Since its creation, the ImageNet-1k benchmark set has played a significant role as a benchmark for ascertaining the accuracy of different deep neural net (DNN) models on the classification problem. Moreover, in recent years it has also served as the principal benchmark for assessing different approaches to DNN training. Finishing a 90-epoch ImageNet-1k training with ResNet-50 on a NVIDIA M40 GPU takes 14 days. This training requires 10 18 single precision operations in total. On the other hand, the world's current fastest supercomputer can finish 2 × 10 17 single precision operations per second. If we can make full use of the computing capability of the fastest supercomputer for DNN training, we should be able to finish the 90-epoch ResNet-50 training in five seconds. Over the last two years, a number of researchers have focused on closing this significant performance gap through scaling DNN training to larger numbers of processors. Most successful approaches to scaling ImageNet training have used the synchronous stochastic gradient descent. However, to scale synchronous stochastic gradient descent one must also increase the batch size used in each iteration. Thus, for many researchers, the focus on scaling DNN training has translated into a focus on developing training algorithms that enable increasing the batch size in data-parallel synchronous stochastic gradient descent without losing accuracy over a fixed number of epochs. As a result, we have seen the batch size and number of processors successfully utilized increase from 1K batch size on 128 processors to 8K batch size on 256 processors over the last two years. The recently published LARS algorithm increased batch size further to 32K for some DNN models. Following up on this work, we wished to confirm that LARS could be used to further scale the number of processors efficiently used in DNN training and, and as a result, further reduce the total training time. In this paper we present the results of this investigation: using LARS we efficiently utilized 1024 CPUs to finish the 100-epoch ImageNet training with AlexNet in 11 minutes with 58.6% accuracy (batch size = 32K), and we utilized 2048 KNLs to finish the 90-epoch ImageNet training with ResNet-50 in 20 minutes without losing accuracy (batch size = 32K). State-of-the-art ImageNet training speed with ResNet-50 is 74.9% top-1 test accuracy in 15 minutes (Akiba, Suzuki, and Fukuda 2017). We got 74.9% top-1 test accuracy in 64 epochs, which only needs 14 minutes. Furthermore, when the batch size is above 16K, our accuracy using LARS is much higher than Facebooks corresponding batch sizes (Figure 1). Our code is available upon request. 15 mins Our version 32K 74.9% 14 mins arXiv:1709.05011v10 [cs.CV]

No abstract

The current mainstream approach of using manual measurements and visual inspections for crop lodging detection is inefficient, time-consuming, and subjective. An innovative method for wheat lodging detection that can overcome or alleviate these shortcomings would be welcomed. This study proposed a systematic approach for wheat lodging detection in research plots (372 experimental plots), which consisted of using unmanned aerial systems (UAS) for aerial imagery acquisition, manual field evaluation, and machine learning algorithms to detect the occurrence or not of lodging. UAS imagery was collected on three different dates (23 and 30 July 2019, and 8 August 2019) after lodging occurred. Traditional machine learning and deep learning were evaluated and compared in this study in terms of classification accuracy and standard deviation. For traditional machine learning, five types of features (i.e. gray level co-occurrence matrix, local binary pattern, Gabor, intensity, and Hu-moment) were extracted and fed into three traditional machine learning algorithms (i.e., random forest (RF), neural network, and support vector machine) for detecting lodged plots. For the datasets on each imagery collection date, the accuracies of the three algorithms were not significantly different from each other. For any of the three algorithms, accuracies on the first and last date datasets had the lowest and highest values, respectively. Incorporating standard deviation as a measurement of performance robustness, RF was determined as the most satisfactory. Regarding deep learning, three different convolutional neural networks (simple convolutional neural network, VGG-16, and GoogLeNet) were tested. For any of the single date datasets, GoogLeNet consistently had superior performance over the other two methods. Further comparisons between RF and GoogLeNet demonstrated that the detection accuracies of the two methods were not significantly different from each other (p > 0.05); hence, the choice of any of the two would not affect the final detection accuracies. However, considering the fact that the average accuracy of GoogLeNet (93%) was larger than RF (91%), it was recommended to use GoogLeNet for wheat lodging detection. This research demonstrated that UAS RGB imagery, coupled with the GoogLeNet machine learning algorithm, can be a novel, reliable, objective, simple, low-cost, and effective (accuracy > 90%) tool for wheat lodging detection.

Introduction !Traditional Chinese medicine (TCM), one of the treasures of the Chinese nation, has made great contributions to the development of both Chinese and world civilization for thousands of years. As traditional medicine and the research technology of TCM develop more rapidly than ever before, much attention is being paid to it worldwide because of its abundant resources, unique curative effects, low toxic and side effects, and more is being invested in its research and development while posing more demanding requirements for the quality of TCM products. However, many problems exist during the conventional TCM herbs production process in China, including unclear germplasm resources, substandard planting and processing techniques, excessive pesticide residue and poor quality of TCM herbs, thus leading to inconsistent quality of Chinese-made TCM products and severely affecting the stability of the curative effects of TCM as well as the lead position of Chinese TCM products in the world drug market. These challenges can be tackled in part through a Good Agricultural Practices (GAP) approach, which aims to standardize the cultivation, collection, and processing of TCM herbs, improving their quality and bringing TCM in line with the international practice, being therefore a key to the modernization and internationalization of TCM. GAP for TCM herbsGAP for TCM herbs, referred to as Chinese crude drugs (CCD) in Chinese official documents, is intended to control various factors affecting the production quality of medicinal plant materials, to standardize various crude drug production processes and even the whole process so as to ensure that TCM herbs are authentic, safe, effective, and consistent in quality [1]. Herein, TCM herbs cover CCD, herbal medicines, ethnodrugs, and introduced botanical drugs. Formulation of GAP for TCM herbsAlong with the worldwide trend for returning to the nature, people are paying more attention to the medical and health care function of traditional medicines, especially TCM, and some countries have formulated relevant legislations concerning the quality control of herbal medicines. Many countries have taken a series of standardized measures concerning quality control of production of raw materials for natural medicines. For example, the Japanese Ministry of Health, Labor, and Welfare revised Medicinal Plants Planting and Quality Evaluation in 1992; the US government issued a draft FDA Guidance for Industry: Botanical Drug Products in 1996; the European Herb Growers Association (Europam) proposed Good Quality Control of Medicinal Plants and Animals in August 1998; later, the European Union (EU) drafted Guidelines for Good Agricul- Abstract !In this paper, we briefly review international Good Agricultural Practice (GAP) regulations related to traditional Chinese medicine herbs (TCM herbs) and the background of the drafting process and the implementation of GAP for TCM herbs in China. We also have summarized progress and achievements since the implementation of GAP for TCM herbs in 200...

No abstract