Filters








7 Hits in 5.5 sec

Sequential Changepoint Detection in Neural Networks with Checkpoints [article]

Michalis K. Titsias, Jakub Sygnowski, Yutian Chen
2020 arXiv   pre-print
We introduce a framework for online changepoint detection and simultaneous model learning which is applicable to highly parametrized models, such as deep neural networks.  ...  We demonstrate the efficiency of our method in challenging continual learning applications with unknown task changepoints, and show improved performance compared to online Bayesian changepoint detection  ...  Section 3 develops our framework for changepoint detection using checkpoints and Section 4 considers applications to continual learning with neural networks.  ... 
arXiv:2010.03053v1 fatcat:wdzbjqnysfeqjm7dkspvducbjy

2019 Index IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems Vol. 38

2019 IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems  
., +, TCAD April 2019 755-766 D TCAD June 2019 1095-1108 Data mining Changepoint-Based Anomaly Detection for Prognostic Diagnosis in a Core Router System.  ...  Xie, Y., +, and Weight Quantization in Spiking Neural Networks for Energy-Efficient Recognition.  ... 
doi:10.1109/tcad.2020.2964359 fatcat:qjr6i73tkrgnrkkmtjexbxberm

On Neural Differential Equations [article]

Patrick Kidger
2022 arXiv   pre-print
In particular, neural differential equations (NDEs) demonstrate that neural networks and differential equation are two sides of the same coin.  ...  Many popular neural network architectures, such as residual networks and recurrent networks, are discretisations.  ...  (t)) ∈ R, (2.3) in which an existing theoretical model is augmented with a neural network correction term.  ... 
arXiv:2202.02435v1 fatcat:vglknmvlgfeddoe2cxohubauxm

Adaptation Strategies for Automated Machine Learning on Evolving Data [article]

Bilge Celik, Joaquin Vanschoren
2021 arXiv   pre-print
These are evaluated empirically on real-world and synthetic data streams with different types of concept drift.  ...  We do this for a variety of AutoML approaches for building machine learning pipelines, including those that leverage Bayesian optimization, genetic programming, and random search with automated stacking  ...  AS-5 Train once -When drift is detected, use the checkpoint option to train the base models incrementally on new data with the same hyperparameters.  ... 
arXiv:2006.06480v2 fatcat:rw3ggc4ozbfmxeafi26lmzw7ju

International Research Conference on Smart Computing and Systems Engineering SCSE 2020 Proceedings [Full Conference Proceedings]

2020 2020 International Research Conference on Smart Computing and Systems Engineering (SCSE)  
ACKNOWLEDGMENT The authors would like to extend their heartfelt gratitude and acknowledgment to all the language experts, Sinhala teachers from different schools and especially, the Sinhala Department in  ...  The above results show the neural network-based methods perform better than the kernel-based methods in anomaly detection in cloud network data.  ...  The above results show that the neural network-based methods perform better than the kernel-based methods in anomaly detection in cloud network data sets we used in our experiments. V.  ... 
doi:10.1109/scse49731.2020.9313027 fatcat:gjk5az2mprgvrpallwh6uhvlfi

Substrate spectrum of PPM1D in the cellular response to

Justus F. Gräf, Ivan Mikicic, Xiaofei Ping, Claudia Scalera, Katharina Mayr, Lukas S. Stelzl, Petra Beli, Sebastian A. Wagner
2022 iScience  
This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article.  ...  Acknowledgments Work in the Beli lab is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) -Project-ID 393547839 -SFB 1361, Project-ID BE 5342/2-1 -FOR 2800 and GRK2526/1 -  ...  We thank Anja Freiwald, Jimmy Chen and Amitkumar Fulzele for assistance with mass spectrometry analysis and Andrea Voigt for technical support.  ... 
doi:10.1016/j.isci.2022.104892 fatcat:7e5ezagzozcxdcmtr3dvupzgdq

Dagstuhl Reports, Volume 11, Issue 7, August 2021, Complete Issue [article]

2021
The overall system is often determined by an interplay of many model aspects (topology, temporal ordering, type of dynamics) and we need to detect which of these interactions aspects are qualitatively  ...  We have investigated the interplay of temporal and multi-way interactions in [3] and found effects, that differ from their projections.  ...  In subsequent studies [6] , Madireddy used I/O changepoint detection to identify noticeable changes in the performance of production file systems and used transfer learning to update I/O performance prediction  ... 
doi:10.4230/dagrep.11.7 fatcat:4b73kynisffo7payzzghe5rfiq