Filters








64 Hits in 2.6 sec

Squirrel: A Switching Hyperparameter Optimizer [article]

Noor Awad, Gresa Shala, Difan Deng, Neeratyoy Mallik, Matthias Feurer, Katharina Eggensperger, Andre' Biedenkapp, Diederick Vermetten, Hao Wang, Carola Doerr, Marius Lindauer, Frank Hutter
2020 arXiv   pre-print
Motivated by the fact that different optimizers work well on different problems, our approach switches between different optimizers.  ...  Squirrel, or here, short, Squirrel.  ...  In our Squirrel framework, we switched between the following components: 1.  ... 
arXiv:2012.08180v2 fatcat:bbokv2bcbza45hsdqt2vikpdeu

ML models for real-time hybrid systems [chapter]

Manuel I. Capel
2021 XLII JORNADAS DE AUTOMÁTICA : LIBRO DE ACTAS  
A correct system design can be systematically obtained from a specification model of a real-time system that integrates hybrid measurements In a realistic industrial environment, this has been carried  ...  An AC motor controller which must be able to maintain a constant air flow through a filter is one of these systems.  ...  basis for obtaining a great improvement over other approaches currently more used for example, obtaining hyperparameters using Bayesian optimization (HPO) [1] .  ... 
doi:10.17979/spudc.9788497498043.752 fatcat:jrnpw6nlubegfgaubnmnsrcbme

Condition Monitoring and Feature Extraction of Stator Current Phasors for Enhanced Fault Diagnosis in AC Drive

Jyothi R, Assistant Professor, Department of Electrical and Electronics Engineering, RV College of Engineering, Bangalore (Karnataka), India., Tejas Holla, Umesh NS, K Uma Rao, Jayapal R, Department of Electrical and Electronics Engineering at RV College of Engineering, Bangalore (Karnataka), India., Assistant Professor, Department of Electrical and Electronics Engineering, RV College of Engineering, Bangalore (Karnataka), India., Professor, Department of Electrical and Electronics Engineering, R.V. College of Engineering, Bangalore (Karnataka), India., Assistant Professor, Department of Electrical and Electronics Engineering, RV College of Engineering, Bangalore (Karnataka), India.
2021 International Journal of Engineering and Advanced Technology  
Variable Frequency Drives (VFD) comprises of Voltage source inverter and a three phase squirrel cage induction motor.  ...  Process industries have seen a paradigm shift from manual control to automatic control.  ...  A diode bridge rectifier is used to obtain dc voltage. The dc link is required to maintain the dc voltage supplied to the inverter. IGBT switches are used in the VSI.  ... 
doi:10.35940/ijeat.a3173.1011121 fatcat:agqxza3vifgjjplkr7dggcsgyi

A Video Captioning Method Based on Multi-Representation Switching for Sustainable Computing

Heechan Kim, Soowon Lee
2021 Sustainability  
To reflect this characteristic explicitly using a deep learning model, we propose a multi-representation switching method.  ...  The proposed multi-representation switching method makes it possible for the three components to extract important information for a given video and description pair efficiently.  ...  The proposed method with the optimal single parameters, θ * s , is expressed as "Multi-Representation Switching with single parameter (MRS-s)" and the methods with the optimal ensemble parameters with  ... 
doi:10.3390/su13042250 fatcat:y5lvtjmdrjbtfigd3ooiu4rnne

Optimization-Induced Graph Implicit Nonlinear Diffusion [article]

Qi Chen, Yifei Wang, Yisen Wang, Jiansheng Yang, Zhouchen Lin
2022 arXiv   pre-print
As a result, our GIND obtains significant improvements on both node-level and graph-level tasks.  ...  More interestingly, we can induce new structural variants by modifying the corresponding optimization objective.  ...  Switching the direction of an edge k = (i, j) is equivalent to multiplying a matrix E k to the left of G, where E k is obtained by switching the (k, k)-th element of an identity matrix to −1.  ... 
arXiv:2206.14418v1 fatcat:mnyu4qqpwbgllpzrzjkzcnrqgq

Deep convolutional neural networks for automatic coil pitches detection systems in induction motors

Hidir Selcuk Nogay
2021 Journal of Electrical Engineering  
A system that can detect the stator coil pitch according to the stator current behavior while the motor is running can significantly shorten this improvement process.  ...  For this purpose, in this study, a deep convolutional neural network (DCNN) model that can automatically estimate IM stator coil pitch angle with an accuracy rate of 100% is designed and applied.  ...  In one study, it was found that an IM performs higher when fed with variable PWM switching frequency by shortening the coil pitch in a double layer stator winding [14] .  ... 
doi:10.2478/jee-2021-0027 fatcat:mgolomzinbeahpgdjxnmhjhy44

A Biased Graph Neural Network Sampler with Near-Optimal Regret [article]

Qingru Zhang, David Wipf, Quan Gan, Le Song
2021 arXiv   pre-print
And unlike prior bandit-GNN use cases, the resulting policy leads to near-optimal regret while accounting for the GNN training dynamics introduced by SGD.  ...  To address these issues, we build upon existing work and treat GNN neighbor sampling as a multi-armed bandit problem but with a newly-designed reward function that introduces some degree of bias designed  ...  GCN GAT Algorithm γ η ∆T γ η ∆T Thanos 0.4 0.01 1000 0.4 0.01 1000 BanditSampler 0.4 0.01 N/A 0.4 0.01 N/A Table 5 : 5 The detailed sampling hyperparameters for Squirrel.  ... 
arXiv:2103.01089v3 fatcat:6qfdpolqp5ag7bt7qvdm3kufty

Thermal Monitoring of Electric Motors: State-of-the-Art Review and Future Challenges

Oliver Wallscheid
2021 IEEE Open Journal of Industry Applications  
In order to optimize the dynamic performance limits of a motor during online operation, important motor temperatures must be known in real time.  ...  In recent years, many promising contributions have been made to this field leading to a vast literature basis.  ...  For many ML models, therefore, a hierarchical optimization task results on two levels [114] : On the top level, suitable hyperparameters of the ML model class have to be found and on the lower level,  ... 
doi:10.1109/ojia.2021.3091870 fatcat:gyp462sx4bgu3lb554eowwuhyq

An End-to-End: Real-Time Solution for Condition Monitoring of Wind Turbine Generators

Adrian Stetco, Juan Melecio Ramirez, Anees Mohammed, Siniša Djurović, Goran Nenadic, John Keane
2020 Energies  
We show how to create, deploy and run the collection of proposed models in a real-time setting using a laptop connected to a test rig via a data acquisition card.  ...  Based on a sampling rate of 5 kHz, predictions are stored in an efficient time series database and monitored using a dynamic visualization framework.  ...  From a Three-Phase Squirrel Cage Induction Motor, they obtained 2600 of healthy and 2600 of faulty bearing samples.  ... 
doi:10.3390/en13184817 fatcat:m6jmsckszzdevbdcwooiqegf3y

Mini Cheetah, the Falling Cat: A Case Study in Machine Learning and Trajectory Optimization for Robot Acrobatics [article]

Vince Kurtz, He Li, Patrick M. Wensing, Hai Lin
2022 arXiv   pre-print
Specifically, we explore how trajectory optimization and machine learning can work together to enable highly dynamic bioinspired behaviors.  ...  We find that a reflex approach, in which a neural network learns entire state trajectories, outperforms a policy approach, in which a neural network learns a mapping from states to control inputs.  ...  Once we are ready to drop the robot, a switch on the remote controller indicates to the robot that it should start trying to detect a fall.  ... 
arXiv:2109.04424v2 fatcat:wp2lmeklqfaxbnks3utp45s4le

Prediction of wind farm reactive power fast variations by adaptive one-dimensional convolutional neural network

Haidar Samet, Saeedeh Ketabipour, Shahabodin Afrasiabi, Mousa Afrasiabi, Mohammad Mohammadi
2021 Computers & electrical engineering  
Link to publication General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of  ...  Additionally, to set the hyperparameters of the designed adaptive deep network, the root mean square propagation (RMSProp) has been adopted from [23] to find the optimal values for hyperparameters.  ...  Despite these advantages, SVC struggles with a major challenge caused by switching limitations, which leads to a time delay (about half-cycle).  ... 
doi:10.1016/j.compeleceng.2021.107480 fatcat:xqnjszr22vh4hpfcunhb5wxccq

Towards Neural Programming Interfaces [article]

Zachary C. Brown, Nathaniel Robinson, David Wingate, Nancy Fulda
2021 arXiv   pre-print
of programs by altering hyperparameters.  ...  In this new paradigm, a specialized neural network (called a Neural Programming Interface or NPI) learns to interface with a pretrained language model by manipulating the hidden activations of the pretrained  ...  Note to train NPI on the avoidance task we simply switch the objective in Equation (5) .  ... 
arXiv:2012.05983v2 fatcat:4tpggy33uzdu3myvq2twly3quy

Increasing Electric Vehicles Reliability by Non-Invasive Diagnosis of Motor Winding Faults

Konrad Górny, Piotr Kuwałek, Wojciech Pietrowski
2021 Energies  
Phase current waveforms obtained from a real object as well as simulation results based on the field-circuit model of an induction motor were used as a diagnostic signal in the research.  ...  The article proposes a proprietary approach to the diagnosis of induction motors allowing increasing the reliability of electric vehicles.  ...  In the process of hyperparameter optimization, a nested five-fold cross-validation was used to avoid a "skewed" assessment of a concrete model with specific hyperparameter values (the model would be evaluated  ... 
doi:10.3390/en14092510 fatcat:iiqjwbo2ibapdicoky5boqtxbe

Image Shape Manipulation from a Single Augmented Training Sample [article]

Yael Vinker, Eliahu Horwitz, Nir Zabari, Yedid Hoshen
2021 arXiv   pre-print
In this paper, we present DeepSIM, a generative model for conditional image manipulation based on a single image.  ...  Our network learns to map between a primitive representation of the image to the image itself.  ...  Optimization: During training, we sample random TPS warps.  ... 
arXiv:2109.06151v3 fatcat:pvc3q3siqjgxbltfwno4o2geuu

Efficient context-dependent model building based on clustering posterior distributions for non-coding sequences

Guy Baele, Yves Van de Peer, Stijn Vansteelandt
2009 BMC Evolutionary Biology  
The selected model's fit is evaluated using a Bayes Factor, which we calculate via model-switch thermodynamic integration.  ...  To obtain an optimal balance between the number of parameters in a context-dependent model and the performance in terms of model fit, we have designed two parameter-reduction strategies and we have shown  ...  hyperparameter in the prior for t b in T.  ... 
doi:10.1186/1471-2148-9-87 pmid:19405957 pmcid:PMC2695821 fatcat:zjcfkg4f6rbl7kdl3hg373ljze
« Previous Showing results 1 — 15 out of 64 results