Back‐dropout transfer learning for action recognition

Transfer learning study on vortex-induced vibration of circular and bending cylinders in laminar flows. Vortex-induced vibrations of circular and bending cylinders have been detected by varying flowrate in rectangle cylinder. The vibration date in different position of the cylinders are acquired by LMS and image information obtained. The dynamic characteristics of vortex shedding are analyzed by the cylinder vibration and flow rate signal. The experimental data of the circular cylinder vibration is used as auxiliary data set and a few data of the bending cylinder vibration as target data set. Both of auxiliary and target data sets compose the training set which is trained by TrAdaboost based on weight iteration adjustment. The result compared with another experimental data of bending cylinder shows that the transfer learning algorithm has a reasonable accuracy.

Transfer learning, which aims to reuse knowledge in different domains, has achieved great success in many scenarios via minimizing domain discrepancy and enhancing feature discriminability. However, there are seldom practical determination methods for measuring the transferability among domains. In this paper, we bring forward a novel meta-transfer feature method (MetaTrans) for this problem. MetaTrans is used to train a model to predict performance improvement ratio from historical transfer learning experiences, and can consider both the Transferability between tasks and the Discriminability emphasized on targets. We apply this method to both shallow and deep transfer learning algorithms, providing a detail explanation for the success of specific transfer learning algorithms. From experimental studies, we find that different transfer learning algorithms have varying dominant factor deciding their success, so we propose a multi-task learning framework which can learn both common and specific experience from historical transfer learning results. The empirical investigations reveal that the knowledge obtained from historical experience can facilitate future transfer learning tasks.

In the field of supervised machine learning, a transfer learning environment is defined as the training data having different distribution characteristics than the testing data. This is due to the lack of available labeled data for the domain of interest, which prompts an alternate domain to be used as the training data. Because there is insufficient labeled data from the domain of interest, validation techniques cannot be reliably used for the algorithm selection process in a transfer learning environment. A transfer learning algorithm is typically comprised of a domain adaptation step followed by a learning step. The learning step is usually implemented using a traditional machine learning algorithm. In this paper, we examine and analyze the impact that the traditional machine learning algorithm (the learning step) has on the overall performance of a transfer learning algorithm. Using the transfer learning test framework, we test five state-of-the-art transfer learning algorithms coupled with seven different traditional learning algorithms for a total of 35 unique transfer learning algorithms. For our experiment, no labeled data from the domain of interest is available for the training process. Since validation techniques cannot be reliably used for the algorithm selection process in a transfer learning environment, it is important for machine learning researchers and practitioners to understand the impact of a traditional machine learner on the overall performance of a transfer learning algorithm.

Breast cancer is a significant factor in female mortality. An early cancer diagnosis leads to a reduction in the breast cancer death rate. With the help of a computer-aided diagnosis system, the efficiency increased, and the cost was reduced for the cancer diagnosis. Traditional breast cancer classification techniques are based on handcrafted features techniques, and their performance relies upon the chosen features. They also are very sensitive to different sizes and complex shapes. However, histopathological breast cancer images are very complex in shape. Currently, deep learning models have become an alternative solution for diagnosis, and have overcome the drawbacks of classical classification techniques. Although deep learning has performed well in various tasks of computer vision and pattern recognition, it still has some challenges. One of the main challenges is the lack of training data. To address this challenge and optimize the performance, we have utilized a transfer learning technique which is where the deep learning models train on a task, and then fine-tune the models for another task. We have employed transfer learning in two ways: Training our proposed model first on the same domain dataset, then on the target dataset, and training our model on a different domain dataset, then on the target dataset. We have empirically proven that the same domain transfer learning optimized the performance. Our hybrid model of parallel convolutional layers and residual links is utilized to classify hematoxylin–eosin-stained breast biopsy images into four classes: invasive carcinoma, in-situ carcinoma, benign tumor and normal tissue. To reduce the effect of overfitting, we have augmented the images with different image processing techniques. The proposed model achieved state-of-the-art performance, and it outperformed the latest methods by achieving a patch-wise classification accuracy of 90.5%, and an image-wise classification accuracy of 97.4% on the validation set. Moreover, we have achieved an image-wise classification accuracy of 96.1% on the test set of the microscopy ICIAR-2018 dataset.

Production forecasting is vital in the oil and gas sector, empowering engineers with insights for effective reservoir management. This paper introduces the concept of Transfer Learning as a powerful tool in the domain of machine-learning-assisted production forecasting that accounts of 3D spatial distributions of three geological properties, namely porosity, permeability, and saturation, two completion parameters, namely hydraulic fracture height and length, and production constraint. Transfer learning efficiently leverages knowledge from one problem to improve generalization on another, especially when data is scarce and computational resources are limited. To demonstrate the utility of transfer learning, we evaluate two scenarios of transfer learning. The first transfer learning scenario demonstrates the generalization of the forecasting to cases with variable hydraulic fracture spacing using limited training data. The second transfer learning scenario demonstrates the generalization of the forecasting to cases with variable natural fracture spacing and natural fracture permeability using limited training data. Source dataset contained 2000 realizations, while the target dataset contained 20, 40, 60, 80, 100, 250, 500, or 1000 realizations to represent the scenarios of limited training-data availability. The study confirms the benefits of transfer learning when the training dataset size is small (generally less than 100 training realizations); however, under large training dataset size (around 500 or more training realizations), transfer learning is not needed. For the first scenario involving variable hydraulic fracture spacing, the use of transfer learning ensured that the source model can be trained on target dataset with 80 realizations for gas rate forecasting at an accuracy of 25% in terms of MAPE, and with 500 realizations for condensate rate forecasting at an accuracy of 12% in terms of MAPE. Similarly, for the second scenario involving variable natural fracture spacing and natural fracture permeability, the use of transfer learning ensured that the source model can be trained on target dataset with 250 realizations for gas rate forecasting at an accuracy of 24% in terms of MAPE, and with 500 realizations for condensate rate forecasting at an accuracy of 23% in terms of MAPE. This illustrates the potential of transfer learning in improving forecasting models with limited data using a well pre-trained model and enhanced hyperparameter tuning of the transfer learning model. For cases with 500 or more training realizations, transfer learning severely underperforms as compared to training a conventional machine-learning model from scratch. The paper explores two cases of transfer learning.

In this work, we design and evaluate a computational learning model that enables a human-robot team to co-develop joint strategies for performing novel tasks that require coordination. The joint strategies are learned through "perturbation training," a human team-training strategy that requires team members to practice variations of a given task to help their team generalize to new variants of that task. We formally define the problem of human-robot perturbation training and develop and evaluate the first end-to-end framework for such training, which incorporates a multi-agent transfer learning algorithm, human-robot co-learning framework and communication protocol. Our transfer learning algorithm, Adaptive Perturbation Training (AdaPT), is a hybrid of transfer and reinforcement learning techniques that learns quickly and robustly for new task variants. We empirically validate the benefits of AdaPT through comparison to other hybrid reinforcement and transfer learning techniques aimed at transferring knowledge from multiple source tasks to a single target task. We also demonstrate that AdaPT's rapid learning supports live interaction between a person and a robot, during which the human-robot team trains to achieve a high level of performance for new task variants. We augment AdaPT with a co-learning framework and a computational bi-directional communication protocol so that the robot can co-train with a person during live interaction. Results from large-scale human subject experiments (n=48) indicate that AdaPT enables an agent to learn in a manner compatible with a human's own learning process, and that a robot undergoing perturbation training with a human results in a high level of team performance. Finally, we demonstrate that human-robot training using AdaPT in a simulation environment produces effective performance for a team incorporating an embodied robot partner.

Kiwifruit has a high decay rate, in part because quality changes during storage cannot be easily monitored in real time. In order to better monitor the shelf life of kiwifruit and understand the quality changing process during storage, internal quality indexes such as hardness, respiratory intensity and TSS(Total Soluble Solid) were considered into the prediction models. The prediction models were constructed based on BPNN (Back Propagation Neural Network), Random Forest (RF) and XGBoost (eXtreme Gradient Boosting) respectively. And transfer learning algorithm was used to construct the quality prediction models with BPNN, RF, and XGBoost algorithms as the base learner. In the experiments, sample data were augmented by adding Gaussian noise, which effectively prevented the model from over-fitting. The experimental results showed that the prediction accuracy of each index based on transfer learning was better than that of individual BPNN, RF and XGBoost. Moreover, the average prediction accuracy of the models based on transfer learning was 96.2%, and that of respiratory intensity was as high as 99.4%. Therefore transfer learning can be used to effectively analyze and predict changes of kiwifruit quality indexes during storage.

Near-infrared spectroscopy mainly reflects the frequency-doubled and total-frequency absorption information of hydrogen-containing groups (O‒H, C‒H, N‒H, S‒H) in organic molecules for near-infrared lights with different wavelengths, so it is applicable to testing of most raw materials and products in the field of petrochemicals. However, the modeling process needs to collect a large number of laboratory analysis data. There are many oil sources in China, and oil properties change frequently. Modeling of each raw material is not only unfeasible but also will affect its engineering application efficiency. In order to achieve rapid modeling of near-infrared spectroscopy and based on historical data of different crude oils under different detection conditions, this paper discusses about the feasibility of the application of transfer learning algorithm and makes it possible that transfer learning can assist in rapid modeling using certain historical data under similar distributions under a small quantity of new data. In consideration of the requirement of transfer learning for certain similarity of different datasets, a transfer learning method based on local similarity feature selection is proposed. The simulation verification of spectral data of 13 crude oils measured by three different probe detection methods is performed. The effectiveness and application scope of the transfer modeling method under different similarity conditions are analyzed.

Multiple Kernel Based Transfer Learning for the Few-Shot Recognition Task in Smart Home Scene

The usage of Artificial intelligence in medical arena has proved to be a game changer in the detection and diagnosis of several medical conditions. In the current digital era, children with stressful medical issues are suffering from Deep Obsessive-Compulsive Disorder (DOCD). This kind of mental stress occurs in children because of the continuous usage of gadgets such as mobile phone, playing games using play stations, watching videos on tablets, etc. In most of the possibilities, single children are the ones affected with several obsessions such as stubborn activities, fighting for selfish priorities and so on. In medical terms, these kinds of complex behavioral changes are identified as DOCD. Genetic behaviors sometimes in a few group of children are also noticed as a modality difference. As symptoms are psychiatric impairment, such a child remains isolated, abnormal silence, being obsessive and repeating irrelevant words, high stress or anxiety. All medical challenges could be treated as healthcare research metrics and the gradual increase in DOCD disorder among children of this generation can be considered too. Early detection of DOCD is essential as it can help in early diagnosis but techniques to do so is unavailable currently. Deep learning-an artificial intelligence method can be utilized to detect DOCD, diagnose and treat it and bring about a positive character in children. Behavior changes in children can be classified and detected using transfer learning algorithms. In COVID-19 pandemic situation, 3% of DOCD has increased to 10-15% as a disorder. This information is retrieved from children by monitoring negative activities, unusual behavior such as nail biting, removing spectacles and placing them in the wrong place, watching tablets, mobile phones and television for more hours. Using Convolutional Neural Networks (CNN), input such as MRI (Magnetic resonance Imaging) is used for experimenting the variations in behavior with the high dimension that are analyzed from the image dataset. Using Transfer Learning with Inception V3-, CNN generalization of misophonia level can be statistically analyzed to avoid overfitting problems. By employing AI techniques, the aggression level can be predicted using data augmentation method with better accuracy and a low error rate than the existing systems. In the research it is observed that using the model employing Inception-V3 transfer learning CNN a better prediction of aggression levels can be achieved in comparison to the existing CNN model used.

Deep learning methods have huge success in task specific feature representation. Transfer learning algorithms are very much effective when large training data is scarce. It has been significantly used for diagnosis of diseases in medical imaging. This article presents a systematic literature review (SLR) by conducting a comparison of a variety of transfer learning approaches with healthcare experts in diagnosing diseases from medical imaging. This study has been compiled by reviewing research studies published in renowned venues between 2014 and 2019. Moreover, the data for the diagnosis performed by health care experts has also been acquired to perform a detailed comparative analysis for a wide range of diseases. The analysis has been performed on the basis of diseases, transfer learning approaches, type of medical imaging used. The comparative analysis is based on performance indices reported in studies which include diagnostic accuracy, true-positive (TP), false-positive (FP), true-negative (TN), false-negative (FN) sensitivity, specificity, and the area under the receiver operating characteristic curve (AUROC). A total of5,188articles were identified out of which 63 studies were included. Among them 21 research studies contain sufficient data to construct the evaluation tables that enable process of test accuracy of transfer learning having sensitivity ranged from 71% to 100% (mean 85.25%) and specificity ranged from 64% to 100% (mean 81.92%). Furthermore, health experts having sensitivity ranged from 33% to 100% (mean 85.27%) and specificity ranged from 82% to 100% (mean 91.63%).This SLR found that diagnostic accuracy of transfer learning is approximately equivalent to the diagnosis of health experts. The results also revealed that convolutional neural networks (CNN) have been extensively used for disease diagnosis from medical imaging. Finally, inappropriate exposure of diseases in transfer learning studies restricts reliable elucidation of the outcomes of diagnostic accuracy.

Deep neural networks (DNNs), e.g., convolutional neural network (CNN), are able to learn effective features from wafer maps for dimensional reduction and feature extraction. However, very large image data are needed to train DNNs to obtain high generalization performance. It is still a difficult task due to the lack of sufficient labeled images with various defects. This article proposes a semisupervised deep transfer learning algorithm called joint feature and label adversarial network (JFLAN). JFLAN uses CNNs to extract transferable features of wafer maps and then introduces a multilayer domain adaptation and pseudolabel learning block based on the generative adversarial network (GAN). This effectively reduces the distribution discrepancy and the among-class distance of the transferable features. Finally, JFLAN transfers knowledge from wafer image source data collected offline and then achieves the goal of significantly improved accuracy of wafer defect recognition and realizes online adaptive defect recognition. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —The defect recognition on wafer maps plays a key role to recognize fault sources of the semiconductor manufacturing processes. Transfer learning is able to use the existing labeled data to assist in the classification of unlabeled data and is very effective to solve the problem of small samples and nonstationary generalization errors. In particular, the infusion of adversarial learning in transfer learning will provide a new idea for deep feature learning. This article provides a novel method based on transfer learning to implement wafer map defect recognition (WMDR) to quickly identify defect root causes for yield enhancement. This article provides a novel way for quality control of semiconductor manufacturing processes based on transfer and adversarial learning.

Analyzing sentiment analysis with deep learning requires massive labeled datasets where such data is not always available. The annotation process is also time-consuming and tedious. Further, even after we train the sentiment analysis, it creates another problem. Because this model is domain-dependent, the performance in another domain estimated to perform poorly. In this paper, we present the transfer learning approach to transfer knowledge gained from the source dataset into the target dataset with the expectation to improve the target model. Multichannel Convolutional Neural Network deploys different n-grams as the input channel in a single CNN model to grasp meaningful features from the text. This method has proven to perform well in sentiment analysis problems. We train our three datasets with different domains using this method as the baseline. The largest dataset then becomes the source model for transfer learning and other datasets as the target. Fine-tuning our source model also needed when retraining it into the target dataset. From the evaluation, we show that several transfer learning strategies outperform the domain-specific model, even when the data is imbalanced. We also highlight certain failing strategies that inflict lousy results on the target model performance.

We propose a novel unsupervised transfer learning framework that utilises unlabelled auxiliary data to quantify and select the most relevant transferrable knowledge for recognising a target object class from the background given very limited training target samples. Unlike existing transfer learning techniques, our method does not assume that auxiliary data are labelled, nor the relationships between target and auxiliary classes are known a priori. Our unsupervised transfer learning is formulated by a novel kernel adaptation transfer (KAT) learning framework, which aims to (a) extract general knowledge about how more structured objects are visually distinctive from cluttered background regardless object class, and (b) more importantly, perform selective transfer of knowledge extracted from the auxiliary data to minimise negative knowledge transfer suffered by existing methods. The effectiveness and efficiency of the proposed approach is demonstrated by performing one-class object recognition (object vs. background) task using the Caltech256 dataset.

Deep convolutional neural networks with ensemble learning and transfer learning for capacity estimation of lithium-ion batteries