Deep neural networks (DNNs) have become the driving force behind recent artificial intelligence (AI) research. An important problem with implementing a neural network is the design of its architecture. Typically, such an architecture is obtained manually by exploring its hyperparameter space and kept fixed during training. This approach is time-consuming and inefficient. Another issue is that modern neural networks often contain millions of parameters, whereas many applications and devices require small inference models. However, efforts to migrate DNNs to such devices typically entail a significant loss of classification accuracy. To address these challenges, we propose a two-step neural network synthesis methodology, called DR+SCANN, that combines two complementary approaches to design compact and accurate DNNs. At the core of our framework is the SCANN methodology that uses three basic architecture-changing operations, namely connection growth, neuron growth, and connection pruning, to synthesize feed-forward architectures with arbitrary structure. SCANN encapsulates three synthesis methodologies that apply a repeated grow-and-prune paradigm to three architectural starting points. DR+SCANN combines the SCANN methodology with dataset dimensionality reduction to alleviate the curse of dimensionality. We demonstrate the efficacy of SCANN and DR+SCANN on various image and non-image datasets. We evaluate SCANN on MNIST and ImageNet benchmarks. In addition, we also evaluate the efficacy of using dimensionality reduction alongside SCANN (DR+SCANN) on nine small to medium-size datasets. We also show that our synthesis methodology yields neural networks that are much better at navigating the accuracy vs. energy efficiency space. This would enable neural network-based inference even on Internet-of-Things sensors.
Representation learning with small labeled data have emerged in many problems, since the success of deep neural networks often relies on the availability of a huge amount of labeled data that is expensive to collect. To address it, many efforts have been made on training sophisticated models with few labeled data in an unsupervised and semi-supervised fashion. In this paper, we will review the recent progresses on these two major categories of methods. A wide spectrum of models will be categorized in a big picture, where we will show how they interplay with each other to motivate explorations of new ideas. We will review the principles of learning the transformation equivariant, disentangled, self-supervised and semi-supervised representations, all of which underpin the foundation of recent progresses. Many implementations of unsupervised and semi-supervised generative models have been developed on the basis of these criteria, greatly expanding the territory of existing autoencoders, generative adversarial nets (GANs) and other deep networks by exploring the distribution of unlabeled data for more powerful representations. We will discuss emerging topics by revealing the intrinsic connections between unsupervised and semi-supervised learning, and propose in future directions to bridge the algorithmic and theoretical gap between transformation equivariance for unsupervised learning and supervised invariance for supervised learning, and unify unsupervised pretraining and supervised finetuning. We will also provide a broader outlook of future directions to unify transformation and instance equivariances for representation learning, connect unsupervised and semi-supervised augmentations, and explore the role of the self-supervised regularization for many learning problems.
The Era of Big Data has forced researchers to explore new distributed solutions for building fuzzy classifiers, which often introduce approximation errors or make strong assumptions to reduce computational and memory requirements. As a result, Big Data classifiers might be expected to be inferior to those designed for standard classification tasks (Small Data) in terms of accuracy and model complexity. To our knowledge, however, there is no empirical evidence to confirm such a conjecture yet. Here, we investigate the extent to which state-of-the-art fuzzy classifiers for Big Data sacrifice performance in favor of scalability. To this end, we carry out an empirical study that compares these classifiers with some of the best performing algorithms for Small Data. Assuming the latter were generally designed for maximizing performance without considering scalability issues, the results of this study provide some intuition around the tradeoff between performance and scalability achieved by current Big Data solutions. Our findings show that, although slightly inferior, Big Data classifiers are gradually catching up with state-of-the-art classifiers for Small data, suggesting that a unified learning algorithm for Big and Small Data might be possible.
With the advent of smartphone indirect ophthalmoscopy, teleophthalmology - the use of specialist ophthalmology assets at a distance from the patient - has experienced a breakthrough, promising enormous benefits especially for healthcare in distant, inaccessible or opthalmologically underserved areas, where specialists are either unavailable or too few in number. However, accurate teleophthalmology requires high-quality ophthalmoscopic imagery. This paper considers three feature families - statistical metrics, gradient-based metrics and wavelet transform coefficient derived indicators - as possible metrics to identify unsharp or blurry images. By using standard machine learning techniques, the suitability of these features for image quality assessment is confirmed, albeit on a rather small data set. With the increased availability and decreasing cost of digital ophthalmoscopy on one hand and the increased prevalence of diabetic retinopathy worldwide on the other, creating tools that can determine whether an image is likely to be diagnostically suitable can play a significant role in accelerating and streamlining the teleophthalmology process. This paper highlights the need for more research in this area, including the compilation of a diverse database of ophthalmoscopic imagery, annotated with quality markers, to train the Point of Acquisition error detection algorithms of the future.
The annotated medical images are usually expensive to be collected. This paper proposes a deep learning method on small data to classify Common Imaging Signs of Lung diseases (CISL) in computed tomography (CT) images. We explore both the real data and the data generated by Generative Adversarial Network (GAN) to improve the reliability and the generalization of learning. First, we use GAN to generate a large number of CISLs from small annotated data, which are difficult to be distinguished from real counterparts. These generated samples are used to pre-train a Convolutional Neural Network (CNN) for classifying CISLs. Second, we fine-tune the CNN classification model with real data. Experiments were conducted on the LISS database of CISLs. We successfully convinced radiologists that our generated CISLs samples were real for 56.7% of our experiments. The pre-trained CNN model achieves 88.4% of mean accuracy of binary classification, and after fine-tuning, the mean accuracy is significantly increased to 95.0%. For multi-classification of all types of CISLs and normal tissues, through the two stages of training, the mean accuracy, sensitivity and specificity are up to about 91.83%, 92.73% and 99.0%, respectively. To our knowledge, this is the best result achieved on the LISS database, which demonstrates that the proposed method is effective and promising for fulfilling deep learning on small data.
We propose a novel data-driven method to learn a mixture of multiple kernels with random features that is certifiabaly robust against adverserial inputs. Specifically, we consider a distributionally robust optimization of the kernel-target alignment with respect to the distribution of training samples over a distributional ball defined by the Kullback-Leibler (KL) divergence. The distributionally robust optimization problem can be recast as a min-max optimization whose objective function includes a log-sum term. We develop a mini-batch biased stochastic primal-dual proximal method to solve the min-max optimization. To debias the minibatch algorithm, we use the Gumbel perturbation technique to estimate the log-sum term. We establish theoretical guarantees for the performance of the proposed multiple kernel learning method. In particular, we prove the consistency, asymptotic normality, stochastic equicontinuity, and the minimax rate of the empirical estimators. In addition, based on the notion of Rademacher and Gaussian complexities, we establish distributionally robust generalization bounds that are tighter than previous known bounds. More specifically, we leverage matrix concentration inequalities to establish distributionally robust generalization bounds. We validate our kernel learning approach for classification with the kernel SVMs on synthetic dataset generated by sampling multvariate Gaussian distributions with differernt variance structures. We also apply our kernel learning approach to the MNIST data-set and evaluate its robustness to perturbation of input images under different adversarial models. More specifically, we examine the robustness of the proposed kernel model selection technique against FGSM, PGM, C\&W, and DDN adversarial perturbations, and compare its performance with alternative state-of-the-art multiple kernel learning paradigms.
Learning predictive models from small high-dimensional data sets is a key problem in high-dimensional statistics. Expert knowledge elicitation can help, and a strong line of work focuses on directly eliciting informative prior distributions for parameters. This either requires considerable statistical expertise or is laborious, as the emphasis has been on accuracy and not on efficiency of the process. Another line of work queries about importance of features one at a time, assuming them to be independent and hence missing covariance information. In contrast, we propose eliciting expert knowledge about pairwise feature similarities, to borrow statistical strength in the predictions, and using sequential decision making techniques to minimize the effort of the expert. Empirical results demonstrate improvement in predictive performance on both simulated and real data, in high-dimensional linear regression tasks, where we learn the covariance structure with a Gaussian process, based on sequential elicitation.
The human mind is still an unknown process of neuroscience in many aspects. Nevertheless, for decades the scientific community has proposed computational models that try to simulate their parts, specific applications, or their behavior in different situations. The most complete model in this line is undoubtedly the LIDA model, proposed by Stan Franklin with the aim of serving as a generic computational architecture for several applications. The present project is inspired by the LIDA model to apply it to the process of movie recommendation, the model called MIRA (Movie Intelligent Recommender Agent) presented percentages of precision similar to a traditional model when submitted to the same assay conditions. Moreover, the proposed model reinforced the precision indexes when submitted to tests with volunteers, proving once again its performance as a cognitive model, when executed with small data volumes. Considering that the proposed model achieved a similar behavior to the traditional models under conditions expected to be similar for natural systems, it can be said that MIRA reinforces the applicability of LIDA as a path to be followed for the study and generation of computational agents inspired by neural behaviors.
Named entity recognition (NER) is an important task in NLP, which is all the more challenging in conversational domain with their noisy facets. Moreover, conversational texts are often available in limited amount, making supervised tasks infeasible. To learn from small data, strong inductive biases are required. Previous work relied on hand-crafted features to encode these biases until transfer learning emerges. Here, we explore a transfer learning method, namely language model pretraining, on NER task in Indonesian conversational texts. We utilize large unlabeled data (generic domain) to be transferred to conversational texts, enabling supervised training on limited in-domain data. We report two transfer learning variants, namely supervised model fine-tuning and unsupervised pretrained LM fine-tuning. Our experiments show that both variants outperform baseline neural models when trained on small data (100 sentences), yielding an absolute improvement of 32 points of test F1 score. Furthermore, we find that the pretrained LM encodes part-of-speech information which is a strong predictor for NER.
The automated construction of coarse-grained models represents a pivotal component in computer simulation of physical systems and is a key enabler in various analysis and design tasks related to uncertainty quantification. Pertinent methods are severely inhibited by the high-dimension of the parametric input and the limited number of training input/output pairs that can be generated when computationally demanding forward models are considered. Such cases are frequently encountered in the modeling of random heterogeneous media where the scale of the microstructure necessitates the use of high-dimensional random vectors and very fine discretizations of the governing equations. The present paper proposes a probabilistic Machine Learning framework that is capable of operating in the presence of Small Data by exploiting aspects of the physical structure of the problem as well as contextual knowledge. As a result, it can perform comparably well under extrapolative conditions. It unifies the tasks of dimensionality and model-order reduction through an encoder-decoder scheme that simultaneously identifies a sparse set of salient lower-dimensional microstructural features and calibrates an inexpensive, coarse-grained model which is predictive of the output. Information loss is accounted for and quantified in the form of probabilistic predictive estimates. The learning engine is based on Stochastic Variational Inference. We demonstrate how the variational objectives can be used not only to train the coarse-grained model, but also to suggest refinements that lead to improved predictions.
In this paper we develop a Bayesian optimization based hyperparameter tuning framework inspired by statistical learning theory for classifiers. We utilize two key facts from PAC learning theory; the generalization bound will be higher for a small subset of data compared to the whole, and the highest accuracy for a small subset of data can be achieved with a simple model. We initially tune the hyperparameters on a small subset of training data using Bayesian optimization. While tuning the hyperparameters on the whole training data, we leverage the insights from the learning theory to seek more complex models. We realize this by using directional derivative signs strategically placed in the hyperparameter search space to seek a more complex model than the one obtained with small data. We demonstrate the performance of our method on the tasks of tuning the hyperparameters of several machine learning algorithms.
The great success achieved by deep neural networks attracts increasing attention from the manufacturing and healthcare communities. However, the limited availability of data and high costs of data collection are the major challenges for the applications in those fields. We propose in this work AISEL, an active image synthesis method for efficient labeling to improve the performance of the small-data learning tasks. Specifically, a complementary AISEL dataset is generated, with labels actively acquired via a physics-based method to incorporate underlining physical knowledge at hand. An important component of our AISEL method is the bidirectional generative invertible network (GIN), which can extract interpretable features from the training images and generate physically meaningful virtual images. Our AISEL method then efficiently samples virtual images not only further exploits the uncertain regions, but also explores the entire image space. We then discuss the interpretability of GIN both theoretically and experimentally, demonstrating clear visual improvements over the benchmarks. Finally, we demonstrate the effectiveness of our AISEL framework on aortic stenosis application, in which our method lower the labeling cost by $90\%$ while achieving a $15\%$ improvement in prediction accuracy.
Surrogate modeling and uncertainty quantification tasks for PDE systems are most often considered as supervised learning problems where input and output data pairs are used for training. The construction of such emulators is by definition a small data problem which poses challenges to deep learning approaches that have been developed to operate in the big data regime. Even in cases where such models have been shown to have good predictive capability in high dimensions, they fail to address constraints in the data implied by the PDE model. This paper provides a methodology that incorporates the governing equations of the physical model in the loss/likelihood functions. The resulting physics-constrained, deep learning models are trained without any labeled data (e.g. employing only input data) and provide comparable predictive responses with data-driven models while obeying the constraints of the problem at hand. This work employs a convolutional encoder-decoder neural network approach as well as a conditional flow-based generative model for the solution of PDEs, surrogate model construction, and uncertainty quantification tasks. The methodology is posed as a minimization problem of the reverse Kullback-Leibler (KL) divergence between the model predictive density and the reference conditional density, where the later is defined as the Boltzmann-Gibbs distribution at a given inverse temperature with the underlying potential relating to the PDE system of interest. The generalization capability of these models to out-of-distribution input is considered. Quantification and interpretation of the predictive uncertainty is provided for a number of problems.
On visual analytics applications, the concept of putting the user on the loop refers to the ability to replace heuristics by user knowledge on machine learning and data mining tasks. On supervised tasks, the user engagement occurs via the manipulation of the training data. However, on unsupervised tasks, the user involvement is limited to changes in the algorithm parametrization or the input data representation, also known as features. Depending on the application domain, different types of features can be extracted from the raw data. Therefore, the result of unsupervised algorithms heavily depends on the type of employed feature. Since there is no perfect feature extractor, combining different features have been explored in a process called feature fusion. The feature fusion is straightforward when the machine learning or data mining task has a cost function. However, when such a function does not exist, user support for combination needs to be provided otherwise the process is impractical. In this paper, we present a novel feature fusion approach that uses small data samples to allows users not only to effortless control the combination of different feature sets but also to interpret the attained results. The effectiveness of our approach is confirmed by a comprehensive set of qualitative and quantitative tests, opening up different possibilities of user-guided analytical scenarios not covered yet. The ability of our approach to providing real-time feedback for the feature fusion is exploited on the context of unsupervised clustering techniques, where the composed groups reflect the semantics of the feature combination.
Diffusion MRI is the modality of choice to study alterations of white matter. In past years, various works have used diffusion MRI for automatic classification of AD. However, classification performance obtained with different approaches is difficult to compare and these studies are also difficult to reproduce. In the present paper, we first extend a previously proposed framework to diffusion MRI data for AD classification. Specifically, we add: conversion of diffusion MRI ADNI data into the BIDS standard and pipelines for diffusion MRI preprocessing and feature extraction. We then apply the framework to compare different components. First, FS has a positive impact on classification results: highest balanced accuracy (BA) improved from 0.76 to 0.82 for task CN vs AD. Secondly, voxel-wise features generally gives better performance than regional features. Fractional anisotropy (FA) and mean diffusivity (MD) provided comparable results for voxel-wise features. Moreover, we observe that the poor performance obtained in tasks involving MCI were potentially caused by the small data samples, rather than by the data imbalance. Furthermore, no extensive classification difference exists for different degree of smoothing and registration methods. Besides, we demonstrate that using non-nested validation of FS leads to unreliable and over-optimistic results: 0.05 up to 0.40 relative increase in BA. Lastly, with proper FR and FS, the performance of diffusion MRI features is comparable to that of T1w MRI. All the code of the framework and the experiments are publicly available: general-purpose tools have been integrated into the Clinica software package (www.clinica.run) and the paper-specific code is available at: https://github.com/aramis-lab/AD-ML.
Sublingual vein is commonly used to diagnose the health status. The width of main sublingual veins gives information of the blood circulation. Therefore, it is necessary to segment the main sublingual veins from the tongue automatically. In general, the dataset in the medical field is small, which is a challenge for training the deep learning model. In order to train the model with a small data set, the proposed method for automatically segmenting the sublingual veins is to re-train U-net model with different sets of the limited number of labels for the same training images. With pre-knowledge of the segmentation, the loss of the trained model will be convergence easier. To improve the performance of the segmentation further, a novel strategy of data augmentation was utilized. The operation for masking output of the model with the input was randomly switched on or switched off in each training step. This approach will force the model to learn the contrast invariance and avoid overfitting. Images of dataset were taken with the developed device using eight near infrared LEDs. The final segmentation results were evaluated on the validation dataset by the IoU metric.
In our study, we demonstrate the synergy effect between convolutional neural networks and the multiplicity of SMILES. The model we propose, the so-called Convolutional Neural Fingerprint (CNF) model, reaches the accuracy of traditional descriptors such as Dragon (Mauri et al. [22]), RDKit (Landrum [18]), CDK2 (Willighagen et al. [43]) and PyDescriptor (Masand and Rastija [20]). Moreover the CNF model generally performs better than highly fine-tuned traditional descriptors, especially on small data sets, which is of great interest for the chemical field where data sets are generally small due to experimental costs, the availability of molecules or accessibility to private databases. We evaluate the CNF model along with SMILES augmentation during both training and testing. To the best of our knowledge, this is the first time that such a methodology is presented. We show that using the multiplicity of SMILES during training acts as a regulariser and therefore avoids overfitting and can be seen as ensemble learning when considered for testing.
In the following short article we adapt a new and popular machine learning model for inference on medical data sets. Our method is based on the Variational AutoEncoder (VAE) framework that we adapt to survival analysis on small data sets with missing values. In our model, the true health status appears as a set of latent variables that affects the observed covariates and the survival chances. We show that this flexible model allows insightful decision-making using a predicted distribution and outperforms a classic survival analysis model.
The task of localizing and categorizing objects in medical images often remains formulated as a semantic segmentation problem. This approach, however, only indirectly solves the coarse localization task by predicting pixel-level scores, requiring ad-hoc heuristics when mapping back to object-level scores. State-of-the-art object detectors on the other hand, allow for individual object scoring in an end-to-end fashion, while ironically trading in the ability to exploit the full pixel-wise supervision signal. This can be particularly disadvantageous in the setting of medical image analysis, where data sets are notoriously small. In this paper, we propose Retina U-Net, a simple architecture, which naturally fuses the Retina Net one-stage detector with the U-Net architecture widely used for semantic segmentation in medical images. The proposed architecture recaptures discarded supervision signals by complementing object detection with an auxiliary task in the form of semantic segmentation without introducing the additional complexity of previously proposed two-stage detectors. We evaluate the importance of full segmentation supervision on two medical data sets, provide an in-depth analysis on a series of toy experiments and show how the corresponding performance gain grows in the limit of small data sets. Retina U-Net yields strong detection performance only reached by its more complex two-staged counterparts. Our framework including all methods implemented for operation on 2D and 3D images is available at github.com/pfjaeger/medicaldetectiontoolkit.
In this paper, we explore creative image generation constrained by small data. To partially automate the creation of cartoon sketches consistent with a specific designer's style, where acquiring a very large original image set is impossible or cost prohibitive, we exploit domain specific knowledge for a huge reduction in original image requirements, creating an effectively infinite number of cartoon giraffes from just nine original drawings. We introduce "expressive augmentations" for cartoon sketches, mathematical transformations that create broad domain appropriate variation, far beyond the usual affine transformations, and we show that chained GANs models trained on the temporal stages of drawing or "complexity layers" can effectively add character appropriate details and finish new drawings in the designer's style. We discuss the application of these tools in design processes for textiles, graphics, architectural elements and interior design.