295 research outputs found
25GMAC/sec/axon photonic neural networks with 7GHz bandwidth optics through channel response-aware training
No full textWe present a channel response-aware Photonic Neural Network (PNN) and demonstrate experimentally its resilience in Inter-Symbol Interference (ISI) when implemented in an integrated neuron. The trained PNN model performs at 25GMAC/sec/axon using only 7GHz-bandwidth photonic axons with 97.37% accuracy in the MNIST dataset
Biosignal time-series analysis
No full textIn this chapter, recent state-of-the-art techniques in biosignal time-series analysis will be presented. We shall start with the problem of patient-specific ECG beat classification where the objective is to discriminate the arrhythmic beats from the normal (healthy) beats of an individual patient. So, we will answer the ultimate question of how to design person-specific, real-time, and accurate monitoring of ECG signals. We shall then move on to the recent solution of a related problem, an early warning system that can alert an individual the instant his/her heart deviates from its normal rhythm. This is a far challenging problem since the detection of the arrhythmia beats should be performed without knowing them.</p
Medical image analysis
No full textThis chapter presents deep learning methodologies for medical imaging tasks. The chapter starts with echocardiography for early detection of myocardial infarction (MI) or commonly known as heart attack. Early and fundamental signs of MI can be visible as the abnormality in one or several segments of the left ventricle (LV) wall, where a segment may move “abnormally” or “nonuniformly.” The primary tool to detect and identify such regional wall motion abnormalities is echocardiography, which is a fast, cost-effective, and lowest risk imaging option. A three-phase approach is introduced, where the entire LV wall is segmented by a deep learning model, and then characteristics of the segmented wall are used to perform early detection of MI robustly and accurately. The second medical imaging task discussed in the chapter is the recognition of coronavirus disease 2019 (COVID-19), which has become a global health concern after it is declared as a pandemic in March 2020. Developing automatic, accurate, and fast algorithms for COVID-19 detection plays a vital role in the prevention of spreading the virus. Deep learning models can provide state-of-the-art performance in many imaging tasks. However, due to data scarcity, these models cannot produce satisfactory results when trained for COVID-19 recognition. To tackle this issue, Convolutional Support Estimator Network (CSEN) is introduced due to its advantage over a scarce-data classification task for robust COVID-19 recognition using chest X-ray images. In order to utilize the CSEN classification scheme, features are extracted from a state-of-the-art deep neural network. Consequently, the introduced network can achieve an elegant performance for COVID-19 recognition
Neuromorphic silicon photonics and hardware-aware deep learning for high-speed inference
No full textThe relentless growth of Artificial Intelligence (AI) workloads has fueled the drive towards non-Von Neuman architectures and custom computing hardware. Neuromorphic photonic engines aspire to synergize the low-power and high-bandwidth credentials of light-based deployments with novel architectures, towards surpassing the computing performance of their electronic counterparts. In this paper, we review recent progress in integrated photonic neuromorphic architectures and analyze the architectural and photonic hardware-based factors that limit their performance. Subsequently, we present our approach towards transforming silicon coherent neuromorphic layouts into high-speed and high-accuracy Deep Learning (DL) engines by combining robust architectures with hardware-aware DL training. Circuit robustness is ensured through a crossbar layout that circumvents insertion loss and fidelity constraints of state-of-the-art linear optical designs. Concurrently, we employ DL training models adapted to the underlying photonic hardware, incorporating noise- and bandwidth-limitations together with the supported activation function directly into Neural Network (NN) training. We validate experimentally the high-speed and high-accuracy advantages of hardware-aware DL models when combined with robust architectures through a SiPho prototype implementing a single column of a 4:4 photonic crossbar. This was utilized as the pen-ultimate hidden layer of a NN, revealing up to 5.93% accuracy improvement at 5GMAC/sec/axon when noise-aware training is enforced and allowing accuracies of 99.15% and 79.8% for the MNIST and CIFAR-10 classification tasks. Channel-aware training was then demonstrated by integrating the frequency response of the photonic hardware in NN training, with its experimental validation with the MNIST dataset revealing an accuracy increase of 12.93% at a record-high rate of 25GMAC/sec/axon
A silicon photonic coherent neuron with 10GMAC/sec processing line-rate
No full textWe demonstrate a novel coherent Si-Pho neuron with 10Gbaud on-chip input-data vector generation capabilities. Its performance as a hidden layer within a neural network has been experimentally validated for the MNIST data-set, yielding 96.19% accuracy
Novel Decision Forest Building Techniques by Utilising Correlation Coefficient Methods
No full textDecision Forests have attracted the academic community’s interest mainly due to their simplicity and transparency. This paper proposes two novel decision forest building techniques, called Maximal Information Coefficient Forest (MICF) and Pearson’s Correlation Coefficient Forest (PCCF). The proposed new algorithms use Pearson’s Correlation Coefficient (PCC) and Maximal Information Coefficient (MIC) as extra measures of the classification capacity score of each feature. Using those approaches, we improve the picking of the most convenient feature at each splitting node, the feature with the greatest Gain Ratio. We conduct experiments on 12 datasets that are available in the publicly accessible UCI machine learning repository. Our experimental results indicate that the proposed methods have the best average ensemble accuracy rank of 1.3 (for MICF) and 3.0 (for PCCF), compared to their closest competitor, Random Forest (RF), which has an average rank of 4.3. Additionally, the results from Friedman and Bonferroni-Dunn tests indicate statistically significant improvement
TrojanDroid:Android Malware Detection for Trojan Discovery Using Convolutional Neural Networks
No full textAndroid platforms are widely used nowadays in different forms such as mobile phones and tablets, and this has made the Android platform an attractive target for hackers. While there are many solutions available for detecting malware on Android devices there aren’t that many that are concentrated on specific malware types. To this extent, this paper delivers a new dataset for Trojan detection for Android apps based on the permissions of the apps, while the second contribution is a neural network architecture that can classify with very high accuracy if an Android app is a genuine app or a Trojan pretending to be a normal app. We have run extensive evaluation tests to validate the performance of the proposed method and we have compared it to other well-known classifiers using well-known evaluation metrics to show its effectiveness
Automatic Accent and Gender Recognition of Regional UK Speakers
No full textWith the ubiquity of voice assistants across the UK and the world, speech recognition of the regional accents across the British Isles has proven challenging due to varying pronunciations. This paper proposes an automated recognition of the geographical origin and gender of a voice sample based on the six regional dialects of the United Kingdom. Twenty six features are extracted from 17,877 voice samples and then used to design, implement and evaluate machine learning classifiers based on Artificial Neural Networks (ANNs), Support Vector Machine (SVM), Random Forest (RF) and k-nearest neighbors (k-NN) algorithms. The results suggest that the proposed approach could be applicable for areas such as e-commerce and the service industry, and it provides a contribution to NLP audio research.</p
Leveraging Quadratic Spherical Mutual Information Hashing for Fast Image Retrieval
No full textInternational audienceSeveral deep supervised hashing techniques have been proposed to allow for querying large image databases. However, it is often overlooked that the process of information retrieval can be modeled using information-theoretic metrics, leading to optimizing various proxies for the problem at hand instead. Contrary to this, we propose a deep supervised hashing algorithm that optimizes the learned codes using an informationtheoretic measure, the Quadratic Mutual Information (QMI). The proposed method is adapted to the needs of large-scale hashing and information retrieval leading to a novel information-theoretic measure, the Quadratic Spherical Mutual Information (QSMI), that is inspired by QMI, but leads to significant better retrieval precision. Indeed, the effectiveness of the proposed method is demonstrated under several different scenarios, using different datasets and network architectures, outperforming existing deep supervised image hashing techniques
- …
