64,302 Hits in 11.8 sec

s-DRN: Stabilized Developmental Resonance Network [article]

In-Ug Yoon, Ue-Hwan Kim, Jong-Hwan
2020 arXiv   pre-print
Online incremental clustering of sequentially incoming data without prior knowledge suffers from changing cluster numbers and tends to fall into local extrema according to given data order.  ...  The comparative studies demonstrate the proposed s-DRN outperforms baselines in terms of stability and accuracy.  ...  His current research interests include structural health monitoring using robotics, artificial intelligence, simultaneous localization and mapping, robot navigation, machine learning, deep learning, and  ... 
arXiv:1912.08541v2 fatcat:zhd22p6wbne5naqcjxj42e5bla

A Defense of (s)crappy Robots [article]

Ryan Jenkins, Mairéad Hurley, Eva Durall, Sebastian Martin
2020 Zenodo  
As makerspaces and STEAM learning environments have become more and more common, many commercial kits for robotics, electric circuits and digital technology have been created for schools and individual  ...  We are inspired by the "shitty robots" made by Simone Giertz and the Hebocon robot sumo contest from Japan to create workshops where learners need to embrace frustration, celebrate moments where things  ...  Scholarship Code: 531) Acknowledgment This research is funded by SSHRC grant #435-2017-0367, with ethics clearance REB #17-088.  ... 
doi:10.5281/zenodo.5501878 fatcat:snjtztj37fbbpeszscwhk63uj4

S-R2F2U-Net: A single-stage model for teeth segmentation [article]

Mrinal Kanti Dhar, Zeyun Yu
2022 arXiv   pre-print
Particularly, S-R2F2U-Net outperforms state-of-the-art models in terms of accuracy and dice score. A hybrid loss function combining the cross-entropy loss and dice loss is used to train the model.  ...  S-R2F2U-Net achieves 97.31% of accuracy and 93.26% of dice score, showing superiority over the state-of-the-art methods. Codes are available at  ...  Teeth segmentation methods can be divided into two categoriestraditional methods which incorporate prior knowledge and image features and deep learning-based methods.  ... 
arXiv:2204.02939v1 fatcat:ysvnej5wznfz7peazrasmbi5xu

Unsupervised MKL in Multi-layer Kernel Machines [article]

Akhil Meethal and Asharaf S and Sumitra S
2021 arXiv   pre-print
Kernel based Deep Learning using multi-layer kernel machines(MKMs) was proposed by Y.Cho and L.K. Saul in .  ...  We propose to use multiple kernels in each layer by taking a convex combination of many kernels following an unsupervised learning strategy.  ...  For a fair comparison with MKMs the I summarizes the results of our empirical study.  ... 
arXiv:2111.13769v1 fatcat:ahmpywdwqfc6bbl4pyjyfn5e3y

Monte Carlo Techniques for Analyzing Deep-Penetration Problems

S. N. Cramer, J. Gonnord, J. S. Hendricks
1986 Nuclear science and engineering  
A review of current methods and difficulties in Monte Carlo deep-penetration calculations is presented.  ...  Finally, current and future work in the area of computer learning and artificial intelligence is discussed in connection with Monte Carlo applications.  ...  have been, and are now under development and testing, several learning, or adaptive techniques where biasing parameters are generated and improved upon during a calculation.  ... 
doi:10.13182/nse86-a18177 fatcat:zwymx3246jcold2fyoaebumtcm

Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500

Christopher Krauss, Xuan Anh Do, Nicolas Huck
2017 European Journal of Operational Research  
Abstract In recent years, machine learning research has gained momentum: New developments in the field of deep learning allow for multiple levels of abstraction and are starting to supersede well-known  ...  This article implements and analyses the effectiveness of deep neural networks (DNN), gradientboosted-trees (GBT), random forests (RAF), and a combination (ENS) of these methods in the context of statistical  ...  Data and software Data For the empirical application, we opt for the S&P 500.  ... 
doi:10.1016/j.ejor.2016.10.031 fatcat:rhplxgkgwzcczbivo2yuoigouu

Photometric redshifts with the Multilayer Perceptron Neural Network: Application to the HDF-S and SDSS

E. Vanzella, S. Cristiani, A. Fontana, M. Nonino, S. Arnouts, E. Giallongo, A. Grazian, G. Fasano, P. Popesso, P. Saracco, S. Zaggia
2004 Astronomy and Astrophysics  
The Multilayer Perceptron (MLP) Artificial Neural Network is used to predict photometric redshifts in the HDF-S from an ultra deep multicolor catalog.  ...  Various possible approaches for the training of the neural network are explored, including the deepest and most complete spectroscopic redshift catalog currently available (the Hubble Deep Field North  ...  This work was partially supported by the ASI grants under the contract number ARS-98-226 and ARS-96-176, by the research contract of the University of Padova "The High redshift Universe: from HST and VLT  ... 
doi:10.1051/0004-6361:20040176 fatcat:alqjl76sxngw5f6o447eod3ila

Inter-temporal R&D and capital investment portfolios for the electricity industry�s low carbon future

Nidhi R. Santen, Mort D. Webster, David Popp, Ignacio P�rez-Arriaga
2017 Energy Journal  
We present a novel method for long-term planning by combining an economic model of endogenous non-linear technical change and a generation capacity planning model with key features of the electricity system  ...  JEL-Code: Q420, Q540, Q550.  ...  Education John S.  ... 
doi:10.5547/01956574.38.6.nsan fatcat:x4rnwritdvghbbojtfpqvawj6y

Actor Prioritized Experience Replay [article]

Baturay Saglam, Furkan B. Mutlu, Dogan C. Cicek, Suleyman S. Kozat
2022 arXiv   pre-print
Motivated by this, we introduce a novel experience replay sampling framework for actor-critic methods, which also regards issues with stability and recent findings behind the poor empirical performance  ...  The introduced algorithm suggests a new branch of improvements to PER and schedules effective and efficient training for both actor and critic networks.  ...  We test LA3P on standard deep reinforcement learning benchmarks in MuJoCo and Box2D and demonstrate that it substantially outperforms the competing methods and improves the state-of-the-art.  ... 
arXiv:2209.00532v1 fatcat:sccq5argqra45k6pdn73x7yqdm

Deep Learning Chromatic and Clique Numbers of Graphs [article]

Jason Van Hulse, Joshua S. Friedman
2021 arXiv   pre-print
In this work, we develop deep learning models to predict the chromatic number and maximum clique size of graphs, both of which represent classical NP-complete combinatorial optimization problems encountered  ...  The experimental results show that deep neural networks, and in particular convolutional neural networks, obtain strong performance on this problem.  ...  Domain knowledge of the problem is not always necessary as the neural network can learn directly from supervised data and examples.  ... 
arXiv:2108.01810v1 fatcat:ohcmkizi5fdh5mhsthv4i22qu4

Robust Single Image Super-Resolution via Deep Networks With Sparse Prior

Ding Liu, Zhaowen Wang, Bihan Wen, Jianchao Yang, Wei Han, Thomas S. Huang
2016 IEEE Transactions on Image Processing  
In this paper, we argue that domain expertise from the conventional sparse coding model can be combined with the key ingredients of deep learning to achieve further improved results.  ...  with models, such as deep neural networks.  ...  CONCLUSIONS We propose a new model for image SR by combining the strengths of sparse coding and deep network, and make considerable improvement over existing deep and shallow SR models both quantitatively  ... 
doi:10.1109/tip.2016.2564643 pmid:27168598 fatcat:475lsaazlnhp5colkspxjo4lje

Teachers and Professional Development: New Contexts, Modes, and Concerns in the Age of Social Media

Christine Greenhow, Arnon Hershkovitz, Alona Forkosh-Baruch, Emilia Askari, Dimitra Tsovaltzi, Christa S. C. Asterhan, Thomas Puhl, Armin Weinberger, Edith Bouton, Joseph L. Polman
2016 International Conference of the Learning Sciences  
affordances and challenges of social network sites and teachers' perceptions and experiences.  ...  The presentations included in this symposium offer a multi-faceted and international view on the topic, highlighting both opportunities and challenges.  ...  Deep learning can be obtained through Argumentative Knowledge Construction (AKC), which is the deliberate practice of elaborating learning material by constructing formally and semantically sound arguments  ... 
dblp:conf/icls/GreenhowHFATAPW16 fatcat:umzwzvlmjzdpnlukqzaaoejx6i

An elicitation instrument for operationalising GQM+Strategies (GQM+S-EI)

Kai Petersen, Cigdem Gencel, Negin Asghari, Stefanie Betz
2014 Empirical Software Engineering  
Conclusions: We conclude that GQM + S-EI can be used for accurately and completely eliciting the information needed by goal driven measurement frameworks.  ...  Objective: The research aims at providing an instrument (called GQM + S-EI), aiding practitioners to accurately elicit information needed by GQM + Strategies (capturing goals, strategies and information  ...  The work also funded partially by ELLIIT, the Strategic Area for ICT research, funded by the Swedish Government.  ... 
doi:10.1007/s10664-014-9306-z fatcat:fhbdkw5vkjdubixkiqxpdekkru

Qualitative v/s. Quantitative Research- A Summarized Review

Sharique Ahmad, Saeeda Wasim, Sumaiya Irfan, Sudarshana Gogoi, Anshika Srivastava, Zarina Farheen
2019 Journal of Evidence Based Medicine and Healthcare  
The purpose of quantitative research is to generate knowledge and create understanding about the social world.  ...  Rather than by logical and statistical procedures, qualitative researchers use multiple systems of inquiry for the study of human phenomena including biography, case study, historical analysis, discourse  ...  Research is the most widely used tool to increase and brushup the stock of knowledge about something and someone.  ... 
doi:10.18410/jebmh/2019/587 doaj:3f1be223c9f949ecb9508038ce4f4df9 fatcat:pcd4qpoui5grthjbga7hac5euy

An Empirical Study on Software Defect Prediction Using CodeBERT Model

Cong Pan, Minyan Lu, Biao Xu
2021 Applied Sciences  
We perform empirical studies using such models in cross-version and cross-project software defect prediction to investigate if using a neural language model like CodeBERT could improve prediction performance  ...  Deep learning-based software defect prediction has been popular these days. Recently, the publishing of the CodeBERT model has made it possible to perform many software engineering tasks.  ...  Transfer learning aims to improve the performance of predictive function f T (·) for the learning task T t by discovering and transferring latent knowledge from D s and T s , where D s = D t and/or T s  ... 
doi:10.3390/app11114793 fatcat:sv6sj7jgz5dgvewfx3ebcxc43u
« Previous Showing results 1 — 15 out of 64,302 results