The speckle pattern produced when a laser is scattered by a disordered medium has recently been shown to give a surprisingly accurate or broadband measurement of wavelength. Here it is shown that deep learning is an ideal approach to analyse wavelength variations using a speckle wavemeter due to its ability to identify trends and overcome low signal to noise ratio in complex datasets. This combination enables wavelength measurement at high precision over a broad operating range in a single step, with a remarkable capability to reject instrumental and environmental noise, which has not been possible with previous approaches. It is demonstrated that the noise rejection capabilities of deep learning provide attometre-scale wavelength precision over an operating range from 488 nm to 976 nm. This dynamic range is six orders of magnitude beyond the state of the art.
The growing application of deep neural networks in safety-critical domains makes the analysis of faults that occur in such systems of enormous importance. In this paper we introduce a large taxonomy of faults in deep learning (DL) systems. We have manually analysed 1059 artefacts gathered from GitHub commits and issues of projects that use the most popular DL frameworks (TensorFlow, Keras and PyTorch) and from related Stack Overflow posts. Structured interviews with 20 researchers and practitioners describing the problems they have encountered in their experience have enriched our taxonomy with a variety of additional faults that did not emerge from the other two sources. Our final taxonomy was validated with a survey involving an additional set of 21 developers, confirming that almost all fault categories (13/15) were experienced by at least 50% of the survey participants.
Neural networks are becoming more and more popular for the analysis of physiological time-series. The most successful deep learning systems in this domain combine convolutional and recurrent layers to extract useful features to model temporal relations. Unfortunately, these recurrent models are difficult to tune and optimize. In our experience, they often require task-specific modifications, which makes them challenging to use for non-experts. We propose U-Time, a fully feed-forward deep learning approach to physiological time series segmentation developed for the analysis of sleep data. U-Time is a temporal fully convolutional network based on the U-Net architecture that was originally proposed for image segmentation. U-Time maps sequential inputs of arbitrary length to sequences of class labels on a freely chosen temporal scale. This is done by implicitly classifying every individual time-point of the input signal and aggregating these classifications over fixed intervals to form the final predictions. We evaluated U-Time for sleep stage classification on a large collection of sleep electroencephalography (EEG) datasets. In all cases, we found that U-Time reaches or outperforms current state-of-the-art deep learning models while being much more robust in the training process and without requiring architecture or hyperparameter adaptation across tasks.
The significant computational requirements of deep learning present a major bottleneck for its large-scale adoption on hardware-constrained IoT-devices. Here, we envision a new paradigm called EdgeAI to address major impediments associated with deploying deep networks at the edge. Specifically, we discuss the existing directions in computation-aware deep learning and describe two new challenges in the IoT era: (1) Data-independent deployment of learning, and (2) Communication-aware distributed inference. We further present new directions from our recent research to alleviate the latter two challenges. Overcoming these challenges is crucial for rapid adoption of learning on IoT-devices in order to truly enable EdgeAI.
The ever-increasing number of Internet of Things (IoT) devices has created a new computing paradigm, called edge computing, where most of the computations are performed at the edge devices, rather than on centralized servers. An edge device is an electronic device that provides connections to service providers and other edge devices; typically, such devices have limited resources. Since edge devices are resource-constrained, the task of launching algorithms, methods, and applications onto edge devices is considered to be a significant challenge. In this paper, we discuss one of the most widely used machine learning methods, namely, Deep Learning (DL) and offer a short survey on the recent approaches used to map DL onto the edge computing paradigm. We also provide relevant discussions about selected applications that would greatly benefit from DL at the edge.
While deep learning has received a surge of interest in a variety of fields in recent years, major deep learning models barely use complex numbers. However, speech, signal and audio data are naturally complex-valued after Fourier Transform, and studies have shown a potentially richer representation of complex nets. In this paper, we propose a Complex Transformer, which incorporates the transformer model as a backbone for sequence modeling; we also develop attention and encoder-decoder network operating for complex input. The model achieves state-of-the-art performance on the MusicNet dataset and an In-phase Quadrature (IQ) signal dataset.
In the past decade, the cyber-crime related to mobile devices has increased. Mobile devices, especially the ones running on Android operating system are particularly interesting to malware creators, as the users often keep the biggest amount of personal information on their mobile devices, such as their contacts, social media profiles, emails, and bank accounts. Both dynamic and static malware analysis is necessary to prevent and detect malware, as both techniques have their benefits and shortcomings. In this paper, we propose a deep learning technique that relies on LSTM and encoder-decoder neural network architectures for dynamic malware analysis based on CPU, memory and battery usage. The proposed system is able to detect and notify users about anomalies in system that is likely consequence of malware behaviour. The method was implemented as a part of OWASP Seraphimdroids anti-malware mechanism and notifies users about anomalies on their devices. The method proved to perform with an F1-score of 79.2%.
The amount of available Earth observation data has increased dramatically in the recent years. Efficiently making use of the entire body information is a current challenge in remote sensing and demands for light-weight problem-agnostic models that do not require region- or problem-specific expert knowledge. End-to-end trained deep learning models can make use of raw sensory data by learning feature extraction and classification in one step solely from data. Still, many methods proposed in remote sensing research require implicit feature extraction through data preprocessing or explicit design of features. In this work, we compare recent deep learning models on crop type classification on raw and preprocessed Sentinel 2 data. We concentrate on the common neural network architectures for time series, i.e., 1D-convolutions, recurrence, a shallow random forest baseline, and focus on the novel self-attention architecture. Our central findings are that data preprocessing still increased the overall classification performance for all models while the choice of model was less crucial. Self-attention and recurrent neural networks, by their architecture, outperformed convolutional neural networks on raw satellite time series. We explore this by a feature importance analysis based on gradient back-propagation that exploits the differentiable nature of deep learning models. Further, we qualitatively show how self-attention scores focus selectively on few classification-relevant observations.
Compressed sensing can increase resolution, and decrease electron dose and scan time of electron microscope point-scan systems with minimal information loss. Building on a history of successful deep learning applications in compressed sensing, we have developed a two-stage multiscale generative adversarial network to supersample scanning transmission electron micrographs with point-scan coverage reduced to 1/16, 1/25, ..., 1/100 px. We propose a novel non-adversarial learning policy to train a unified generator for multiple coverages and introduce an auxiliary network to homogenize prioritization of training data with varied signal-to-noise ratios. This achieves root mean square errors of 3.23% and 4.54% at 1/16 px and 1/100 px coverage, respectively; within 1% of errors for networks trained for each coverage individually. Detailed error distributions are presented for unified and individual coverage generators, including errors per output pixel. In addition, we present a baseline one-stage network for a single coverage and investigate numerical precision for web serving. Source code, training data, and pretrained models are publicly available at https://github.com/Jeffrey-Ede/DLSS-STEM
Bayesian neural networks (BNNs) have developed into useful tools for probabilistic modelling due to recent advances in variational inference enabling large scale BNNs. However, BNNs remain brittle and hard to train, especially: (1) when using deep architectures consisting of many hidden layers and (2) in situations with large weight variances. We use signal propagation theory to quantify these challenges and propose self-stabilising priors. This is achieved by a reformulation of the ELBO to allow the prior to influence network signal propagation. Then, we develop a stabilising prior, where the distributional parameters of the prior are adjusted before each forward pass to ensure stability of the propagating signal. This stabilised signal propagation leads to improved convergence and robustness making it possible to train deeper networks and in more noisy settings.
In computer chip manufacturing, the study of etch patterns on silicon wafers, or metrology, occurs on the nano-scale and is therefore subject to large variation from small, yet significant, perturbations in the manufacturing environment. An enormous amount of information can be gathered from a single etch process, a sequence of actions taken to produce an etched wafer from a blank piece of silicon. Each final wafer, however, is costly to take measurements from, which limits the number of examples available to train a predictive model. Part of the significance of this work is the success we saw from the models despite the limited number of examples. In order to accommodate the high dimensional process signatures, we isolated important sensor variables and applied domain-specific summarization on the data using multiple feature engineering techniques. We used a neural network architecture consisting of the summarized inputs, a single hidden layer of 4032 units, and an output layer of one unit. Two different models were learned, corresponding to the metrology measurements in the dataset, Recess and Remaining Mask. The outputs are related abstractly and do not form a two dimensional space, thus two separate models were learned. Our results approach the error tolerance of the microscopic imaging system. The model can make predictions for a class of etch recipes that include the correct number of etch steps and plasma reactors with the appropriate sensors, which are chambers containing an ionized gas that determine the manufacture environment. Notably, this method is not restricted to some maximum process length due to the summarization techniques used. This allows the method to be adapted to new processes that satisfy the aforementioned requirements. In order to automate semiconductor manufacturing, models like these will be needed throughout the process to evaluate production quality.
In recent years, the interest in Big Data sources has been steadily growing within the Official Statistic community. The Italian National Institute of Statistics (Istat) is currently carrying out several Big Data pilot studies. One of these studies, the ICT Big Data pilot, aims at exploiting massive amounts of textual data automatically scraped from the websites of Italian enterprises in order to predict a set of target variables (e.g. e-commerce) that are routinely observed by the traditional ICT Survey. In this paper, we show that Deep Learning techniques can successfully address this problem. Essentially, we tackle a text classification task: an algorithm must learn to infer whether an Italian enterprise performs e-commerce from the textual content of its website. To reach this goal, we developed a sophisticated processing pipeline and evaluated its performance through extensive experiments. Our pipeline uses Convolutional Neural Networks and relies on Word Embeddings to encode raw texts into grayscale images (i.e. normalized numeric matrices). Web-scraped texts are huge and have very low signal to noise ratio: to overcome these issues, we adopted a framework known as False Positive Reduction, which has seldom (if ever) been applied before to text classification tasks. Several original contributions enable our processing pipeline to reach good classification results. Empirical evidence shows that our proposal outperforms all the alternative Machine Learning solutions already tested in Istat for the same task.
The lack of a large amount of training data has always been the constraining factor in solving a lot of problems in machine learning, making One Shot Learning one of the most intriguing ideas in machine learning. It aims to learn information about object categories from one, or only a few training examples. This process of learning in deep learning is usually accomplished by proper objective function, i.e; loss function and embeddings extraction i.e; architecture. In this paper, we discussed about metrics based deep learning architectures for one shot learning such as Siamese neural networks and present a method to improve on their accuracy using Kafnets (kernel-based non-parametric activation functions for neural networks) by learning proper embeddings with relatively less number of epochs. Using kernel activation functions, we are able to achieve strong results which exceed those of ReLU based deep learning models in terms of embeddings structure, loss convergence, and accuracy.
Generalization is essential for deep learning. In contrast to previous works claiming that Deep Neural Networks (DNNs) have an implicit regularization implemented by the stochastic gradient descent, we demonstrate explicitly Bayesian regularizations in a specific category of DNNs, i.e., Convolutional Neural Networks (CNNs). First, we introduce a novel probabilistic representation for the hidden layers of CNNs and demonstrate that CNNs correspond to Bayesian networks with the serial connection. Furthermore, we show that the hidden layers close to the input formulate prior distributions, thus CNNs have explicitly Bayesian regularizations based on the Bayesian regularization theory. In addition, we clarify two recently observed empirical phenomena that are inconsistent with traditional theories of generalization. Finally, we validate the proposed theory on a synthetic dataset
Credit investigation is critical for financial services. Whereas, traditional methods are often restricted as the employed data hardly provide sufficient, timely and reliable information. With the prevalence of smart mobile devices, peoples' geographic footprints can be automatically and constantly collected nowadays, which provides an unprecedented opportunity for credit investigations. Inspired by the observation that locations are somehow related to peoples' credit level, this research aims to enhance credit investigation with users' geographic footprints. To this end, a two-stage credit investigation framework is designed, namely CreditPrint. In the first stage, CreditPrint explores regions' credit characteristics and learns a credit-aware embedding for each region by considering both each region's individual characteristics and cross-region relationships with graph convolutional networks. In the second stage, a hierarchical attention-based credit assessment network is proposed to aggregate the credit indications from a user's multiple trajectories covering diverse regions. The results on real-life user mobility datasets show that CreditPrint can increase the credit investigation accuracy by up to 10% compared to baseline methods.
Real artificial intelligence always has been focused on by many machine learning researchers, especially in the area of deep learning. However deep neural network is hard to be understood and explained, and sometimes, even metaphysics. The reason is, we believe that: the network is essentially a perceptual model. Therefore, we believe that in order to complete complex intelligent activities from simple perception, it is necessary to con-struct another interpretable logical network to form accurate and reasonable responses and explanations to external things. Researchers like Bolei Zhou and Quanshi Zhang have found many explanatory rules for deep feature extraction aimed at the feature extraction stage of convolution neural network. However, although researchers like Marco Gori have also made great efforts to improve the interpretability of the fully connected layers of the network, the problem is also very difficult. This paper firstly analyzes its reason. Then a method of constructing logical network based on the fully connected layers and extracting logical relation between input and output of the layers is proposed. The game process between perceptual learning and logical abstract cognitive learning is implemented to improve the interpretable performance of deep learning process and deep learning model. The benefits of our approach are illustrated on benchmark data sets and in real-world experiments.
The field of high-energy physics (HEP), along with many scientific disciplines, is currently experiencing a dramatic influx of new methodologies powered by modern machine learning techniques. Over the last few years, a growing body of HEP literature has focused on identifying promising applications of deep learning in particular, and more recently these techniques are starting to be realized in an increasing number of experimental measurements. The overall conclusion from this impressive and extensive set of studies is that rarer and more complex physics signatures can be identified with the new set of powerful tools from deep learning. However, there is an unstudied systematic risk associated with combining the traditional HEP workflow and deep learning with high-dimensional data. In particular, calibrating and validating the response of deep neural networks is in general not experimentally feasible, and therefore current methods may be biased in ways that are not covered by current uncertainty estimates. By borrowing ideas from AI safety, we illustrate these potential issues and propose a method to bound the size of unaccounted for uncertainty. In addition to providing a pragmatic diagnostic, this work will hopefully begin a dialogue within the community about the robust application of deep learning to experimental analyses.
Kuzushiji, a cursive writing style, had been used in Japan for over a thousand years starting from the 8th century. Over 3 millions books on a diverse array of topics, such as literature, science, mathematics and even cooking are preserved. However, following a change to the Japanese writing system in 1900, Kuzushiji has not been included in regular school curricula. Therefore, most Japanese natives nowadays cannot read books written or printed just 150 years ago. Museums and libraries have invested a great deal of effort into creating digital copies of these historical documents as a safeguard against fires, earthquakes and tsunamis. The result has been datasets with hundreds of millions of photographs of historical documents which can only be read by a small number of specially trained experts. Thus there has been a great deal of interest in using Machine Learning to automatically recognize these historical texts and transcribe them into modern Japanese characters. Nevertheless, several challenges in Kuzushiji recognition have made the performance of existing systems extremely poor. To tackle these challenges, we propose KuroNet, a new end-to-end model which jointly recognizes an entire page of text by using a residual U-Net architecture which predicts the location and identity of all characters given a page of text (without any pre-processing). This allows the model to handle long range context, large vocabularies, and non-standardized character layouts. We demonstrate that our system is able to successfully recognize a large fraction of pre-modern Japanese documents, but also explore areas where our system is limited and suggest directions for future work.
We generalize a standard benchmark of reinforcement learning, the classical cartpole balancing problem, to the quantum regime by stabilizing a particle in an unstable potential through measurement and feedback. We use state-of-the-art deep reinforcement learning to stabilize a quantum cartpole and find that our deep learning approach performs comparably to or better than other strategies in standard control theory. Our approach also applies to measurement-feedback cooling of quantum oscillators, showing the applicability of deep learning to general continuous-space quantum control.
Understanding dynamic 3D environment is crucial for robotic agents and many other applications. We propose a novel neural network architecture called $MeteorNet$ for learning representations for dynamic 3D point cloud sequences. Different from previous work that adopts a grid-based representation and applies 3D or 4D convolutions, our network directly processes point clouds. We propose two ways to construct spatiotemporal neighborhoods for each point in the point cloud sequence. Information from these neighborhoods is aggregated to learn features per point. We benchmark our network on a variety of 3D recognition tasks including action recognition, semantic segmentation and scene flow estimation. MeteorNet shows stronger performance than previous grid-based methods while achieving state-of-the-art performance on Synthia. MeteorNet also outperforms previous baseline methods that are able to process at most two consecutive point clouds. To the best of our knowledge, this is the first work on deep learning for dynamic raw point cloud sequences.