For many NLP applications, such as question answering and summarisation, the goal is to select the best solution from a large space of candidates to meet a particular user's needs. To address the lack of user-specific training data, we propose an interactive text ranking approach that actively selects pairs of candidates, from which the user selects the best. Unlike previous strategies, which attempt to learn a ranking across the whole candidate space, our method employs Bayesian optimisation to focus the user's labelling effort on high quality candidates and integrates prior knowledge in a Bayesian manner to cope better with small data scenarios. We apply our method to community question answering (cQA) and extractive summarisation, finding that it significantly outperforms existing interactive approaches. We also show that the ranking function learned by our method is an effective reward function for reinforcement learning, which improves the state of the art for interactive summarisation.
Bayesian neural networks (BNN) are the probabilistic model that combines the strengths of both neural network (NN) and stochastic processes. As a result, BNN can combat overfitting and perform well in applications where data is limited. Earthquake rupture study is such a problem where data is insufficient, and scientists have to rely on many trial and error numerical or physical models. Lack of resources and computational expenses, often, it becomes hard to determine the reasons behind the earthquake rupture. In this work, a BNN has been used (1) to combat the small data problem and (2) to find out the parameter combinations responsible for earthquake rupture and (3) to estimate the uncertainty associated with earthquake rupture. Two thousand rupture simulations are used to train and test the model. A simple 2D rupture geometry is considered where the fault has a Gaussian geometric heterogeneity at the center, and eight parameters vary in each simulation. The test F1-score of BNN (0.8334), which is 2.34% higher than plain NN score. Results show that the parameters of rupture propagation have higher uncertainty than the rupture arrest. Normal stresses play a vital role in determining rupture propagation and are also the highest source of uncertainty, followed by the dynamic friction coefficient. Shear stress has a moderate role, whereas the geometric features such as the width and height of the fault are least significant and uncertain.
Humans excel in continuously learning with small data without forgetting how to solve old problems. However, neural networks require large datasets to compute latent representations across different tasks while minimizing a loss function. For example, a natural language understanding (NLU) system will often deal with emerging entities during its deployment as interactions with users in realistic scenarios will generate new and infrequent names, events, and locations. Here, we address this scenario by introducing an RL trainable controller that disentangles the representation learning of a neural encoder from its memory management role. Our proposed solution is straightforward and simple: we train a controller to execute an optimal sequence of reading and writing operations on an external memory with the goal of leveraging diverse activations from the past and provide accurate predictions. Our approach is named Learning to Control (LTC) and allows few-shot learning with two degrees of memory plasticity. We experimentally show that our system obtains accurate results for few-shot learning of entity recognition in the Stanford Task-Oriented Dialogue dataset.
The quality of visual input is very important for both human and machine perception. Consequently many processing techniques exist that deal with different distortions. Usually image processing is applied freely and lacks redundancy regarding safety. We propose a novel image comparison method called the Combined Distances of Contrast Distributions (CD2) to protect against errors that arise during processing. Based on the distribution of image contrasts a new reduced-reference image quality assessment (IQA) method is introduced. By combining various distance functions excellent performance on IQA benchmarks is achieved with only a small data and computation overhead.
Deep learning has achieved impressive prediction performance in the field of sequence learning recently. Dissolved oxygen prediction, as a kind of time-series forecasting, is suitable for this technique. Although many researchers have developed hybrid models or variant models based on deep learning techniques, there is no comprehensive and sound comparison among the deep learning models in this field currently. Plus, most previous studies focused on one-step forecasting by using a small data set. As the convenient access to high-frequency data, this paper compares multi-step deep learning forecasting by using walk-forward validation. Specifically, we test Convolutional Neural Network (CNN), Temporal Convolutional Network (TCN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional Recurrent Neural Network (BiRNN) based on the real-time data recorded automatically at a fixed observation point in the Yangtze River from 2012 to 2016. By comparing the average accumulated statistical metrics of root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination in each time step, We find for multi-step time series forecasting, the average performance of each time step does not decrease linearly. GRU outperforms other models with significant advantages.
As a mental disorder progresses, it may affect brain structure, but brain function expressed in brain dynamics is affected much earlier. Capturing the moment when brain dynamics express the disorder is crucial for early diagnosis. The traditional approach to this problem via training classifiers either proceeds from handcrafted features or requires large datasets to combat the $m>>n$ problem when a high dimensional fMRI volume only has a single label that carries learning signal. Large datasets may not be available for a study of each disorder, or rare disorder types or sub-populations may not warrant for them. In this paper, we demonstrate a self-supervised pre-training method that enables us to pre-train directly on fMRI dynamics of healthy control subjects and transfer the learning to much smaller datasets of schizophrenia. Not only we enable classification of disorder directly based on fMRI dynamics in small data but also significantly speed up the learning when possible. This is encouraging evidence of informative transfer learning across datasets and diagnostic categories.
In drug-discovery-related tasks such as virtual screening, machine learning is emerging as a promising way to predict molecular properties. Conventionally, molecular fingerprints (numerical representations of molecules) are calculated through rule-based algorithms that map molecules to a sparse discrete space. However, these algorithms perform poorly for shallow prediction models or small datasets. To address this issue, we present SMILES Transformer. Inspired by Transformer and pre-trained language models from natural language processing, SMILES Transformer learns molecular fingerprints through unsupervised pre-training of the sequence-to-sequence language model using a huge corpus of SMILES, a text representation system for molecules. We performed benchmarks on 10 datasets against existing fingerprints and graph-based methods and demonstrated the superiority of the proposed algorithms in small-data settings where pre-training facilitated good generalization. Moreover, we define a novel metric to concurrently measure model accuracy and data efficiency.
Automatic question generation aims at the generation of questions from a context, with the corresponding answers being sub-spans of the given passage. Whereas, most of the methods mostly rely on heuristic rules to generate questions, more recently also neural network approaches have been proposed. In this work, we propose a variant of the self-attention Transformer network architectures model to generate meaningful and diverse questions. To this end, we propose an easy to use model consisting of the conjunction of the Transformer decoder GPT-2 model with Transformer encoder BERT for the downstream task for question answering. The model is trained in an end-to-end fashion, where the language model is trained to produce a question-answer-aware input representation that facilitates to generate an answer focused question. Our result of neural question generation from text on the SQuAD 1.1 dataset suggests that our method can produce semantically correct and diverse questions. Additionally, we assessed the performance of our proposed method for the downstream task of question answering. The analysis shows that our proposed generation & answering collaboration framework relatively improves both tasks and is particularly powerful in the semi-supervised setup. The results further suggest a robust and comparably lean pipeline facilitating question generation in the small-data regime.
In the majority of molecular optimization tasks, predictive machine learning (ML) models are limited due to the unavailability and cost of generating big experimental datasets on the specific task. To circumvent this limitation, ML models are trained on big theoretical datasets or experimental indicators of molecular suitability that are either publicly available or inexpensive to acquire. These approaches produce a set of candidate molecules which have to be ranked using limited experimental data or expert knowledge. Under the assumption that structure is related to functionality, here we use a molecular fragment-based graphical autoencoder to generate unique structural fingerprints to efficiently search through the candidate set. We demonstrate that fragment-based graphical autoencoding reduces the error in predicting physical characteristics such as the solubility and partition coefficient in the small data regime compared to other extended circular fingerprints and string based approaches. We further demonstrate that this approach is capable of providing insight into real world molecular optimization problems, such as searching for stabilization additives in organic semiconductors by accurately predicting 92% of test molecules given 69 training examples. This task is a model example of black box molecular optimization as there is minimal theoretical and experimental knowledge to accurately predict the suitability of the additives.
This paper tackles the challenge presented by small-data to the task of Bayesian inference. A novel methodology, based on manifold learning and manifold sampling, is proposed for solving this computational statistics problem under the following assumptions: 1) neither the prior model nor the likelihood function are Gaussian and neither can be approximated by a Gaussian measure; 2) the number of functional input (system parameters) and functional output (quantity of interest) can be large; 3) the number of available realizations of the prior model is small, leading to the small-data challenge typically associated with expensive numerical simulations; the number of experimental realizations is also small; 4) the number of the posterior realizations required for decision is much larger than the available initial dataset. The method and its mathematical aspects are detailed. Three applications are presented for validation: The first two involve mathematical constructions aimed to develop intuition around the method and to explore its performance. The third example aims to demonstrate the operational value of the method using a more complex application related to the statistical inverse identification of the non-Gaussian matrix-valued random elasticity field of a damaged biological tissue (osteoporosis in a cortical bone) using ultrasonic waves.
The utility of linguistic annotation in neural machine translation seemed to had been established in past papers. The experiments were however limited to recurrent sequence-to-sequence architectures and relatively small data settings. We focus on the state-of-the-art Transformer model and use comparably larger corpora. Specifically, we try to promote the knowledge of source-side syntax using multi-task learning either through simple data manipulation techniques or through a dedicated model component. In particular, we train one of Transformer attention heads to produce source-side dependency tree. Overall, our results cast some doubt on the utility of multi-task setups with linguistic information. The data manipulation techniques, recommended in previous works, prove ineffective in large data settings. The treatment of self-attention as dependencies seems much more promising: it helps in translation and reveals that Transformer model can very easily grasp the syntactic structure. An important but curious result is, however, that identical gains are obtained by using trivial "linear trees" instead of true dependencies. The reason for the gain thus may not be coming from the added linguistic knowledge but from some simpler regularizing effect we induced on self-attention matrices.
Early detection of melanoma is difficult for the human eye but a crucial step towards reducing its death rate. Computerized detection of these melanoma and other skin lesions is necessary. The central research question in this paper is "How to segment skin lesion images using a neural network with low available data?". This question is divided into three sub questions regarding best performing network structure, training data and training method. First theory associated with these questions is discussed. Literature states that U-net CNN structures have excellent performances on the segmentation task, more training data increases network performance and utilizing transfer learning enables networks to generalize to new data better. To validate these findings in the literature two experiments are conducted. The first experiment trains a network on data sets of different size. The second experiment proposes twelve network structures and trains them on the same data set. The experimental results support the findings in the literature. The FCN16 and FCN32 networks perform best in the accuracy, intersection over union and mean BF1 Score metric. Concluding from these results the skin lesion segmentation network is a fully convolutional structure with a skip architecture and an encoder depth of either one or two. Weights of this network should be initialized using transfer learning from the pre trained VGG16 network. Training data should be cropped to reduce complexity and augmented during training to reduce the likelihood of overfitting.
This paper proposes a deep neural network approach for predicting multiphase flow in heterogeneous domains with high computational efficiency. The deep neural network model is able to handle permeability heterogeneity in high dimensional systems, and can learn the interplay of viscous, gravity, and capillary forces from small data sets. Using the example of carbon dioxide (CO2) storage, we demonstrate that the model can generate highly accurate predictions of a CO2 saturation distribution given a permeability field, injection duration, injection rate, and injection location. The trained neural network model has an excellent ability to interpolate and to a limited extent, the ability to extrapolate beyond the training data ranges. To improve the prediction accuracy when the neural network model needs to extrapolate, we propose a transfer learning (fine-tuning) procedure that can quickly teach the neural network model new information without going through massive data collection and retraining. Based on this trained neural network model, a web-based tool is provided that allows users to perform CO2-water multiphase flow calculations online. With the tools provided in this paper, the deep neural network approach can provide a computationally efficient substitute for repetitive forward multiphase flow simulations, which can be adopted to the context of history matching and uncertainty quantification.
This paper explores object detection in the small data regime, where only a limited number of annotated bounding boxes are available due to data rarity and annotation expense. This is a common challenge today with machine learning being applied to many new tasks where obtaining training data is more challenging, e.g. in medical images with rare diseases that doctors sometimes only see once in their life-time. In this work we explore this problem from a generative modeling perspective by learning to generate new images with associated bounding boxes, and using these for training an object detector. We show that simply training previously proposed generative models does not yield satisfactory performance due to them optimizing for image realism rather than object detection accuracy. To this end we develop a new model with a novel unrolling mechanism that jointly optimizes the generative model and a detector such that the generated images improve the performance of the detector. We show this method outperforms the state of the art on two challenging datasets, disease detection and small data pedestrian detection, improving the average precision on NIH Chest X-ray by a relative 20% and localization accuracy by a relative 50%.
In this paper, we introduce a new gait segmentation method based on accelerometer data and develop a new distance function between two time series, showing novel and effectiveness in simultaneously identifying user and adversary. Comparing with the normally used Neural Network methods, our approaches use geometric features to extract walking cycles more precisely and employ a new similarity metric to conduct user-adversary identification. This new technology for simultaneously identify user and adversary contributes to cybersecurity beyond user-only identification. In particular, the new technology is being applied to cell phone recorded walking data and performs an accuracy of $98.79\%$ for 6 classes classification (user-adversary identification) and $99.06\%$ for binary classification (user only identification). In addition to walking signal, our approach works on walking up, walking down and mixed walking signals. This technology is feasible for both large and small data set, overcoming the current challenges facing to Neural Networks such as tuning large number of hyper-parameters for large data sets and lacking of training data for small data sets. In addition, the new distance function developed here can be applied in any signal analysis.
We have witnessed rapid advances in both face presentation attack models and presentation attack detection (PAD) in recent years. Compared to widely studied 2D face presentation attacks (e.g. printed photos and video replays), 3D face presentation attacks are more challenging because face recognition systems (FRS) is more easily confused by the 3D characteristics of materials similar to real faces. Existing 3D face spoofing databases, mostly based on 3D facial masks, are restricted to small data size and suffer from poor authenticity due to the difficulty and expense of mask production. In this work, we introduce a wax figure face database (WFFD) as a novel and super-realistic 3D face presentation attack. This database contains 2300 image pairs (totally 4600) and 745 subjects including both real and wax figure faces with high diversity from online collections. On one hand, our experiments have demonstrated the spoofing potential of WFFD on three popular FRSs. On the other hand, we have developed a multi-feature voting scheme for wax figure face detection (anti-spoofing), which combines three discriminative features at the decision level. The proposed detection method was compared against several face PAD approaches and found to outperform other competing methods. Surprisingly, our fusion-based detection method achieves an Average Classification Error Rate (ACER) of 11.73\% on the WFFD database, which is even better than human-based detection.
Gaussian Processes (GPs) are known to provide accurate predictions and uncertainty estimates even with small amounts of labeled data by capturing similarity between data points through their kernel function. However traditional GP kernels are not very effective at capturing similarity between high dimensional data points. Neural networks can be used to learn good representations that encode intricate structures in high dimensional data, and can be used as inputs to the GP kernel. However the huge data requirement of neural networks makes this approach ineffective in small data settings. To solves the conflicting problems of representation learning and data efficiency, we propose to learn deep kernels on probabilistic embeddings by using a probabilistic neural network. Our approach maps high-dimensional data to a probability distribution in a low dimensional subspace and then computes a kernel between these distributions to capture similarity. To enable end-to-end learning, we derive a functional gradient descent procedure for training the model. Experiments on a variety of datasets show that our approach outperforms the state-of-the-art in GP kernel learning in both supervised and semi-supervised settings. We also extend our approach to other small-data paradigms such as few-shot classification where it outperforms previous approaches on mini-Imagenet and CUB datasets.
Kernel learning methods are among the most effective learning methods and have been vigorously studied in the past decades. However, when tackling with complicated tasks, classical kernel methods are not flexible or "rich" enough to describe the data and hence could not yield satisfactory performance. In this paper, via Random Fourier Features (RFF), we successfully incorporate the deep architecture into kernel learning, which significantly boosts the flexibility and richness of kernel machines while keeps kernels' advantage of pairwise handling small data. With RFF, we could establish a deep structure and make every kernel in RFF layers could be trained end-to-end. Since RFF with different distributions could represent different kernels, our model has the capability of finding suitable kernels for each layer, which is much more flexible than traditional kernel-based methods where the kernel is pre-selected. This fact also helps yield a more sophisticated kernel cascade connection in the architecture. On small datasets (less than 1000 samples), for which deep learning is generally not suitable due to overfitting, our method achieves superior performance compared to advanced kernel methods. On large-scale datasets, including non-image and image classification tasks, our method also has competitive performance.
Efficient quality control is inevitable in the manufacturing of light-emitting diodes (LEDs). Because defective LED chips may be traced back to different causes, a time and cost-intensive electrical and optical contact measurement is employed. Fast photoluminescence measurements, on the other hand, are commonly used to detect wafer separation damages but also hold the potential to enable an efficient detection of all kinds of defective LED chips. On a photoluminescence image, every pixel corresponds to an LED chip's brightness after photoexcitation, revealing performance information. But due to unevenly distributed brightness values and varying defect patterns, photoluminescence images are not yet employed for a comprehensive defect detection. In this work, we show that fully convolutional networks can be used for chip-wise defect detection, trained on a small data-set of photoluminescence images. Pixel-wise labels allow us to classify each and every chip as defective or not. Being measurement-based, labels are easy to procure and our experiments show that existing discrepancies between training images and labels do not hinder network training. Using weighted loss calculation, we were able to equalize our highly unbalanced class categories. Due to the consistent use of skip connections and residual shortcuts, our network is able to predict a variety of structures, from extensive defect clusters up to single defective LED chips.
We show how complexity theory can be introduced in machine learning to help bring together apparently disparate areas of current research. We show that this new approach requires less training data and is more generalizable as it shows greater resilience to random attacks. We investigate the shape of the discrete algorithmic space when performing regression or classification using a loss function parametrized by algorithmic complexity, demonstrating that the property of differentiation is not necessary to achieve results similar to those obtained using differentiable programming approaches such as deep learning. In doing so we use examples which enable the two approaches to be compared (small, given the computational power required for estimations of algorithmic complexity). We find and report that (i) machine learning can successfully be performed on a non-smooth surface using algorithmic complexity; (ii) that parameter solutions can be found using an algorithmic-probability classifier, establishing a bridge between a fundamentally discrete theory of computability and a fundamentally continuous mathematical theory of optimization methods; (iii) a formulation of an algorithmically directed search technique in non-smooth manifolds can be defined and conducted; (iv) exploitation techniques and numerical methods for algorithmic search to navigate these discrete non-differentiable spaces can be performed; in application of the (a) identification of generative rules from data observations; (b) solutions to image classification problems more resilient against pixel attacks compared to neural networks; (c) identification of equation parameters from a small data-set in the presence of noise in continuous ODE system problem, (d) classification of Boolean NK networks by (1) network topology, (2) underlying Boolean function, and (3) number of incoming edges.