27,333 Hits in 5.2 sec

Efficient Identification of Approximate Best Configuration of Training in Large Datasets

Silu Huang, Chi Wang, Bolin Ding, Surajit Chaudhuri
Given a set of configurations and a large dataset randomly split into training and testing set, we study how to efficiently identify the best configuration with approximately the highest testing accuracy  ...  A configuration of training refers to the combinations of feature engineering, learner, and its associated hyperparameters.  ...  Discussion and Conclusion We studied the problem of efficiently finding approximate best configuration among a given set of training configurations for a large dataset.  ... 
doi:10.1609/aaai.v33i01.33013862 fatcat:mxryrdmfv5emrar23uujlxhhma

Deep Convolutional Networks in System Identification [article]

Carl Andersson, Antônio H. Ribeiro, Koen Tiels, Niklas Wahlström, Thomas B. Schön
2019 arXiv   pre-print
In this paper, we establish connections between the deep learning and the system identification communities.  ...  We end the paper with an experimental study where we provide results on two real-world problems, the well-known Silverbox dataset and a newer dataset originating from ground vibration experiments on an  ...  This is called mini-batch gradient descent and is a crucial component for efficient training of a neural network when the dataset is large.  ... 
arXiv:1909.01730v2 fatcat:dcarmnrilfbzdaqvl6h3rhvway

Robust data-driven approach for predicting the configurational energy of high entropy alloys [article]

Jiaxin Zhang, Xianglin Liu, Sirui Bi, Junqi Yin, Guannan Zhang and Markus Eisenbach
2019 arXiv   pre-print
NbMoTaW, NbMoTaWV and NbMoTaWTi where it is demonstrated how dataset size affects the confidence we can place in statistical estimates of configurational energy when data are sparse.  ...  In this study, a robust data-driven framework based on Bayesian approaches is proposed and demonstrated on the accurate and efficient prediction of configurational energy of high entropy alloys.  ...  The RMSE mean value of m = 13 and m = 7 is reduced to approximate the best shells as dataset size increases (see Fig. 12 (a)) but the variations (standard deviation) are still substantially large.  ... 
arXiv:1908.03665v1 fatcat:yckjuoiky5fgroomnmj6f54p6e

Practical and sample efficient zero-shot HPO [article]

Fela Winkelmolen, Nikita Ivkin, H. Furkan Bozkurt, Zohar Karnin
2020 arXiv   pre-print
Current techniques for obtaining this list are computationally expensive as they rely on running training jobs on a diverse collection of datasets and a large collection of randomly drawn HPs.  ...  The first is based on a surrogate model and adaptively chooses pairs of dataset, configuration to query.  ...  For each of the above algorithms we pre-compute the miss-classification rate for a large number of classification datasets using a large number of randomly sampled HP configurations.  ... 
arXiv:2007.13382v1 fatcat:cmi6dtoj7zghhfp4tyhfgmegnq

Learning face similarities for face verification using hybrid convolutional neural networks

Fadhlan Hafizhelmi Kamaru Zaman, Juliana Johari, Ahmad Ihsan Mohd Yassin
2019 Indonesian Journal of Electrical Engineering and Computer Science  
There are 3 face pairing configurations discussed in this paper.  ...  Besides significant degradation due to images that have large variations in pose, illumination, expression, aging, and occlusions, it also suffers from large-scale ever-expanding data needed to perform  ...  Provided that the training data is large enough to model the similarity commonly found in faces, we expect the performance would comparable in the case of out-of-sample inferencing. Figure 9 .  ... 
doi:10.11591/ijeecs.v16.i3.pp1333-1342 fatcat:dp54m75475cknag7ces33kw6ue

Stable Hash Generation for Efficient Privacy-Preserving Face Identification

Daile Osorio-Roig, Christian Rathgeb, Pawel Drozdowski, Christoph Busch
2021 IEEE Transactions on Biometrics Behavior and Identity Science  
For the best configuration, the experimental evaluation carried out over closed-set and open-set settings shows the feasibility of the proposed technique for the use in large-scale facial identification  ...  More precisely, they have been utilised in identification systems performing exhaustive searches, thereby leading to degradations of the computational efficiency.  ...  Therefore, the penetration rate p of our system in terms of percent (%) for a large dataset would be p = 0.03 for the best case and p = 1.53 for the worst case.  ... 
doi:10.1109/tbiom.2021.3100639 fatcat:g6lszbksxneslggvujmnq2le54

Towards efficient automated singer identification in large music databases

Jialie Shen, Bin Cui, John Shepherd, Kian-Lee Tan
2006 Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval - SIGIR '06  
Automated singer identification is important in organising, browsing and retrieving data in large music databases.  ...  Extensive experimental results conducted on a large music database demonstrate the superiority of our method over state-of-the-art approaches.  ...  The best identification accuracy achieved on a small dataset, containing a 10 artist set, is approximately 70%.  ... 
doi:10.1145/1148170.1148184 dblp:conf/sigir/ShenCST06 fatcat:pziiybemmzb6po4au5ivrx2oii

Automatic identification and characterization of the epiretinal membrane in OCT images

Sergio Baamonde, Joaquim de Moura, Jorge Novo, Pablo Charlón, Marcos Ortega
2019 Biomedical Optics Express  
Finally, selected classifiers were trained and compared using different metrics, providing in the best configuration an accuracy of 89.35%.  ...  This work presents an automatic methodology for the identification of the ERM presence in OCT scans.  ...  Table 1 showcases the result of the classification process. The results highlighting the best configuration for each type of classifier are represented in bold.  ... 
doi:10.1364/boe.10.004018 pmid:31452992 pmcid:PMC6701536 fatcat:vpwxtdg55ffa5n72qnznwbdk5q

Probabilistic Recurrent State-Space Models [article]

Andreas Doerr, Christian Daniel, Martin Schiegg, Duy Nguyen-Tuong, Stefan Schaal, Marc Toussaint, Sebastian Trimpe
2018 arXiv   pre-print
The effectiveness of the proposed PR-SSM is evaluated on a set of real-world benchmark datasets in comparison to state-of-the-art probabilistic model learning methods.  ...  State-space models (SSMs) are a highly expressive model class for learning patterns in time series data and for system identification. Deterministic versions of SSMs (e.g.  ...  Acknowledgements This research was supported in part by National Science Foundation grants IIS-1205249, IIS-1017134, EECS-0926052, the Office of Naval Research, the Okawa Foundation, and the Max-Planck-Society  ... 
arXiv:1801.10395v2 fatcat:ysvlttbtmbejxmkpormsmmerma

A novel framework for efficient automated singer identification in large music databases

Jialie Shen, John Shepherd, Bin Cui, Kian-Lee Tan
2009 ACM Transactions on Information Systems  
An extensive experimental study on a large music database demonstrates the superiority of our method over state-of-the-art approaches in terms of effectiveness, efficiency, scalability, and robustness.  ...  One such characteristic, useful in a range of applications, is the identification of the singer in a musical piece.  ...  Keith van Rijsbergen in the Department of Computing Science at the University of Glasgow for his valuable advice.  ... 
doi:10.1145/1508850.1508856 fatcat:eyvez56zrvhinhw5vurenw4fpq

Making a Science of Model Search [article]

J. Bergstra and D. Yamins and D. D. Cox
2012 arXiv   pre-print
Many computer vision algorithms depend on a variety of parameter choices and settings that are typically hand-tuned in the course of evaluating the algorithm.  ...  Our approach yields state of the art results on three disparate computer vision problems: a face-matching verification task (LFW), a face identification task (PubFig83) and an object recognition task (  ...  Random Search: LFW and PubFig83 Random search in a large space of biologically-inspired models has been shown to be an effective approach to face verification [23] and identification [25] .  ... 
arXiv:1209.5111v1 fatcat:nxnbsjvurnhzjc6yycqh7yyrem

Sample Efficient Graph-Based Optimization with Noisy Observations [article]

Tan Nguyen, Ali Shameli, Yasin Abbasi-Yadkori, Anup Rao, Branislav Kveton
2020 arXiv   pre-print
We define a notion of convexity, and we show that a variant of best-arm identification can find a near-optimal solution after a small number of queries that is independent of the size of the graph.  ...  We show effectiveness of the greedy algorithm with restarts and the simulated annealing on problems of graph-based nearest neighbor classification as well as a web document re-ranking application.  ...  The computational complexity of this method is O(n), and is not practical in large-scale problems. An approximate nearest neighbor method returns an approximate solution in a sublinear time.  ... 
arXiv:2006.02672v1 fatcat:zoebur3sfnhvlih7zqeovevce4

Face identification with top-push constrained generalized low-rank approximation of matrices

Yuanjian Chen, Yanna Zhao, Yunlong He, Fangzhou Xu, Weikuan Jia, Jian Lian, Yuanjie Zheng
2019 IEEE Access  
The effectiveness of TFL is verified on four publicly available face datasets.  ...  To this end, we formulate the learning process under the framework of generalized low-rank approximation of matrices (GLRAM) supervised with a top-push constraint.  ...  TABLE 3 . 3 Comparison of identification accuracy. The values in parentheses denote the dimension of feature vectors for the best identification accuracy.  ... 
doi:10.1109/access.2019.2947164 fatcat:kpvcjfpferd7pejzirwmr4s3z4

Large-scale EMM identification based on geometry-constrained visual word correspondence voting

Xin Yang, Qiong Liu, Chunyuan Liao, Kwang-Ting Cheng, Andreas Girgensohn
2011 Proceedings of the 1st ACM International Conference on Multimedia Retrieval - ICMR '11  
To address the challenges posed by large datasets and variation in camera-phone-captured query images, we introduce a novel image matching scheme based on geometrically consistent correspondences.  ...  We present a large-scale Embedded Media Marker (EMM) identification system which allows users to retrieve relevant dynamic media associated with a static paper document via camera-phones.  ...  In addition, as the camera-phone may capture any part of the page, the system would need to characterize and index the entire document pages [11] , resulting in high time/memory cost for large datasets  ... 
doi:10.1145/1991996.1992031 dblp:conf/mir/YangLLCG11 fatcat:bnxbxhs6szf4ddtbcufhfm5aam

Support vector machine classification on a biased training set: Multi-jet background rejection at hadron colliders

Federico Sforza, Vittorio Lippi
2013 Nuclear Instruments and Methods in Physics Research Section A : Accelerators, Spectrometers, Detectors and Associated Equipment  
This is possible thanks to the feedback of a signal-background template fit performed on a validation sample and included both in the optimization process and in the input variable selection.  ...  The procedure is applied to a real case of interest at hadron collider experiments: the reduction and the estimate of the multi-jet background in the W→ e ν plus jets data sample collected by the CDF experiment  ...  Wolfe for the review of the manuscript.  ... 
doi:10.1016/j.nima.2013.04.046 fatcat:fnwhbvgpc5dp7nhifvvwcbou4u
« Previous Showing results 1 — 15 out of 27,333 results