Few-shot learning aims to recognize novel classes from a few examples. Although significant progress has been made in the image domain, few-shot video classification is relatively unexplored. We argue that previous methods underestimate the importance of video feature learning and propose to learn spatiotemporal features using a 3D CNN. Proposing a two-stage approach that learns video features on base classes followed by fine-tuning the classifiers on novel classes, we show that this simple baseline approach outperforms prior few-shot video classification methods by over 20 points on existing benchmarks. To circumvent the need of labeled examples, we present two novel approaches that yield further improvement. First, we leverage tag-labeled videos from a large dataset using tag retrieval followed by selecting the best clips with visual similarities. Second, we learn generative adversarial networks that generate video features of novel classes from their semantic embeddings. Moreover, we find existing benchmarks are limited because they only focus on 5 novel classes in each testing episode and introduce more realistic benchmarks by involving more novel classes, i.e. few-shot learning, as well as a mixture of novel and base classes, i.e. generalized few-shot learning. The experimental results show that our retrieval and feature generation approach significantly outperform the baseline approach on the new benchmarks.
Chinese text recognition is more challenging than Latin text due to the large amount of fine-grained Chinese characters and the great imbalance over classes, which causes a serious overfitting problem. We propose to apply Maximum Entropy Regularization to regularize the training process, which is to simply add a negative entropy term to the canonical cross-entropy loss without any additional parameters and modification of a model. We theoretically give the convergence probability distribution and analyze how the regularization influence the learning process. Experiments on Chinese character recognition, Chinese text line recognition and fine-grained image classification achieve consistent improvement, proving that the regularization is beneficial to generalization and robustness of a recognition model.
Recently, the concept of teaching has been introduced into machine learning, in which a teacher model is used to guide the training of a student model (which will be used in real tasks) through data selection, loss function design, etc. Learning to reweight, which is a specific kind of teaching that reweights training data using a teacher model, receives much attention due to its simplicity and effectiveness. In existing learning to reweight works, the teacher model only utilizes shallow/surface information such as training iteration number and loss/accuracy of the student model from training/validation sets, but ignores the internal states of the student model, which limits the potential of learning to reweight. In this work, we propose an improved data reweighting algorithm, in which the student model provides its internal states to the teacher model, and the teacher model returns adaptive weights of training samples to enhance the training of the student model. The teacher model is jointly trained with the student model using meta gradients propagated from a validation set. Experiments on image classification with clean/noisy labels and neural machine translation empirically demonstrate that our algorithm makes significant improvement over previous methods.
In this paper, detection of deception attack on deep neural network (DNN) based image classification in autonomous and cyber-physical systems is considered. Several studies have shown the vulnerability of DNN to malicious deception attacks. In such attacks, some or all pixel values of an image are modified by an external attacker, so that the change is almost invisible to the human eye but significant enough for a DNN-based classifier to misclassify it. This paper first proposes a novel pre-processing technique that facilitates the detection of such modified images under any DNN-based image classifier as well as the attacker model. The proposed pre-processing algorithm involves a certain combination of principal component analysis (PCA)-based decomposition of the image, and random perturbation based detection to reduce computational complexity. Next, an adaptive version of this algorithm is proposed where a random number of perturbations are chosen adaptively using a doubly-threshold policy, and the threshold values are learnt via stochastic approximation in order to minimize the expected number of perturbations subject to constraints on the false alarm and missed detection probabilities. Numerical experiments show that the proposed detection scheme outperforms a competing algorithm while achieving reasonably low computational complexity.
Several text classification tasks such as sentiment analysis, news categorization, multi-label classification and opinion classification are challenging problems even for modern deep learning networks. Recently, Capsule Networks (CapsNets) are proposed for image classification. It has been shown that CapsNets have several advantages over Convolutional Neural Networks (CNNs), while their validity in the domain of text has been less explored. In this paper, we propose a novel hybrid architecture viz., BGCapsule, which is a Capsule model preceded by an ensemble of Bidirectional Gated Recurrent Units (BiGRU) for several text classification tasks. We employed an ensemble of Bidirectional GRUs for feature extraction layer preceding the primary capsule layer. The hybrid architecture, after performing basic pre-processing steps, consists of five layers: an embedding layer based on GloVe, a BiGRU based ensemble layer, a primary capsule layer, a flatten layer and fully connected ReLU layer followed by a fully connected softmax layer. In order to evaluate the effectiveness of BGCapsule, we conducted extensive experiments on five benchmark datasets (ranging from 10,000 records to 700,000 records) including Movie Review (MR Imdb 2005), AG News dataset, Dbpedia ontology dataset, Yelp Review Full dataset and Yelp review polarity dataset. These benchmarks cover several text classification tasks such as news categorization, sentiment analysis, multiclass classification, multi-label classification and opinion classification. We found that our proposed architecture (BGCapsule) achieves better accuracy compared to the existing methods without the help of any external linguistic knowledge such as positive sentiment keywords and negative sentiment keywords. Further, BGCapsule converged faster compared to other extant techniques.
Replacing normal convolutions with group convolutions can significantly increase the computational efficiency of modern deep convolutional networks, which has been widely adopted in compact network architecture designs. However, existing group convolutions undermine the original network structures by cutting off some connections permanently resulting in significant accuracy degradation. In this paper, we propose dynamic group convolution (DGC) that adaptively selects which part of input channels to be connected within each group for individual samples on the fly. Specifically, we equip each group with a small feature selector to automatically select the most important input channels conditioned on the input images. Multiple groups can adaptively capture abundant and complementary visual/semantic features for each input image. The DGC preserves the original network structure and has similar computational efficiency as the conventional group convolution simultaneously. Extensive experiments on multiple image classification benchmarks including CIFAR-10, CIFAR-100 and ImageNet demonstrate its superiority over the existing group convolution techniques and dynamic execution methods. The code is available at https://github.com/zhuogege1943/dgc.
Quantization is essential to simplify DNN inference in edge applications. Existing uniform and non-uniform quantization methods, however, exhibit an inherent conflict between the representing range and representing resolution, and thereby result in either underutilized bit-width or significant accuracy drop. Moreover, these methods encounter three drawbacks: i) the absence of a quantitative metric for in-depth analysis of the source of the quantization errors; ii) the limited focus on the image classification tasks based on CNNs; iii) the unawareness of the real hardware and energy consumption reduced by lowering the bit-width. In this paper, we first define two quantitative metrics, i.e., the Clipping Error and rounding error, to analyze the quantization error distribution. We observe that the boundary- and rounding- errors vary significantly across layers, models and tasks. Consequently, we propose a novel quantization method to quantize the weight and activation. The key idea is to Approximate the Uniform quantization by Adaptively Superposing multiple Non-uniform quantized values, namely AUSN. AUSN is consist of a decoder-free coding scheme that efficiently exploits the bit-width to its extreme, a superposition quantization algorithm that can adapt the coding scheme to different DNN layers, models and tasks without extra hardware design effort, and a rounding scheme that can eliminate the well-known bit-width overflow and re-quantization issues. Theoretical analysis~(see Appendix A) and accuracy evaluation on various DNN models of different tasks show the effectiveness and generalization of AUSN. The synthesis~(see Appendix B) results on FPGA show $2\times$ reduction of the energy consumption, and $2\times$ to $4\times$ reduction of the hardware resource.
The emergence of powerful artificial intelligence is defining new research directions in neuroscience. To date, this research has focused largely on deep neural networks trained using supervised learning, in tasks such as image classification. However, there is another area of recent AI work which has so far received less attention from neuroscientists, but which may have profound neuroscientific implications: deep reinforcement learning. Deep RL offers a comprehensive framework for studying the interplay among learning, representation and decision-making, offering to the brain sciences a new set of research tools and a wide range of novel hypotheses. In the present review, we provide a high-level introduction to deep RL, discuss some of its initial applications to neuroscience, and survey its wider implications for research on brain and behavior, concluding with a list of opportunities for next-stage research.
The observation of sparse trainable sub-networks within over-parametrized networks - also known as Lottery Tickets (LTs) - has prompted inquiries around their trainability, scaling, uniqueness, and generalization properties. Across 28 combinations of image classification tasks and architectures, we discover differences in the connectivity structure of LTs found through different iterative pruning techniques, thus disproving their uniqueness and connecting emergent mask structure to the choice of pruning. In addition, we propose a consensus-based method for generating refined lottery tickets. This lottery ticket denoising procedure, based on the principle that parameters that always go unpruned across different tasks more reliably identify important sub-networks, is capable of selecting a meaningful portion of the architecture in an embarrassingly parallel way, while quickly discarding extra parameters without the need for further pruning iterations. We successfully train these sub-networks to performance comparable to that of ordinary lottery tickets.
Generative Adversarial Networks (GANs) based semi-supervised learning (SSL) approaches are shown to improve classification performance by utilizing a large number of unlabeled samples in conjunction with limited labeled samples. However, their performance still lags behind the state-of-the-art non-GAN based SSL approaches. We identify that the main reason for this is the lack of consistency in class probability predictions on the same image under local perturbations. Following the general literature, we address this issue via label consistency regularization, which enforces the class probability predictions for an input image to be unchanged under various semantic-preserving perturbations. In this work, we introduce consistency regularization into the vanilla semi-GAN to address this critical limitation. In particular, we present a new composite consistency regularization method which, in spirit, leverages both local consistency and interpolation consistency. We demonstrate the efficacy of our approach on two SSL image classification benchmark datasets, SVHN and CIFAR-10. Our experiments show that this new composite consistency regularization based semi-GAN significantly improves its performance and achieves new state-of-the-art performance among GAN-based SSL approaches.
Diseases and pests cause huge economic loss to the apple industry every year. The identification of various apple diseases is challenging for the farmers as the symptoms produced by different diseases may be very similar, and may be present simultaneously. This paper is an attempt to provide the timely and accurate detection and identification of apple diseases. In this study, we propose a deep learning based approach for identification and classification of apple diseases. The first part of the study is dataset creation which includes data collection and data labelling. Next, we train a Convolutional Neural Network (CNN) model on the prepared dataset for automatic classification of apple diseases. CNNs are end-to-end learning algorithms which perform automatic feature extraction and learn complex features directly from raw images, making them suitable for wide variety of tasks like image classification, object detection, segmentation etc. We applied transfer learning to initialize the parameters of the proposed deep model. Data augmentation techniques like rotation, translation, reflection and scaling were also applied to prevent overfitting. The proposed CNN model obtained encouraging results, reaching around 97.18% of accuracy on our prepared dataset. The results validate that the proposed method is effective in classifying various types of apple diseases and can be used as a practical tool by farmers.
Deep neural networks (DNNs) have achieved the state of the art performance in numerous fields. However, DNNs need high computation times, and people always expect better performance in a lower computation. Therefore, we study the human somatosensory system and design a neural network (SpinalNet) to achieve higher accuracy with fewer computations. Hidden layers in traditional NNs receive inputs in the previous layer, apply activation function, and then transfer the outcomes to the next layer. In the proposed SpinalNet, each layer is split into three splits: 1) input split, 2) intermediate split, and 3) output split. Input split of each layer receives a part of the inputs. The intermediate split of each layer receives outputs of the intermediate split of the previous layer and outputs of the input split of the current layer. The number of incoming weights becomes significantly lower than traditional DNNs. The SpinalNet can also be used as the fully connected or classification layer of DNN and supports both traditional learning and transfer learning. We observe significant error reductions with lower computational costs in most of the DNNs. Traditional learning on the VGG-5 network with SpinalNet classification layers provided the state-of-the-art (SOTA) performance on QMNIST, Kuzushiji-MNIST, EMNIST (Letters, Digits, and Balanced) datasets. Traditional learning with ImageNet pre-trained initial weights and SpinalNet classification layers provided the SOTA performance on STL-10, Fruits 360, Bird225, and Caltech-101 datasets. The scripts of the proposed SpinalNet are available at the following link: https://github.com/dipuk0506/SpinalNet
There has been a large literature of neural architecture search, but most existing work made use of heuristic rules that largely constrained the search flexibility. In this paper, we first relax these manually designed constraints and enlarge the search space to contain more than $10^{160}$ candidates. In the new space, most existing differentiable search methods can fail dramatically. We then propose a novel algorithm named Gradual One-Level Differentiable Neural Architecture Search (GOLD-NAS) which introduces a variable resource constraint to one-level optimization so that the weak operators are gradually pruned out from the super-network. In standard image classification benchmarks, GOLD-NAS can find a series of Pareto-optimal architectures within a single search procedure. Most of the discovered architectures were never studied before, yet they achieve a nice tradeoff between recognition accuracy and model complexity. We believe the new space and search algorithm can advance the search of differentiable NAS.
Breast cancer is the most common cancer in women. Classification of cancer/non-cancer patients with clinical records requires high sensitivity and specificity for an acceptable diagnosis test. The state-of-the-art classification model - Convolutional Neural Network (CNN), however, cannot be used with clinical data that are represented in 1-D format. CNN has been designed to work on a set of 2-D matrices whose elements show some correlation with neighboring elements such as in image data. Conversely, the data examples represented as a set of 1-D vectors -- apart from the time series data -- cannot be used with CNN, but with other classification models such as Artificial Neural Networks or RandomForest. We have proposed some novel preprocessing methods of data wrangling that transform a 1-D data vector, to a 2-D graphical image with appropriate correlations among the fields to be processed on CNN. We tested our methods on Wisconsin Original Breast Cancer (WBC) and Wisconsin Diagnostic Breast Cancer (WDBC) datasets. To our knowledge, this work is novel on non-image to image data transformation for the non-time series data. The transformed data processed with CNN using VGGnet-16 shows competitive results for the WBC dataset and outperforms other known methods for the WDBC dataset.
The search cost of neural architecture search (NAS) has been largely reduced by weight-sharing methods. These methods optimize a super-network with all possible edges and operations, and determine the optimal sub-network by discretization, \textit{i.e.}, pruning off weak candidates. The discretization process, performed on either operations or edges, incurs significant inaccuracy and thus the quality of the final architecture is not guaranteed. This paper presents discretization-aware architecture search (DA\textsuperscript{2}S), with the core idea being adding a loss term to push the super-network towards the configuration of desired topology, so that the accuracy loss brought by discretization is largely alleviated. Experiments on standard image classification benchmarks demonstrate the superiority of our approach, in particular, under imbalanced target network configurations that were not studied before.
In this work, we evaluate two different image clustering objectives, k-means clustering and correlation clustering, in the context of Triplet Loss induced feature space embeddings. Specifically, we train a convolutional neural network to learn discriminative features by optimizing two popular versions of the Triplet Loss in order to study their clustering properties under the assumption of noisy labels. Additionally, we propose a new, simple Triplet Loss formulation, which shows desirable properties with respect to formal clustering objectives and outperforms the existing methods. We evaluate all three Triplet loss formulations for K-means and correlation clustering on the CIFAR-10 image classification dataset.
In this paper we propose C2G-Net, a pipeline for image classification that exploits the morphological properties of images containing a large number of similar objects like biological cells. C2G-Net consists of two components: (1) Cell2Grid, an image compression algorithm that identifies objects using segmentation and arranges them on a grid, and (2) DeepLNiNo, a CNN architecture with less than 10,000 trainable parameters aimed at facilitating model interpretability. To test the performance of C2G-Net we used multiplex immunohistochemistry images for predicting relapse risk in colon cancer. Compared to conventional CNN architectures trained on raw images, C2G-Net achieved similar prediction accuracy while training time was reduced by 85% and its model was is easier to interpret.
Handwritten digit and letter recognition is one of the oldest and a very important topic in the field of pattern recognition. Handwritten digit and letter recognition poses different problem because of different writing styles, similarity in structure and angle of orientation. Therefore it is very important to find effective method for recognition and classification of digit and letter. Handwritten digit and letter recognition has various applications such as number plate recognition, extracting business card information, bank check processing, postal address processing, passport processing, signature processing etc. This paper propose a method of handwritten digit and letter recognition using feature extraction based on hybrid Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT). These extracted features are passed to K-Nearest Neighbour (KNN) and Support Vector Machine (SVM) classifiers for classification. Standard MNIST and EMNIST letter dataset are used for this experiment. Firstly MNIST digit and EMNSIT letter dataset are binarized and later stray pixels are removed. Features are extracted using hybrid Discrete Wavelet Transform and Discrete Cosine Transform. KNN and SVM classifiers are used for classification purpose. The proposed method was able to obtain a highest accuracy of 97.74% for digit and 89.51% for letter using SVM classifier.
Training neural networks with auxiliary tasks is a common practice for improving the performance on a main task of interest. Two main challenges arise in this multi-task learning setting: (i) designing useful auxiliary tasks; and (ii) combining auxiliary tasks into a single coherent loss. Here, we propose a novel framework, AuxiLearn, that targets both challenges based on implicit differentiation. First, when useful auxiliaries are known, we propose learning a network that combines all losses into a single coherent objective function. This network can learn non-linear interactions between tasks. Second, when no useful auxiliary task is known, we describe how to learn a network that generates a meaningful, novel auxiliary task. We evaluate AuxiLearn in a series of tasks and domains, including image segmentation and learning with attributes in the low data regime, and find that it consistently outperforms competing methods.
This paper proposes a novel domain adaptation algorithm to handle the challenges posed by the satellite and aerial imagery, and demonstrates its effectiveness on the built-up region segmentation problem. Built-up area estimation is an important component in understanding the human impact on the environment, the effect of public policy, and general urban population analysis. The diverse nature of aerial and satellite imagery and lack of labeled data covering this diversity makes machine learning algorithms difficult to generalize for such tasks, especially across multiple domains. On the other hand, due to the lack of strong spatial context and structure, in comparison to the ground imagery, the application of existing unsupervised domain adaptation methods results in the sub-optimal adaptation. We thoroughly study the limitations of existing domain adaptation methods and propose a weakly-supervised adaptation strategy where we assume image-level labels are available for the target domain. More specifically, we design a built-up area segmentation network (as encoder-decoder), with an image classification head added to guide the adaptation. The devised system is able to address the problem of visual differences in multiple satellite and aerial imagery datasets, ranging from high resolution (HR) to very high resolution (VHR). A realistic and challenging HR dataset is created by hand-tagging the 73.4 sq-km of Rwanda, capturing a variety of build-up structures over different terrain. The developed dataset is spatially rich compared to existing datasets and covers diverse built-up scenarios including built-up areas in forests and deserts, mud houses, tin, and colored rooftops. Extensive experiments are performed by adapting from the single-source domain, to segment out the target domain. We achieve high gains ranging 11.6%-52% in IoU over the existing state-of-the-art methods.