian assumption  ...  In this paper, we argue that the lack of robustness for CNN is caused by the softmax layer, which is a totally discriminative model and based on the assumption of closed world (i.e., with a fixed number  ...  Most rejection strategies are based on the probabilities (confidences) produced by the softmax layer of CNN model.  ... 
doi:10.1109/cvpr.2018.00366 dblp:conf/cvpr/YangZYL18 fatcat:uyaggmmmsvaydkgbpjapc5o5m4

Can Learning Vector Quantization be an Alternative to SVM and Deep Learning? - Recent Trends and Advanced Variants of Learning Vector Quantization for Classification Learning

Thomas Villmann, Andrea Bohnsack, Marika Kaden
2017 Journal of Artificial Intelligence and Soft Computing Research  
Learning vector quantization (LVQ) is one of the most powerful approaches for prototype based classification of vector data, intuitively introduced by Kohonen.  ...  Thus, the intention of the paper is to provide a comprehensive overview of the stateof- the-art serving as a starting point to search for an appropriate LVQ variant in case of a given specific classification  ...  An attempt to incorporate reject options into the SVM model was provided in [105] , which is based on geometric considerations.  ... 
doi:10.1515/jaiscr-2017-0005 fatcat:a3bnek56tfgy5e2umpb62d4vae
« Previous Showing results 1 — 15 out of 280,310 results