In this paper, we propose an image compression algorithm called Microshift. We employ an algorithm hardware co-design methodology, yielding a hardware-friendly compression approach with low power consumption. In our method, the image is first micro-shifted, then the sub-quantized values are further compressed. Two methods, the FAST and MRF model, are proposed to recover the bit-depth by exploiting the spatial correlation of natural images. Both methods can decompress images progressively. Our compression algorithm compresses images to 1.25 bits per pixel on average with PSNR of 33.16 dB, outperforming other on-chip compression algorithms. Then, we propose a hardware architecture and implement the algorithm on an FPGA and ASIC. The results on the VLSI design further validate the low hardware complexity and high power efficiency, showing our method is promising, particularly for low-power wireless vision sensor networks.
The conventional approach to pre-process data for compression is to apply transforms such as the Fourier, the Karhunen-Lo\`{e}ve, or wavelet transforms. One drawback from adopting such an approach is that it is independent of the use of the compressed data, which may induce significant optimality losses when measured in terms of final utility (instead of being measured in terms of distortion). We therefore revisit this paradigm by tayloring the data pre-processing operation to the utility function of the decision-making entity using the compressed (and therefore noisy) data. More specifically, the utility function consists of an Lp-norm, which is very relevant in the area of smart grids. Both a linear and a non-linear use-oriented transforms are designed and compared with conventional data pre-processing techniques, showing that the impact of compression noise can be significantly reduced.
Autonomous vehicles require fleet-wide data collection for continuous algorithm development and validation. The Smart Black Box (SBB) intelligent event data recorder has been proposed as a system for prioritized high-bandwidth data capture. This paper extends the SBB by applying anomaly detection and action detection methods for generalized event-of-interest (EOI) detection. An updated SBB pipeline is proposed for the real-time capture of driving video data. A video dataset is constructed to evaluate the SBB on real-world data for the first time. SBB performance is assessed by comparing the compression of normal and anomalous data and by comparing our prioritized data recording with a FIFO strategy. Results show that SBB data compression can increase the anomalous-to-normal memory ratio by ~25%, while the prioritized recording strategy increases the anomalous-to-normal count ratio when compared to a FIFO strategy. We compare the real-world dataset SBB results to a baseline SBB given ground-truth anomaly labels and conclude that improved general EOI detection methods will greatly improve SBB performance.
Broadband, high resolution and rapid measurement of dual-comb spectroscopy (DCS) generates a large amount of data stream. We numerically demonstrate significant data compression of DCS spectra by using a compressive sensing technique. Our numerical simulation shows a compression rate of more than 100 with 3% error in mole fraction estimation of mid-infrared (MIR) DCS of two molecular species in a broadband (~30 THz) and high resolution (~115 MHz) condition. We also numerically demonstrate a massively parallel MIR DCS spectrum of 10 different molecular species can be reconstructed with a compression rate of 10.5 with a transmittance error of 0.003 from the original spectrum.
A data compression system capable of providing real-time streaming of high-resolution continuous point-on-wave (CPOW) and phasor measurement unit (PMU) measurements is proposed. Referred to as adaptive subband compression (ASBC), the proposed technique partitions the signal space into subbands and adaptively compresses subband signals based on each subband's active bandwidth. The proposed technique conforms to existing industry phasor measurement standards, making it suitable for streaming high-resolution CPOW and PMU data either in continuous or burst on-demand/event-triggered modes. Experiments on synthetic and real data show that ASBC reduces the CPOW sampling rates by several orders of magnitude for real-time streaming while maintaining the precision required by industry standards.
Sparse representation has been widely used in data compression, signal and image denoising, dimensionality reduction and computer vision. While overcomplete dictionaries are required for sparse representation of multidimensional data, orthogonal bases represent one-dimensional data well. In this paper, we propose a data-driven sparse representation using orthonormal bases under the lossless compression constraint. We show that imposing such constraint under the Minimum Description Length (MDL) principle leads to a unique and optimal sparse representation for one-dimensional data, which results in discriminative features useful for data discovery.
The lossless data compression algorithm based on Bayesian Attention Networks is derived from first principles. Bayesian Attention Networks are defined by introducing an attention factor per a training sample loss as a function of two sample inputs, from training sample and prediction sample. By using a sharpened Jensen's inequality we show that the attention factor is completely defined by a correlation function of the two samples w.r.t. the model weights. Due to the attention factor the solution for a prediction sample is mostly defined by a few training samples that are correlated with the prediction sample. Finding a specific solution per prediction sample couples together the training and the prediction. To make the approach practical we introduce a latent space to map each prediction sample to a latent space and learn all possible solutions as a function of the latent space along with learning attention as a function of the latent space and a training sample. The latent space plays a role of the context representation with a prediction sample defining a context and a learned context dependent solution used for the prediction.
While the beta-VAE family is aiming to find disentangled representations and acquire human-interpretable generative factors, like what an ICA (from the linear domain) does, we propose Full Encoder, a novel unified autoencoder framework as a correspondence to PCA in the non-linear domain. The idea is to train an autoencoder with one latent variable first, then involve more latent variables progressively to refine the reconstruction results. The Full Encoder is also a latent variable predictive model that the latent variables acquired are stable and robust, as they always learn the same representation regardless of the network initial states. Full Encoder can be used to determine the degrees of freedom in a simple non-linear system and can be useful for data compression or anomaly detection. Full Encoder can also be combined with the beta-VAE framework to sort out the importance of the generative factors, providing more insights for non-linear system analysis. These qualities will make FE useful for analyzing real-life industrial non-linear systems. To validate, we created a toy dataset with a custom-made non-linear system to test it and compare its properties to those of VAE and beta-VAE's.
Quantum machine learning -- and specifically Variational Quantum Algorithms (VQAs) -- offers a powerful, flexible paradigm for programming near-term quantum computers, with applications in chemistry, metrology, materials science, data science, and mathematics. Here, one trains an ansatz, in the form of a parameterized quantum circuit, to accomplish a task of interest. However, challenges have recently emerged suggesting that deep ansatzes are difficult to train, due to flat training landscapes caused by randomness or by hardware noise. This motivates our work, where we present a variable structure approach to build ansatzes for VQAs. Our approach, called VAns (Variable Ansatz), applies a set of rules to both grow and (crucially) remove quantum gates in an informed manner during the optimization. Consequently, VAns is ideally suited to mitigate trainability and noise-related issues by keeping the ansatz shallow. We employ VAns in the variational quantum eigensolver for condensed matter and quantum chemistry applications, in the quantum autoencoder for data compression and in unitary compilation problems showing successful results in all cases.
The underlying theme of this paper is to explore the various facets of power systems data through the lens of graph signal processing (GSP), laying down the foundations of the Grid-GSP framework. Grid-GSP provides an interpretation for the spatio-temporal properties of voltage phasor measurements, by showing how the well-known power systems modeling supports a generative low-pass graph filter model for the state variables, namely the voltage phasors. Using the model we formalize the empirical observation that voltage phasor measurement data lie in a low-dimensional subspace and tie their spatio-temporal structure to generator voltage dynamics. The Grid-GSP generative model is then successfully employed to investigate the problems pertaining to the grid of data sampling and interpolation, network inference, detection of anomalies and data compression. Numerical results on a large synthetic grid that mimics the real-grid of the state of Texas, ACTIVSg2000, and on real-world measurements from ISO-New England verify the efficacy of applying Grid-GSP methods to electric grid data.
We propose a new simple approach for image compression: instead of storing the RGB values for each pixel of an image, we store the weights of a neural network overfitted to the image. Specifically, to encode an image, we fit it with an MLP which maps pixel locations to RGB values. We then quantize and store the weights of this MLP as a code for the image. To decode the image, we simply evaluate the MLP at every pixel location. We found that this simple approach outperforms JPEG at low bit-rates, even without entropy coding or learning a distribution over weights. While our framework is not yet competitive with state of the art compression methods, we show that it has various attractive properties which could make it a viable alternative to other neural data compression approaches.
Linear models are used in online decision making, such as in machine learning, policy algorithms, and experimentation platforms. Many engineering systems that use linear models achieve computational efficiency through distributed systems and expert configuration. While there are strengths to this approach, it is still difficult to have an environment that enables researchers to interactively iterate and explore data and models, as well as leverage analytics solutions from the open source community. Consequently, innovation can be blocked. Conditionally sufficient statistics is a unified data compression and estimation strategy that is useful for the model development process, as well as the engineering deployment process. The strategy estimates linear models from compressed data without loss on the estimated parameters and their covariances, even when errors are autocorrelated within clusters of observations. Additionally, the compression preserves almost all interactions with the the original data, unlocking better productivity for both researchers and engineering systems.
Latent variable models have been successfully applied in lossless compression with the bits-back coding algorithm. However, bits-back suffers from an increase in the bitrate equal to the KL divergence between the approximate posterior and the true posterior. In this paper, we show how to remove this gap asymptotically by deriving bits-back coding algorithms from tighter variational bounds. The key idea is to exploit extended space representations of Monte Carlo estimators of the marginal likelihood. Naively applied, our schemes would require more initial bits than the standard bits-back coder, but we show how to drastically reduce this additional cost with couplings in the latent space. When parallel architectures can be exploited, our coders can achieve better rates than bits-back with little additional cost. We demonstrate improved lossless compression rates in a variety of settings, especially in out-of-distribution or sequential data compression.
We study the neural network (NN) compression problem, viewing the tension between the compression ratio and NN performance through the lens of rate-distortion theory. We choose a distortion metric that reflects the effect of NN compression on the model output and derive the tradeoff between rate (compression) and distortion. In addition to characterizing theoretical limits of NN compression, this formulation shows that \emph{pruning}, implicitly or explicitly, must be a part of a good compression algorithm. This observation bridges a gap between parts of the literature pertaining to NN and data compression, respectively, providing insight into the empirical success of model pruning. Finally, we propose a novel pruning strategy derived from our information-theoretic formulation and show that it outperforms the relevant baselines on CIFAR-10 and ImageNet datasets.
Unleashing the full potential of massive MIMO in FDD mode by reducing the overhead of CSI feedback has recently garnered attention. Numerous deep learning for massive MIMO CSI feedback approaches have demonstrated their efficiency and potential. However, most existing methods improve accuracy at the cost of computational complexity and the accuracy decreases significantly as the CSI compression rate increases. This paper presents a novel neural network CLNet tailored for CSI feedback problem based on the intrinsic properties of CSI. CLNet proposes a forge complex-valued input layer to process signals and utilizes attention mechanism to enhance the performance of the network. The experiment result shows that CLNet outperforms the state-of-the-art method by average accuracy improvement of 5.41\% in both outdoor and indoor scenarios with average 24.1\% less computational overhead. Codes for deep learning-based CSI feedback CLNet are available at GitHub.
In a large class of deep learning models, including vector embedding models such as word and database embeddings, we observe that floating point exponent values cluster around a few unique values, permitting entropy based data compression. Entropy coding compresses fixed-length values with variable-length codes, encoding most probable values with fewer bits. We propose the EFloat compressed floating point number format that uses a variable field boundary between the exponent and significand fields. EFloat uses entropy coding on exponent values and signs to minimize the average width of the exponent and sign fields, while preserving the original FP32 exponent range unchanged. Saved bits become part of the significand field increasing the EFloat numeric precision by 4.3 bits on average compared to other reduced-precision floating point formats. EFloat makes 8-bit and even smaller floats practical without sacrificing the exponent range of a 32-bit floating point representation. We currently use the EFloat format for saving memory capacity and bandwidth consumption of large vector embedding models such as those used for database embeddings. Using the RMS error as metric, we demonstrate that EFloat provides higher accuracy than other floating point formats with equal bit budget. The EF12 format with 12-bit budget has less end-to-end application error than the 16-bit BFloat16. EF16 with 16-bit budget has an RMS-error 17 to 35 times less than BF16 RMS-error for a diverse set of embedding models. When making similarity and dissimilarity queries, using the NDCG ranking metric, EFloat matches the result quality of prior floating point representations with larger bit budgets.
We present a non-intrusive model reduction framework for linear poroelasticity problems in heterogeneous porous media using proper orthogonal decomposition (POD) and neural networks, based on the usual offline-online paradigm. As the conductivity of porous media can be highly heterogeneous and span several orders of magnitude, we utilize the interior penalty discontinuous Galerkin (DG) method as a full order solver to handle discontinuity and ensure local mass conservation during the offline stage. We then use POD as a data compression tool and compare the nested POD technique, in which time and uncertain parameter domains are compressed consecutively, to the classical POD method in which all domains are compressed simultaneously. The neural networks are finally trained to map the set of uncertain parameters, which could correspond to material properties, boundary conditions, or geometric characteristics, to the collection of coefficients calculated from an $L^2$ projection over the reduced basis. We then perform a non-intrusive evaluation of the neural networks to obtain coefficients corresponding to new values of the uncertain parameters during the online stage. We show that our framework provides reasonable approximations of the DG solution, but it is significantly faster. Moreover, the reduced order framework can capture sharp discontinuities of both displacement and pressure fields resulting from the heterogeneity in the media conductivity, which is generally challenging for intrusive reduced order methods. The sources of error are presented, showing that the nested POD technique is computationally advantageous and still provides comparable accuracy to the classical POD method. We also explore the effect of different choices of the hyperparameters of the neural network on the framework performance.
For effective human-robot teaming, it is important for the robots to be able to share their visual perception with the human operators. In a harsh remote collaboration setting, data compression techniques such as autoencoder can be utilized to obtain and transmit the data in terms of latent variables in a compact form. In addition, to ensure real-time runtime performance even under unstable environments, an anytime estimation approach is desired that can reconstruct the full contents from incomplete information. In this context, we propose a method for imputation of latent variables whose elements are partially lost. To achieve the anytime property with only a few dimensions of variables, exploiting prior information of the category-level is essential. A prior distribution used in variational autoencoders is simply assumed to be isotropic Gaussian regardless of the labels of each training datapoint. This type of flattened prior makes it difficult to perform imputation from the category-level distributions. We overcome this limitation by exploiting a category-specific multi-modal prior distribution in the latent space. The missing elements of the partially transferred data can be sampled, by finding a specific modal according to the remaining elements. Since the method is designed to use partial elements for anytime estimation, it can also be applied for data over-compression. Based on the experiments on the ModelNet and Pascal3D datasets, the proposed approach shows consistently superior performance over autoencoder and variational autoencoder up to 70% data loss.
Neural data compression has been shown to outperform classical methods in terms of $RD$ performance, with results still improving rapidly. At a high level, neural compression is based on an autoencoder that tries to reconstruct the input instance from a (quantized) latent representation, coupled with a prior that is used to losslessly compress these latents. Due to limitations on model capacity and imperfect optimization and generalization, such models will suboptimally compress test data in general. However, one of the great strengths of learned compression is that if the test-time data distribution is known and relatively low-entropy (e.g. a camera watching a static scene, a dash cam in an autonomous car, etc.), the model can easily be finetuned or adapted to this distribution, leading to improved $RD$ performance. In this paper we take this concept to the extreme, adapting the full model to a single video, and sending model updates (quantized and compressed using a parameter-space prior) along with the latent representation. Unlike previous work, we finetune not only the encoder/latents but the entire model, and - during finetuning - take into account both the effect of model quantization and the additional costs incurred by sending the model updates. We evaluate an image compression model on I-frames (sampled at 2 fps) from videos of the Xiph dataset, and demonstrate that full-model adaptation improves $RD$ performance by ~1 dB, with respect to encoder-only finetuning.
Principal Component Analysis (PCA) is well known for its capability of dimension reduction and data compression. However, when using PCA for compressing/reconstructing images, images need to be recast to vectors. The vectorization of images makes some correlation constraints of neighboring pixels and spatial information lost. To deal with the drawbacks of the vectorizations adopted by PCA, we used small neighborhoods of each pixel to form compounded pixels and use a tensorial version of PCA, called TPCA (Tensorial Principal Component Analysis), to compress and reconstruct a compounded image of compounded pixels. Our experiments on public data show that TPCA compares favorably with PCA in compressing and reconstructing images. We also show in our experiments that the performance of TPCA increases when the order of compounded pixels increases.