Estimating the entropy of a discrete random variable is a fundamental problem in information theory and related fields. This problem has many applications in various domains, including machine learning, statistics and data compression. Over the years, a variety of estimation schemes have been suggested. However, despite significant progress, most methods still struggle when the sample is small, compared to the variable's alphabet size. In this work, we introduce a practical solution to this problem, which extends the work of McAllester and Statos (2020). The proposed scheme uses the generalization abilities of cross-entropy estimation in deep neural networks (DNNs) to introduce improved entropy estimation accuracy. Furthermore, we introduce a family of estimators for related information-theoretic measures, such as conditional entropy and mutual information. We show that these estimators are strongly consistent and demonstrate their performance in a variety of use-cases. First, we consider large alphabet entropy estimation. Then, we extend the scope to mutual information estimation. Next, we apply the proposed scheme to conditional mutual information estimation, as we focus on independence testing tasks. Finally, we study a transfer entropy estimation problem. The proposed estimators demonstrate improved performance compared to existing methods in all tested setups.
Petabytes of data are generated each day by emerging Internet of Things (IoT), but only few of them can be finally collected and used for Machine Learning (ML) purposes due to the apprehension of data & privacy leakage, which seriously retarding ML's growth. To alleviate this problem, Federated learning is proposed to perform model training by multiple clients' combined data without the dataset sharing within the cluster. Nevertheless, federated learning introduces massive communication overhead as the synchronized data in each epoch is of the same size as the model, and thereby leading to a low communication efficiency. Consequently, variant methods mainly focusing on the communication rounds reduction and data compression are proposed to reduce the communication overhead of federated learning. In this paper, we propose Overlap-FedAvg, a framework that parallels the model training phase with model uploading & downloading phase, so that the latter phase can be totally covered by the former phase. Compared to vanilla FedAvg, Overlap-FedAvg is further developed with a hierarchical computing strategy, a data compensation mechanism and a nesterov accelerated gradients~(NAG) algorithm. Besides, Overlap-FedAvg is orthogonal to many other compression methods so that they can be applied together to maximize the utilization of the cluster. Furthermore, the theoretical analysis is provided to prove the convergence of the proposed Overlap-FedAvg framework. Extensive experiments on both conventional and recurrent tasks with multiple models and datasets also demonstrate that the proposed Overlap-FedAvg framework substantially boosts the federated learning process.
Autoencoders are a class of artificial neural networks which have gained a lot of attention in the recent past. Using the encoder block of an autoencoder the input image can be compressed into a meaningful representation. Then a decoder is employed to reconstruct the compressed representation back to a version which looks like the input image. It has plenty of applications in the field of data compression and denoising. Another version of Autoencoders (AE) exist, called Variational AE (VAE) which acts as a generative model like GAN. Recently, an optimizer was introduced which is known as lookahead optimizer which significantly enhances the performances of Adam as well as SGD. In this paper, we implement Convolutional Autoencoders (CAE) and Convolutional Variational Autoencoders (CVAE) with lookahead optimizer (with Adam) and compare them with the Adam (only) optimizer counterparts. For this purpose, we have used a movie dataset comprising of natural images for the former case and CIFAR100 for the latter case. We show that lookahead optimizer (with Adam) improves the performance of CAEs for reconstruction of natural images.
Clustering is a ubiquitous problem in data science and signal processing. In many applications where we observe noisy signals, it is common practice to first denoise the data, perhaps using wavelet denoising, and then to apply a clustering algorithm. In this paper, we develop a sparse convex wavelet clustering approach that simultaneously denoises and discovers groups. Our approach utilizes convex fusion penalties to achieve agglomeration and group-sparse penalties to denoise through sparsity in the wavelet domain. In contrast to common practice which denoises then clusters, our method is a unified, convex approach that performs both simultaneously. Our method yields denoised (wavelet-sparse) cluster centroids that both improve interpretability and data compression. We demonstrate our method on synthetic examples and in an application to NMR spectroscopy.
Wireless signal strength based localization can enable robust localization for robots using inexpensive sensors. For this, a location-to-signal-strength map has to be learned for each access point in the environment. Due to the ubiquity of Wireless networks in most environments, this can result in tens or hundreds of maps. To reduce the dimensionality of this problem, we employ autoencoders, which are a popular unsupervised approach for feature extraction and data compression. In particular, we propose the use of sparse autoencoders that learn latent spaces that preserve the relative distance between inputs. Distance invariance between input and latent spaces allows our system to successfully learn compact representations that allow precise data reconstruction but also have a low impact on localization performance when using maps from the latent space rather than the input space. We demonstrate the feasibility of our approach by performing experiments in outdoor environments.
Spatial linear transforms that process multiple parallel analog signals to simplify downstream signal processing find widespread use in multi-antenna communication systems, machine learning inference, data compression, audio and ultrasound applications, among many others. In the past, a wide range of mixed-signal as well as digital spatial transform circuits have been proposed---it is, however, a longstanding question whether analog or digital transforms are superior in terms of throughput, power, and area. In this paper, we focus on Hadamard transforms and perform a systematic comparison of state-of-the-art analog and digital circuits implementing spatial transforms in the same 65\,nm CMOS technology. We analyze the trade-offs between throughput, power, and area, and we identify regimes in which mixed-signal or digital Hadamard transforms are preferable. Our comparison reveals that (i) there is no clear winner and (ii) analog-to-digital conversion is often dominating area and energy efficiency---and not the spatial transform.
This paper presents a wireless neural recording system featuring energy-efficient data compression and encryption. An ultra-high efficiency is achieved by leveraging compressed sensing (CS) for simultaneous data compression and encryption. CS enables sub-Nyquist sampling of neural signals by taking advantage of its intrinsic sparsity. It simultaneously encrypts the data with the sampling matrix being the cryptographic key. To share the key over an insecure wireless channel, we implement an elliptic-curve cryptography (ECC) based key exchanging protocol. The CS operation is executed in a custom-designed IC fabricated in 180nm CMOS technology. Mixed-signal circuits are designed to optimize the power efficiency of the matrix-vector multiplication (MVM) of the CS operation. The ECC algorithm is implemented in a low-power Cortex-M0 microcontroller (MCU). To be protected from timing and power analysis attacks, the implementation avoids possible data-dependent branches and also employs a randomized ECC initialization. At a compression ratio of 8x, the average correlated coefficient between the reconstructed signals and the uncompressed signals is 0.973, while the ciphertext-only attacks (CoA) achieve no better than 0.054 over 200,000 attacks. The prototype achieves a 35x power saving compared with conventional implementation in low-power MCUs. This work demonstrates a promising solution for future chronic neural recording systems with requirements in high energy efficiency and security.
This paper proposes compressed domain signal processing (CSP) multiple input multiple output (MIMO) radar, a MIMO radar approach that achieves substantial sample complexity reduction by exploiting the idea of CSP. CSP MIMO radar involves two levels of data compression followed by target detection at the compressed domain. First, compressive sensing is applied at the receive antennas, followed by a Capon beamformer which is designed to suppress clutter. Exploiting the sparse nature of the beamformer output, a second compression is applied to the filtered data. Target detection is subsequently conducted by formulating and solving a hypothesis testing problem at each grid point of the discretized angle space. The proposed approach enables an 8-fold reduction of the sample complexity in some settings as compared to a conventional compressed sensing (CS) MIMO radar thus enabling faster target detection. Receiver operating characteristic (ROC) curves of the proposed detector are provided. Simulation results show that the proposed approach outperforms recovery-based compressed sensing algorithms.
Improving efficiency of healthcare systems is a top national interest worldwide. However, the need of delivering scalable healthcare services to the patients while reducing costs is a challenging issue. Among the most promising approaches for enabling smart healthcare (s-health) are edge-computing capabilities and next-generation wireless networking technologies that can provide real-time and cost-effective patient remote monitoring. In this paper, we present our vision of exploiting multi-access edge computing (MEC) for s-health applications. We envision a MEC-based architecture and discuss the benefits that it can bring to realize in-network and context-aware processing so that the s-health requirements are met. We then present two main functionalities that can be implemented leveraging such an architecture to provide efficient data delivery, namely, multimodal data compression and edge-based feature extraction for event detection. The former allows efficient and low distortion compression, while the latter ensures high-reliability and fast response in case of emergency applications. Finally, we discuss the main challenges and opportunities that edge computing could provide and possible directions for future research.
Rate-distortion (RD) theory is at the heart of lossy data compression. Here we aim to model the generalized RD (GRD) trade-off between the visual quality of a compressed video and its encoding profiles (e.g., bitrate and spatial resolution). We first define the theoretical functional space $\mathcal{W}$ of the GRD function by analyzing its mathematical properties.We show that $\mathcal{W}$ is a convex set in a Hilbert space, inspiring a computational model of the GRD function, and a method of estimating model parameters from sparse measurements. To demonstrate the feasibility of our idea, we collect a large-scale database of real-world GRD functions, which turn out to live in a low-dimensional subspace of $\mathcal{W}$. Combining the GRD reconstruction framework and the learned low-dimensional space, we create a low-parameter eigen GRD method to accurately estimate the GRD function of a source video content from only a few queries. Experimental results on the database show that the learned GRD method significantly outperforms state-of-the-art empirical RD estimation methods both in accuracy and efficiency. Last, we demonstrate the promise of the proposed model in video codec comparison.
Underwater vehicles are utilized in various applications including underwater data-collection missions. The tethered connection constrains the mission both in distance traveled and number of vehicles that can run in the same area, while the addition of acoustic communications onto the vehicles grants them several functionalities. However, due to the low bandwidth of the underwater acoustic channel-which leads to low data rates-and the time overhead imposed by both the channel propagation delay and the processing delay by the acoustic modems, efficient protocols are required. In this paper, an implicit data-compression and transmission protocol is proposed to carry out environmental monitoring missions such as adaptive sampling of physical and chemical parameters in the water. In a semi-autonomous manner between the vehicle and the control center, both sides keep silent in data transmission as long as they can estimate and predict the actions of the other side, unless environmental data and/or kinematic data are found to be unpredictable. Our design puts the human in the loop to send high-level control commands. Experiments were conducted using an autonomous vehicle with WHOI micro-modems in the Raritan River, Somerset, Carnegie Lake in Princeton, and in the Marine Park in Red Bank, all in New Jersey.
Data compression often subtracts prediction and encodes the difference (residue) e.g. assuming Laplace distribution, for example for images, videos, audio, or numerical data. Its performance is strongly dependent on the proper choice of width (scale parameter) of this parametric distribution, can be improved if optimizing it based on local situation like context. For example in popular LOCO-I \cite{loco} (JPEG-LS) lossless image compressor there is used 3 dimensional context quantized into 365 discrete possibilities treated independently. This article discusses inexpensive approaches for exploiting their dependencies with autoregressive ARCH-like context dependent models for parameters of parametric distribution for residue, also evolving in time for adaptive case. For example tested such 4 or 11 parameter models turned out to provide similar performance as 365 parameter LOCO-I model for 48 tested images. Beside smaller headers, such reduction of number of parameters can lead to better generalization. In contrast to context quantization approaches, parameterized models also allow to directly use higher dimensional contexts, for example using information from all 3 color channels, further pixels, some additional region classifiers, or from interleaving multi-scale scanning - for which there is proposed Haar upscale scan combining advantages of Haar wavelets with possibility of scanning exploiting local contexts.
Quantization is an essential step in the analog-to-digital conversion process and it is very important in all modern telecommunication systems. In this paper, a novel chaotic uniform quantizer is proposed and its application for speech coding is presented. The proposed system consists of three stages: two PCM coders separated by an XOR operation with a chaotic sequence, where the first step is used for continuous signal sampling and second stage performs data encryption, while the third stage provides additional data compression. The performance of the presented quantizer for Laplacian distributed signals and real speech signals is investigated and compared with that of the well-known uniform and non-uniform quantizers. Simulation results show that the proposed quantizer provides secured data with higher levels of SQNR compared to others.
This paper deals with parameter estimation from extreme measurements. While being a special case of parameter estimation from partial data, in scenarios where only one sample from a given set of K measurements can be extracted, choosing only the minimum or the maximum (i.e., extreme) value from that set is of special interest because of the ultra-low energy, storage, and processing power required to extract extreme values from a given data set. We present a new methodology to analyze the performance of parameter estimation from extreme measurements. In particular, we present a general close-form approximation for the Cramer-Rao Lower Bound on the parameter estimation error, based on extreme values. We demonstrate our methodology on the case where the original measurements are exponential distributed, which is related to many practical applications. The analysis shows that the maximum values carry most of the information about the parameter of interest and that the additional information in the minimum is negligible. Moreover, it shows that for small sets of iid measurements (e.g. K=15) the use of the maximum can provide data compression with factor 15 while keeping about 50% of the information stored in the complete set.
The Magnetic Resonance Imaging (MRI) processing chain starts with a critical acquisition stage that provides raw data for reconstruction of images for medical diagnosis. This flow usually includes a near-lossless data compression stage that enables digital storage and/or transmission in binary formats. In this work we propose a framework for joint optimization of the MRI reconstruction and lossy compression, producing compressed representations of medical images that achieve improved trade-offs between quality and bit-rate. Moreover, we demonstrate that lossy compression can even improve the reconstruction quality compared to settings based on lossless compression. Our method has a modular optimization structure, implemented using the alternating direction method of multipliers (ADMM) technique and the state-of-the-art image compression technique (BPG) as a black-box module iteratively applied. This establishes a medical data compression approach compatible with a lossy compression standard of choice. A main novelty of the proposed algorithm is in the total-variation regularization added to the modular compression process, leading to decompressed images of higher quality without any additional processing at/after the decompression stage. Our experiments show that our regularization-based approach for joint MRI reconstruction and compression often achieves significant PSNR gains between 4 to 9 dB at high bit-rates compared to non-regularized solutions of the joint task. Compared to regularization-based solutions, our optimization method provides PSNR gains between 0.5 to 1 dB at high bit-rates, which is the range of interest for medical image compression.
Density estimation, compression and data generation are crucial tasks in artificial intelligence. Variational Auto-Encoders (VAEs) constitute a single framework to achieve these goals. Here, we present a novel class of generative models, called self-supervised Variational Auto-Encoder (selfVAE), that utilizes deterministic and discrete variational posteriors. This class of models allows to perform both conditional and unconditional sampling, while simplifying the objective function. First, we use a single self-supervised transformation as a latent variable, where a transformation is either downscaling or edge detection. Next, we consider a hierarchical architecture, i.e., multiple transformations, and we show its benefits compared to the VAE. The flexibility of selfVAE in data reconstruction finds a particularly interesting use case in data compression tasks, where we can trade-off memory for better data quality, and vice-versa. We present performance of our approach on three benchmark image data (Cifar10, Imagenette64, and CelebA).
Tensor completion refers to the task of estimating the missing data from an incomplete measurement or observation, which is a core problem frequently arising from the areas of big data analysis, computer vision, and network engineering. Due to the multidimensional nature of high-order tensors, the matrix approaches, e.g., matrix factorization and direct matricization of tensors, are often not ideal for tensor completion and recovery. In this paper, we introduce a unified low-rank and sparse enhanced Tucker decomposition model for tensor completion. Our model possesses a sparse regularization term to promote a sparse core tensor of the Tucker decomposition, which is beneficial for tensor data compression. Moreover, we enforce low-rank regularization terms on factor matrices of the Tucker decomposition for inducing the low-rankness of the tensor with a cheap computational cost. Numerically, we propose a customized ADMM with enough easy subproblems to solve the underlying model. It is remarkable that our model is able to deal with different types of real-world data sets, since it exploits the potential periodicity and inherent correlation properties appeared in tensors. A series of computational experiments on real-world data sets, including internet traffic data sets, color images, and face recognition, demonstrate that our model performs better than many existing state-of-the-art matricization and tensorization approaches in terms of achieving higher recovery accuracy.
The impressive growth of data throughput in optical microscopy has triggered a widespread use of supervised learning (SL) models running on compressed image datasets for efficient automated analysis. However, since lossy image compression risks to produce unpredictable artifacts, quantifying the effect of data compression on SL applications is of pivotal importance to assess their reliability, especially for clinical use. We propose an experimental method to evaluate the tolerability of image compression distortions in 2D and 3D cell segmentation SL tasks: predictions on compressed data are compared to the raw predictive uncertainty, which is numerically estimated from the raw noise statistics measured through sensor calibration. We show that predictions on object- and image-specific segmentation parameters can be altered by up to 15% and more than 10 standard deviations after 16-to-8 bits downsampling or JPEG compression. In contrast, a recently developed lossless compression algorithm provides a prediction spread which is statistically equivalent to that stemming from raw noise, while providing a compression ratio of up to 10:1. By setting a lower bound to the SL predictive uncertainty, our technique can be generalized to validate a variety of data analysis pipelines in SL-assisted fields.
In many ultrasonic imaging systems, data acquisition and image formation are performed on separate computing devices. Data transmission is becoming a bottleneck, thus, efficient data compression is essential. Compression rates can be improved by considering the fact that many image formation methods rely on approximations of wave-matter interactions, and only use the corresponding part of the data. Tailored data compression could exploit this, but extracting the useful part of the data efficiently is not always trivial. In this work, we tackle this problem using deep neural networks, optimized to preserve the image quality of a particular image formation method. The Delay-And-Sum (DAS) algorithm is examined which is used in reflectivity-based ultrasonic imaging. We propose a novel encoder-decoder architecture with vector quantization and formulate image formation as a network layer for end-to-end training. Experiments demonstrate that our proposed data compression tailored for a specific image formation method obtains significantly better results as opposed to compression agnostic to subsequent imaging. We maintain high image quality at much higher compression rates than the theoretical lossless compression rate derived from the rank of the linear imaging operator. This demonstrates the great potential of deep ultrasonic data compression tailored for a specific image formation method.
Recent deep learning based video synthesis approaches, in particular with applications that can forge identities such as "DeepFake", have raised great security concerns. Therefore, corresponding deep forensic methods are proposed to tackle this problem. However, existing methods are either based on unexplainable deep networks which greatly degrades the principal interpretability factor to media forensic, or rely on fragile image statistics such as noise pattern, which in real-world scenarios can be easily deteriorated by data compression. In this paper, we propose an fully-interpretable video forensic method that is designed specifically to expose deep-faked videos. To enhance generalizability on videos with various content, we model the temporal motion of multiple specific spatial locations in the videos to extract a robust and reliable representation, called Co-Motion Pattern. Such kind of conjoint pattern is mined across local motion features which is independent of the video contents so that the instance-wise variation can also be largely alleviated. More importantly, our proposed co-motion pattern possesses both superior interpretability and sufficient robustness against data compression for deep-faked videos. We conduct extensive experiments to empirically demonstrate the superiority and effectiveness of our approach under both classification and anomaly detection evaluation settings against the state-of-the-art deep forensic methods.