APL Machine Learning最新文献_第4页

Brains and bytes: Trends in neuromorphic technology 大脑和字节:神经形态技术的趋势

APL Machine Learning Pub Date : 2023-06-01 DOI: 10.1063/5.0162712

A. Mehonic, J. Eshraghian

引用次数: 3

Multiplexed gradient descent: Fast online training of modern datasets on hardware neural networks without backpropagation 多路梯度下降:无反向传播的硬件神经网络上现代数据集的快速在线训练

APL Machine Learning Pub Date : 2023-06-01 DOI: 10.1063/5.0157645

Adam N. McCaughan, Bakhrom G. Oripov, Natesh Ganesh, Sae Woo Nam, Andrew Dienstfrey, Sonia M. Buckley

{"title":"Multiplexed gradient descent: Fast online training of modern datasets on hardware neural networks without backpropagation","authors":"Adam N. McCaughan, Bakhrom G. Oripov, Natesh Ganesh, Sae Woo Nam, Andrew Dienstfrey, Sonia M. Buckley","doi":"10.1063/5.0157645","DOIUrl":"https://doi.org/10.1063/5.0157645","url":null,"abstract":"We present multiplexed gradient descent (MGD), a gradient descent framework designed to easily train analog or digital neural networks in hardware. MGD utilizes zero-order optimization techniques for online training of hardware neural networks. We demonstrate its ability to train neural networks on modern machine learning datasets, including CIFAR-10 and Fashion-MNIST, and compare its performance to backpropagation. Assuming realistic timescales and hardware parameters, our results indicate that these optimization techniques can train a network on emerging hardware platforms orders of magnitude faster than the wall-clock time of training via backpropagation on a standard GPU, even in the presence of imperfect weight updates or device-to-device variations in the hardware. We additionally describe how it can be applied to existing hardware as part of chip-in-the-loop training or integrated directly at the hardware level. Crucially, because the MGD framework is model-free it can be applied to nearly any hardware platform with tunable parameters, and its gradient descent process can be optimized to compensate for specific hardware limitations, such as slow parameter-update speeds or limited input bandwidth.","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135220143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Long-term forecasts of residential energy profiles based on Conv2D and LSTM models (electricity- and gas-based households) 基于Conv2D和LSTM模型(用电户和燃气户)的住宅能源状况长期预测

APL Machine Learning Pub Date : 2023-06-01 DOI: 10.1063/5.0137443

Zaccheus Olaofe Olaofe

{"title":"Long-term forecasts of residential energy profiles based on Conv2D and LSTM models (electricity- and gas-based households)","authors":"Zaccheus Olaofe Olaofe","doi":"10.1063/5.0137443","DOIUrl":"https://doi.org/10.1063/5.0137443","url":null,"abstract":"For power system operation and expansion of grid-import systems, an accurate forecast model plays an essential role in the better management of household electricity demands. With the aim of finding an accurate forecast model in the proper representation of various household energy profiles, our research objective is centered on the development of a reliable forecast system for a group of 24-household energy consumers. In this energy study, we proposed long-term forecasts of (1) residential energy profiles within the multi-classification framework and (2) energy costing of the household demands using the Keras two-dimensional convolutional neural network (Conv2D) model and long short-term memory (LSTM) models. These high-level Keras neural networks are built to extract multivariate features for household energy consumption modeling and forecasting. The proposed forecast systems utilized a similar model hyperparameter configuration, while the forecast skills are validated with spatial–temporal variation datasets of ten remote locations. The actual costs of household demand and supply are estimated and compared with Conv2D predictions. The finding results (hourly and seasonal predictions and model evaluation) revealed that Conv2D and LSTM forecast systems are promising for household energy forecast solutions. Experimental results of the Conv2D predictive system achieved better forecast skills [correlation coefficient (0.727–0.994) and root mean square error (0.190–0.868)] than LSTM forecasts (0.308–0.987 and 0.278–1.212). However, experimental findings revealed that forecast skills of the predictive systems in residential energy demand predictions are highly influenced by the (1) quality of input datasets, (2) model hyperparameter tuning approach, and (3) learning rate of selected network optimizer(s).","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132682014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Designing composition ratio of magnetic alloy multilayer for transverse thermoelectric conversion by Bayesian optimization 利用贝叶斯优化设计横向热电转换用磁性合金多层材料的组成比

APL Machine Learning Pub Date : 2023-05-26 DOI: 10.1063/5.0140332

N. Chiba, K. Masuda, K. Uchida, Y. Miura

{"title":"Designing composition ratio of magnetic alloy multilayer for transverse thermoelectric conversion by Bayesian optimization","authors":"N. Chiba, K. Masuda, K. Uchida, Y. Miura","doi":"10.1063/5.0140332","DOIUrl":"https://doi.org/10.1063/5.0140332","url":null,"abstract":"We demonstrated the effectiveness of the machine learning method combined with first-principles calculations for the enhancement of the anomalous Nernst effect (ANE) of multilayers. The composition ratio of CoNi homogeneous alloy superlattices was optimized by Bayesian optimization so as to maximize the transverse thermoelectric conductivity (αxy). The nonintuitive optimal composition with a large αxy of ∼10 A K−1 m−1 was identified through the two-step Bayesian optimization using rough and fine candidate pools. The Berry curvature and band dispersion analyses revealed that αxy is enhanced by the appearance of the flat band near the Fermi level due to the multilayer formation. The magnitude of the energy derivative of the anomalous Hall conductivity increases owing to the large Berry curvature near the flat band along the R-M high symmetry line, which emerges only in the optimized superlattice, leading to the αxy enhancement. The effective method verified here will broaden the choices of ANE materials to more complex systems and, therefore, lead to the development of transverse thermoelectric conversion technologies.","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"174 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122855944","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Scalable wavelength-multiplexing photonic reservoir computing 可扩展波长复用光子库计算

APL Machine Learning Pub Date : 2023-05-24 DOI: 10.1063/5.0158939

Ruiyue Li, Yi-Wei Shen, Bao-De Lin, Jingyi Yu, Xuming He, Chengyuan Wang

引用次数: 1

A machine learning-based prediction of crystal orientations for multicrystalline materials 基于机器学习的多晶材料晶体取向预测

APL Machine Learning Pub Date : 2023-05-24 DOI: 10.1063/5.0138099

Kyoka Hara, T. Kojima, K. Kutsukake, Hiroaki Kudo, N. Usami

{"title":"A machine learning-based prediction of crystal orientations for multicrystalline materials","authors":"Kyoka Hara, T. Kojima, K. Kutsukake, Hiroaki Kudo, N. Usami","doi":"10.1063/5.0138099","DOIUrl":"https://doi.org/10.1063/5.0138099","url":null,"abstract":"We established a rapid, low-cost, and accurate technique to measure crystallographic orientations in multicrystalline materials by optical images and machine learning. A long short-term memory neural network was trained with pairs of light reflection patterns and the correct orientations of each grain, successfully predicting orientation with an error median of 8.61°. The model was improved by diverse data taken from various incident light angles and by data augmentation. When trained on different incident angles, the model was capable of estimating different orientations. This is related to the geometrical configuration of the incident light angles and surface facets of the crystal. The failure in certain orientations is thought to be complemented by supplementary data taken from different incident angles. Combining data from multiple incident angles, we acquired an error median of 4.35°. Data augmentation was successfully performed, reducing error by an additional 35%. This technique can provide the crystallographic orientations of a 15 × 15 cm2 sized wafer in less than 8 min, while baseline techniques such as electron backscatter diffraction and Laue scanner may take more than 10 h. The rapid and accurate measurement can accelerate data collection for full-sized ingots, helping us gain a comprehensive understanding of crystal growth. We believe that our technique will contribute to controlling crystalline structure for the fabrication of high-performance materials.","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"96 1 Suppl 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129038719","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

Intelligent performance inference: A graph neural network approach to modeling maximum achievable throughput in optical networks 智能性能推断:一种图神经网络方法来模拟光网络中最大可实现吞吐量

APL Machine Learning Pub Date : 2023-05-19 DOI: 10.1063/5.0137426

Robin Matzner, Ruijie Luo, G. Zervas, P. Bayvel

{"title":"Intelligent performance inference: A graph neural network approach to modeling maximum achievable throughput in optical networks","authors":"Robin Matzner, Ruijie Luo, G. Zervas, P. Bayvel","doi":"10.1063/5.0137426","DOIUrl":"https://doi.org/10.1063/5.0137426","url":null,"abstract":"One of the key performance metrics for optical networks is the maximum achievable throughput for a given network. Determining it, however, is a nondeterministic polynomial time (NP) hard optimization problem, often solved via computationally expensive integer linear programming (ILP) formulations. These are infeasible to implement as objectives, even on very small node scales of a few tens of nodes. Alternatively, heuristics are used although these, too, require considerable computation time for a large number of networks. There is, thus, a need for an ultra-fast and accurate performance evaluation of optical networks. For the first time, we propose the use of a geometric deep learning model, message passing neural networks (MPNNs), to learn the relationship between node and edge features, the network structure, and the maximum achievable network throughput. We demonstrate that MPNNs can accurately predict the maximum achievable throughput while reducing the computational time by up to five-orders of magnitude compared to the ILP for small networks (10–15 nodes) and compared to a heuristic for large networks (25–100 nodes)—proving their suitability for the design and optimization of optical networks on different time- and distance-scales.","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122263355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Glass transition of amorphous polymeric materials informed by machine learning 基于机器学习的非晶聚合物材料的玻璃化转变

APL Machine Learning Pub Date : 2023-05-04 DOI: 10.1063/5.0137357

Anwen Hu, Yongdi Huang, Qionghai Chen, Wanhui Huang, Xiaohui Wu, Lihong Cui, Yining Dong, Jun Liu

{"title":"Glass transition of amorphous polymeric materials informed by machine learning","authors":"Anwen Hu, Yongdi Huang, Qionghai Chen, Wanhui Huang, Xiaohui Wu, Lihong Cui, Yining Dong, Jun Liu","doi":"10.1063/5.0137357","DOIUrl":"https://doi.org/10.1063/5.0137357","url":null,"abstract":"The glass transition temperature (Tg) is used to determine thermophysical properties of polymer materials and is often considered one of the most important descriptors. Methods for predicting various physical properties of materials based on machine learning algorithms and key molecular descriptors are efficient and accurate. However, it still needs improvements because an overly complex model is less practical and difficult to generalize. In addition, obtaining a large number of samples to achieve accurate predictions remains a challenge due to the complex and lengthy experimental process. In this work, based on Tg of 100 polymers, we use a feature selection algorithm combining FeatureWiz and the least absolute shrinkage and selection operator to quickly select molecular descriptors that are minimally redundant and maximally relevant to Tg. The processed dataset is interpolated from the original dataset using the nearest neighbor interpolation algorithm to solve the data deficiency problem. Finally, the synthetic minority oversampling technique algorithm is used to solve the data imbalance problem. The augmented dataset is used to construct the extreme gradient boosting prediction model to achieve good prediction accuracy. The experimental results demonstrate the robustness of the proposed model and the accuracy of its prediction results.","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127408431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Label free identification of different cancer cells using deep learning-based image analysis 使用基于深度学习的图像分析对不同的癌细胞进行无标签识别

APL Machine Learning Pub Date : 2023-05-02 DOI: 10.1063/5.0141730

Karl Gardner, Rutwik Joshi, Md Nayeem Hasan Kashem, T. Pham, Qiugang Lu, Wei Li

{"title":"Label free identification of different cancer cells using deep learning-based image analysis","authors":"Karl Gardner, Rutwik Joshi, Md Nayeem Hasan Kashem, T. Pham, Qiugang Lu, Wei Li","doi":"10.1063/5.0141730","DOIUrl":"https://doi.org/10.1063/5.0141730","url":null,"abstract":"Cancer diagnostics is an important field of cancer recovery and survival with many expensive procedures needed to administer the correct treatment. Machine Learning (ML) approaches can help with the diagnostic prediction from circulating tumor cells in liquid biopsy or from a primary tumor in solid biopsy. After predicting the metastatic potential from a deep learning model, doctors in a clinical setting can administer a safe and correct treatment for a specific patient. This paper investigates the use of deep convolutional neural networks for predicting a specific cancer cell line as a tool for label free identification. Specifically, deep learning strategies for weight initialization and performance metrics are described, with transfer learning and the accuracy metric utilized in this work. The equipment used for prediction involves brightfield microscopy without the use of chemical labels, advanced instruments, or time-consuming biological techniques, giving an advantage over current diagnostic methods. In the procedure, three different binary datasets of well-known cancer cell lines were collected, each having a difference in metastatic potential. Two different classification models were adopted (EfficientNetV2 and ResNet-50) with the analysis given for each stage in the ML architecture. The training results for each model and dataset are provided and systematically compared. We found that the test set accuracy showed favorable performance for both ML models with EfficientNetV2 accuracy reaching up to 99%. These test results allowed EfficientNetV2 to outperform ResNet-50 at an average percent increase of 3.5% for each dataset. The high accuracy obtained from the predictions demonstrates that the system can be retrained on a large-scale clinical dataset.","PeriodicalId":229559,"journal":{"name":"APL Machine Learning","volume":"93 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126347886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep learning of nonlinear flame fronts development due to Darrieus–Landau instability 基于达里厄-朗道不稳定性的非线性火焰锋面发展的深度学习

APL Machine Learning Pub Date : 2023-04-11 DOI: 10.1063/5.0139857

Rixin Yu

引用次数: 0