155,713 Hits in 4.9 sec

On Overfitting and Asymptotic Bias in Batch Reinforcement Learning with Partial Observability (Extended Abstract)

Vincent Francois-Lavet, Guillaume Rabusseau, Joelle Pineau, Damien Ernst, Raphael Fonteneau
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
When an agent has limited information on its environment, the suboptimality of an RL algorithm can be decomposed into the sum of two terms: a term related to an asymptotic bias (suboptimality with unlimited  ...  data) and a term due to overfitting (additional suboptimality due to limited data).  ...  A potential direction is based on nonnegative matrix factorization (NMF) [Lee and Seung, 1999] .  ... 
doi:10.24963/ijcai.2020/695 dblp:conf/ijcai/0001Z20 fatcat:yx2wihhuobgmjjh4aevkbr33g4

Correlation Function in Deep Redshift Space as a Cosmological Probe

Takahiko Matsubara
2004 Astrophysical Journal  
The survey area on the sky should be smaller at deep redshifts than at shallow redshifts to keep the number density as dense as possible.  ...  Assuming future redshift surveys of z 3 which are within reach of the present-day technology, achievable error bounds on cosmological parameters are estimated by calculating the Fisher matrix.  ...  The bias is the most uncertain factor in galaxy redshift surveys.  ... 
doi:10.1086/424561 fatcat:qla3jwbw2jdsdmeuqkllej7kji

Deep Galaxy: Classification of Galaxies based on Deep Convolutional Neural Networks [article]

Nour Eldeen M. Khalifa, Mohamed Hamed N. Taha, Aboul Ella Hassanien, I. M. Selim
2017 arXiv   pre-print
In this paper, a deep convolutional neural network architecture for galaxies classification is presented.  ...  The proposed deep galaxies architecture consists of 8 layers, one main convolutional layer for features extraction with 96 filters, followed by two principles fully connected layers for classification.  ...  % 91.64 10 2016 Galaxy Image Classification using Non- Negative Matrix Factorization Used a method based on Non- Negative matrix factorization for images of galax- ies in the Zsolt frei  ... 
arXiv:1709.02245v1 fatcat:tsxso6536fdmtkpwi5s4sjjzym

Deep matrix factorizations [article]

Pierre De Handschutter, Nicolas Gillis, Xavier Siebert
2020 arXiv   pre-print
Recently, deep matrix factorization (deep MF) was introduced to deal with the extraction of several layers of features and has been shown to reach outstanding performances on unsupervised tasks.  ...  Constrained low-rank matrix approximations have been known for decades as powerful linear dimensionality reduction techniques to be able to extract the information contained in large data sets in a relevant  ...  Most deep MF models assume the non-negativity of several factors of the decomposition and therefore extend some NMF ideas. This paper serves as a survey on the recent literature on deep MF.  ... 
arXiv:2010.00380v2 fatcat:5d6zleu6w5gh7nmduv2zxu7ep4

Acoustic studies of spatial gradients in the Baltic: Implications for fish distribution

A Orlowski
1999 ICES Journal of Marine Science  
Both methods are designed for studying the spatial structure of abiotic and biotic factors. Firstly, a method for estimation of vertical gradients in environmental factors is defined.  ...  This paper presents two approaches to treating acoustic, biologic, and hydrographic data, collected during surveys of significantly large spatial units of the marine ecosystem.  ...  In matrix macrosounding the whole area surveyed is divided (see Figure 5 ) into elementary units (rectangles) forming the matrix of columns and rows.  ... 
doi:10.1006/jmsc.1999.0484 fatcat:3dnwnmexovaitejpfeqfvg4pz4

A Survey of Model Compression and Acceleration for Deep Neural Networks [article]

Yu Cheng, Duo Wang, Pan Zhou, Tao Zhang
2020 arXiv   pre-print
Therefore, a natural thought is to perform model compression and acceleration in deep networks without significantly decreasing the model performance.  ...  After that, we survey the evaluation matrices, the main datasets used for evaluating the model performance, and recent benchmark efforts.  ...  and used deep networks at test Baseline Models Representative Works Alexnet [1] structural matrix [30] , [31] , [33] low-rank factorization [41] Network in network [76] low-rank factorization  ... 
arXiv:1710.09282v9 fatcat:frwedew2gfe3rjif5ds75jqay4

A Deep Learning Approach to Analyze Airline Customer Propensities: The Case of South Korea

So-Hyun Park, Mi-Yeon Kim, Yeon-Ji Kim, Young-Ho Park
2022 Applied Sciences  
However, their application to the airline industry has traditionally focused solely on surveys; hence, there is a lack of attention paid to deep learning techniques based on survey results.  ...  For this, we applied deep learning techniques to the survey data collected from the users who have used mostly Korean airplanes.  ...  Therefore, there is a limitation in that the derived results are limited to the following linear format: factor A has an effect of value C on factor B.  ... 
doi:10.3390/app12041916 fatcat:nw7y2alafzfbfk65de753wc3n4

A Survey on Concept Factorization: From Shallow to Deep Representation Learning [article]

Zhao Zhang, Yan Zhang, Mingliang Xu, Li Zhang, Yi Yang, Shuicheng Yan
2021 arXiv   pre-print
As a relatively new paradigm for representation learning, Concept Factorization (CF) has attracted a great deal of interests in the areas of machine learning and data mining for over a decade.  ...  In this paper, we therefore survey the recent advances on CF methodologies and the potential benchmarks by categorizing and summarizing the current methods.  ...  Thus, in this survey paper, we aim to present a comprehensive survey on the concept factorization algorithms.  ... 
arXiv:2007.15840v3 fatcat:ahun2mogmfapxe4mqhqlsakyku

How to Build a Graph-Based Deep Learning Architecture in Traffic Domain: A Survey [article]

Jiexia Ye, Juanjuan Zhao, Kejiang Ye, Chengzhong Xu
2020 arXiv   pre-print
To provide a comprehensive and clear picture of such emerging trend, this survey carefully examines various graph-based deep learning architectures in many traffic applications.  ...  We first give guidelines to formulate a traffic problem based on graph and construct graphs from various kinds of traffic datasets.  ...  To our best knowledge, we are the first to provide a comprehensive survey on graph-based deep learning works in traffic domain.  ... 
arXiv:2005.11691v6 fatcat:uiso5cg6cvhvnfmtisvuxapfqi

The Structure of the Nucleon

Shu-Kun Lin, Wolfram Weise
2001 Molecules  
From the contents: electromagnetic structure of the nucleon, weak probes of nucleon structure, deep inelastic lepton scattering on the nucleon; elements of QCD, aspects of non-perturbative QCD, Chiral  ...  Weise present a unique report on the extensive empirical studies, theoretical foundations and the different models of the nucleon.  ...  Myon Capture Near Threshold Pion Electroproduction Neutral Current Interactions DEEP-INELASTIC LEPTON SCATTERING ON THE NUCLEON The Parton Model Scaling Violations Sum Rules Neutrino Deep-Inelastic  ... 
doi:10.3390/61201041 fatcat:whhbttph5bfpxi5mgwxnzjem7u

Precision cosmology with a combination of wide and deep Sunyaev-Zel'dovich cluster surveys

Satej Khedekar, Subhabrata Majumdar, Sudeep Das
2010 Physical Review D  
We show that by dividing up a cluster survey into a wide and a deep survey, one can essentially recover the cosmological information that would be diluted in a single survey of the same duration due to  ...  A variable depth survey with a few thousand clusters is as effective at constraining cosmological parameters as a single depth survey with a much larger cluster sample.  ...  We find that a two component survey improves the constraints on w by a factor of 4 for the case when cluster evolution is unknown.  ... 
doi:10.1103/physrevd.82.041301 fatcat:lq7tmeqcwvbjhhxuozjxzy663u

Graph embedding techniques, applications, and performance: A survey

Palash Goyal, Emilio Ferrara
2018 Knowledge-Based Systems  
We then present three categories of approaches based on factorization methods, random walks, and deep learning, with examples of representative algorithms in each category and analysis of their performance  ...  In this survey, we provide a comprehensive and structured analysis of various graph embedding techniques proposed in the literature.  ...  Approaches to factorize the representative matrix vary based on the matrix properties.  ... 
doi:10.1016/j.knosys.2018.03.022 fatcat:wpud5byxxndllmhqdnhkljvcga

Traditional Dimensionality Reduction Techniques Using Deep Learning

2019 International journal of recent technology and engineering  
This paper gives a review about the traditional methods used in Machine algorithm for reducing the dimension and proposes a view, how deep learning can be used for dimensionality reduction.  ...  The main aim of the Dimensionality Reduction algorithms such as Principal Component Analysis (PCA), Random Projection (RP) and Non Negative Matrix Factorization (NMF) is used to decrease the inappropriate  ...  [44] Nonnegative Matrix Factorization (NMF) is one of the most promising behaviors for dimension reduction in unsupervised learning, and is extended from two-matrix to triple-matrix factorization.  ... 
doi:10.35940/ijrte.c6110.098319 fatcat:ignipcp2hzd53jysfenx4gytni

Determination of z ∼ 0.8 neutral hydrogen fluctuations using the 21 cm intensity mapping autocorrelation

E. R. Switzer, K. W. Masui, K. Bandura, L.-M. Calin, T.-C. Chang, X.-L. Chen, Y.-C. Li, Y.-W. Liao, A. Natarajan, U.-L. Pen, J. B. Peterson, J. R. Shaw (+1 others)
2013 Monthly Notices of the Royal Astronomical Society Letters  
Our previous measurements of the cross-correlation of 21 cm intensity and the WiggleZ galaxy survey provide a lower bound.  ...  Even in the presence of residual foregrounds, the auto-power can still be interpreted as an upper bound on the 21 cm signal.  ...  Computations were performed on the GPC supercomputer at the SciNet HPC Consortium. SciNet is funded by the Canada Foundation for Innovation.  ... 
doi:10.1093/mnrasl/slt074 fatcat:cbgqpgw5s5earfxogmba64acqe

Constraining the evolution of dark energy with a combination of galaxy cluster observables

Sheng Wang, Justin Khoury, Zoltán Haiman, Morgan May
2004 Physical Review D  
For the X-ray and SZ surveys, constraints on dark energy parameters are improved by a factor of two by combining the cluster data with cosmic microwave background (CMB) anisotropy measurements by Planck  ...  , but degrade by a factor of two if the survey is required to solve simultaneously for cosmological and cluster structure evolution parameters.  ...  Acknowledgments We thank Joseph Hennawi, Wayne Hu, Joseph Mohr, and the members of the DUO team for insightful discussions, Jochen Weller for useful comments on the manuscript, Gil Holder and Joseph Mohr  ... 
doi:10.1103/physrevd.70.123008 fatcat:duoyxvzi25bariyrv7k65tcuqq
« Previous Showing results 1 — 15 out of 155,713 results