Filters








12,153 Hits in 3.2 sec

Hyperparameter Transfer Across Developer Adjustments [article]

Danny Stoll, Jörg K.H. Franke, Diane Wagner, Simon Selg, Frank Hutter
2020 arXiv   pre-print
In this work, we remedy this situation and propose a new research framework: hyperparameter transfer across adjustments (HT-AA).  ...  This question poses a challenging problem, as developer adjustments can change which hyperparameter settings perform well, or even the hyperparameter search space itself.  ...  HYPERPARAMETER TRANSFER ACROSS ADJUSTMENTS After presenting a broad introduction to the topic, we now provide a detailed description of hyperparameter transfer across developer adjustments (HT-AA).  ... 
arXiv:2010.13117v1 fatcat:7kelwh3uwvceffakz3c23oqqwa

Efficacy of the Image Augmentation Method using CNN Transfer Learning in Identification of Timber Defect

Teo Hong Chun, Ummi Rabaah Hashim, Sabrina Ahmad, Lizawati Salahuddin, Ngo Hea Choon, Kasturi Kanchymalay
2022 International Journal of Advanced Computer Science and Applications  
According to the results, the ResNet50 algorithm, which has its basis in the transfer learning methodology, outclasses other CNN algorithms (ShuffleNet, AlexNet, MobileNetV2, NASNetMobile, and GoogLeNet  ...  an augmentation methodology not just addresses the issue of a limited dataset but also enhances CNN classification output by 5.78% with the support of T-test that demonstrates a significant difference across  ...  II CLASSIFICATION PERFORMANCE OF CNN ALGORITHMS ACROSS TIMBER SPECIES WITH MULTIPLE HYPERPARAMETERS SETTINGS.  ... 
doi:10.14569/ijacsa.2022.0130514 fatcat:ayz55suk75dwzdgjmmmbxp3soa

AgroAId: A Mobile App System for Visual Classification of Plant Species and Diseases Using Deep Learning and TensorFlow Lite

Mariam Reda, Rawan Suwwan, Seba Alkafri, Yara Rashed, Tamer Shanableh
2022 Informatics  
learning approach, and hyperparameter optimizations.  ...  In this paper, we develop a mobile plant care support system ("AgroAId"), which incorporates computer vision technology to classify a plant's [species–disease] combination from an input plant leaf image  ...  and adjusted base network retraining portions -developed using the transfer learning approaches outlined in [20] -to improve on the accuracies of the referenced models.  ... 
doi:10.3390/informatics9030055 fatcat:ciiu3vwopjbapkympneci4bpdy

Transfer Learning versus Multiagent Learning regarding Distributed Decision-Making in Highway Traffic

Mark Schutera, Niklas Goby, Dirk Neumann, Markus Reischl
2018 International Joint Conference on Artificial Intelligence  
In the first step traffic agents are trained by means of a deep reinforcement learning approach, being deployed inside an elitist evolutionary algorithm for hyperparameter search.  ...  The resulting architectures and training parameters are then utilized in order to either train a single autonomous traffic agent and transfer the learned weights onto a multi-agent scenario or else to  ...  Stefan Elser from Research and Development, as well as the whole Data Science Team at ZF Friedrichshafen AG, for supporting this research.  ... 
dblp:conf/ijcai/SchuteraG0R18 fatcat:kpbploe6nvaqvnwf6qzlbhgtva

Transfer Learning versus Multi-agent Learning regarding Distributed Decision-Making in Highway Traffic [article]

Mark Schutera, Niklas Goby, Dirk Neumann, Markus Reischl
2018 arXiv   pre-print
In the first step traffic agents are trained by means of a deep reinforcement learning approach, being deployed inside an elitist evolutionary algorithm for hyperparameter search.  ...  The resulting architectures and training parameters are then utilized in order to either train a single autonomous traffic agent and transfer the learned weights onto a multi-agent scenario or else to  ...  Stefan Elser from Research and Development, as well as the whole Data Science Team at ZF Friedrichshafen AG, for supporting this research.  ... 
arXiv:1810.08515v1 fatcat:5geimpxlafddrhifkfadrdom7e

Hyperparameter Optimization with Neural Network Pruning [article]

Kangil Lee, Junho Yim
2022 arXiv   pre-print
As service development using deep learning models has gradually become competitive, many developers highly demand rapid hyperparameter optimization algorithms.  ...  Since the deep learning model is highly dependent on hyperparameters, hyperparameter optimization is essential in developing deep learning model-based applications, even if it takes a long time.  ...  This trend is consistent across the three datasets.  ... 
arXiv:2205.08695v1 fatcat:zfdkblp2mnc7fpnauxxkpxn5ya

Transfer learning reveals sequence determinants of regulatory element accessibility [article]

Marco Salvatore, Marc Horlacher, Ole Winther, Robin Andersson
2022 bioRxiv   pre-print
Here, we develop ChromTransfer, a transfer learning method that uses a pre-trained, cell-type agnostic model of open chromatin regions as a basis for fine-tuning on regulatory sequences.  ...  Hyperparameters were adjusted to yield the best performance on the validation set.  ...  As a step towards establishing transfer learning for modeling the regulatory code, we here develop ChromTransfer, a transfer learning scheme for single-task modeling of the DNA sequence determinants of  ... 
doi:10.1101/2022.08.05.502903 fatcat:b5454nvkgre2pn7loa3lrm26qi

Towards empirical force fields that match experimental observables [article]

Thorben Fröhlking and Mattia Bernetti and Nicola Calonaci and Giovanni Bussi
2020 arXiv   pre-print
to provide transferable information.  ...  Validation of the selected optimized model against new data that, importantly, has not been used to adjust neither parameters nor hyperparameters, is best practice in this case as well. λ (n) E CV (1)  ... 
arXiv:2004.01630v4 fatcat:erviaidwmrbpfbl2uisrzt565e

Predicting Reaction Conditions from Limited Data through Active Transfer Learning

Eunjae Shim, Joshua Kammeraad, Ziping Xu, Ambuj Tewari, Tim Cernak, Paul Zimmerman
2022 Chemical Science  
Transfer and active learning have the potential to accelerate the development of new chemical reactions, using prior data and new experiments to inform models that adapt to the target area...  ...  Moreover, with Please do not adjust margins Please do not adjust margins subsequent collection of target reaction data, different hyperparameter choices may be favorable for the source model.  ...  These Please do not adjust margins Please do not adjust margins transfer ROC-AUC scores were compared with scores of models that were prepared appropriately.  ... 
doi:10.1039/d1sc06932b pmid:35756521 pmcid:PMC9172577 fatcat:yp2qs24i4ngnlj5op4uzs65tk4

Transfer Learning Across Patient Variations with Hidden Parameter Markov Decision Processes [article]

Taylor Killian, George Konidaris, Finale Doshi-Velez
2016 arXiv   pre-print
We update the HiP-MDP framework and extend it to more robustly develop personalized medicine strategies for HIV treatment.  ...  Hidden Parameter Markov Decision Processes (HiP-MDPs) tackle this transfer-learning problem by embedding these tasks into a low-dimensional space.  ...  In Sec. 3 we formalize the adjustments to the HiP-MDP framework and in Sec. 5 we present the performance of the adjusted HiP-MDP on developing personalized treatment strategies within HIV simulators.  ... 
arXiv:1612.00475v1 fatcat:o3tl6244gzbytod3g6boewcbs4

Width Transfer: On the (In)variance of Width Optimization [article]

Ting-Wu Chin, Diana Marculescu, Ari S. Morcos
2021 arXiv   pre-print
We show that width transfer works well across various width optimization algorithms and networks.  ...  In this work, we propose width transfer, a technique that harnesses the assumptions that the optimized widths (or channel counts) are regular across sizes and depths.  ...  Given width optimization is a fast-developing field, we study the transferability in the solution space to have a more general result.  ... 
arXiv:2104.13255v1 fatcat:fjboanmix5gmzll57m4ms6yo7u

Towards Automatic Actor-Critic Solutions to Continuous Control [article]

Jake Grigsby, Jin Yong Yoo, Yanjun Qi
2021 arXiv   pre-print
Empirically, we show that our agent outperforms well-tuned hyperparameter settings in popular benchmarks from the DeepMind Control Suite.  ...  However, these algorithms rely on a number of design tricks and hyperparameters, making their application to new domains difficult and computationally expensive.  ...  The highest-fitness members are randomly paired with the lowest-fitness members to transfer and then perturb their hyperparameter values. Network parameters and optimizer states are also transferred.  ... 
arXiv:2106.08918v2 fatcat:2hy6rrfmoffx3be5xsdgr3krjq

A flexible transfer learning framework for Bayesian optimization with convergence guarantee

Tinu Theckel Joy, Santu Rana, Sunil Gupta, Svetha Venkatesh
2019 Expert systems with applications  
We provide a mechanism to compute the noise level from the data to automatically adjust for different relatedness between the source and target tasks.  ...  We propose a novel transfer learning method for Bayesian optimization where we leverage the knowledge from an already completed source optimization task for the optimization of a target task.  ...  Addressing this, we develop a new framework for transfer learning.  ... 
doi:10.1016/j.eswa.2018.08.023 fatcat:jtoe44k2irfidltug5om5eawfe

Adversarial Learning for Zero-Shot Stance Detection on Social Media [article]

Emily Allaway, Malavika Srikanth, Kathleen McKeown
2021 arXiv   pre-print
In addition, we extend zero-shot stance detection to new topics, highlighting future directions for zero-shot transfer.  ...  In this work, we propose a new model for zero-shot stance detection on Twitter that uses adversarial learning to generalize across topics.  ...  Hyperparameters We tune the hyperparameters for our adversarial model using uniform sampling on the development set with 20 search trials.  ... 
arXiv:2105.06603v1 fatcat:qzn7hhuzp5efpintjpm743adya

Automated Detection of Greenhouse Structures Using Cascade Mask R-CNN

Haeng Yeol Oh, Muhammad Sarfraz Khan, Seung Bae Jeon, Myeong-Hun Jeong
2022 Applied Sciences  
Our proposed model is regional-based because it was optimized for the Republic of Korea via transfer learning and hyperparameter tuning, which improved the efficiency of the automated detection of greenhouse  ...  Similarly, the F1-score of the proposed Cascade Mask R-CNN model was 62.07, which outperformed those of the baseline mask R-CNN and the Mask R-CNN with hyperparameter tuning and transfer learning considered  ...  Model mAP F1-Score baseline Mask R-CNN 70.77 52.33 Mask R-CNN (hyperparameter tuning and transfer learning) 81.70 59.13 Cascade Mask R-CNN (hyperparameter tuning and transfer learning) 83.60 62.07  ... 
doi:10.3390/app12115553 fatcat:yllauvmdqbealf6fodcrjvrppu
« Previous Showing results 1 — 15 out of 12,153 results