Bit-Swap improves lossless compression for hierarchical latent variable models.
problem Efficient lossless compression for latent variable models with hierarchical structure.
method Generalizes bits-back coding to hierarchical latent variable models with Markov chain structure.
result Achieves superior lossless compression rates for hierarchical latent variable models.
Generalizes bits back coding for time-series models with latent Markov structures.
problem Efficiently compressing time-series data with latent Markov structures.
method Extends bits back coding to time-series models with latent Markov structures, including HMMs and LGSSMs.
result Effective for small scale models, promising for larger scale settings like video compression.
A new method compresses MNIST dataset with VAE at near optimal rate.
problem Practical lossless compression with latent variable models.
method Bits Back with ANS (BB-ANS) for near optimal compression rate.
result Achieved superior compression rates to standard methods.
New method improves image compression using bits-back coding.
problem Lossy image compression with deep latent variable models.
method Iterative inference, stochastic annealing, bits-back coding.
result New state-of-the-art performance on lossy image compression.
SHVC improves image compression with fewer parameters.
problem Challenges in VAE compression, especially with bits-back coding.
method Introduces autoregressive sub-pixel convolution and autoregressive initial bits.
result Achieves state-of-the-art compression performance with fewer model parameters.
New technique for flow models achieves theoretical compression lengths.
problem No guaranteed computationally efficient codes for flow models.
method Local bits-back coding for flow models.
result Efficient algorithms achieve theoretical codelengths for flow models.
A new method, REC, compresses images by encoding their latent representations efficiently.
problem Efficiently compressing single images with latent representations.
method Relative Entropy Coding (REC) that directly encodes latent representations with codelength close to relative entropy.
result REC is more efficient for single image compression compared to previous methods and is competitive for lossy compression.
This paper simplifies ANS for statisticians, making it easier to use.
problem Statisticians struggle to understand ANS and its versatility.
method Present ANS from a latent variable model perspective and guide step-by-step implementation.
result Made ANS more accessible for statisticians through a Python implementation and library.
The state-of-the-art hardware platforms for training Deep Neural Networks (DNNs) are moving from traditional single precision (32-bit) computations towards 16 bits of precision -- in large part due to the high energy efficiency and smaller bit storage associated with using reduced-precision representations. However, un…
Paper proposes training deep neural networks with 8-bit floating point precision.
problem Challenges in training deep neural networks at 8-bit precision due to higher precision and dynamic range requirements.
method Proposes a method to train deep neural networks using 8-bit floating point for weights, activations, errors, and gradients. Introduces an enhanced loss scaling method and stochastic rounding technique.
result Demonstrates state-of-the-art accuracy across multiple datasets and workloads compared to full precision baseline.
Develops a method for lossless compression using latent variable models.
problem Lossless compression of large datasets.
method Bits back with asymmetric numeral systems (BB-ANS) using latent variable models.
result Achieves state-of-the-art lossless compression of full-size colour images.
While deep neural networks are a highly successful model class, their large memory footprint puts considerable strain on energy consumption, communication bandwidth, and storage requirements. Consequently, model size reduction has become an utmost goal in deep learning. A typical approach is to train a set of determini…
ActNN reduces neural network training memory by 2-bit quantization.
problem Limited memory in training neural networks.
method Randomly quantizes activations to 2 bits, proving convergence and proposing mixed-precision strategies.
result Reduces activation memory footprint by 12x with negligible accuracy loss.
Task offloading is an emerging technology in fog-enabled networks. It allows users to transmit tasks to neighbor fog nodes so as to utilize the computing resources of the networks. In this paper, we investigate a stochastic task offloading model and propose a multi-armed bandit framework to formulate this model. We con…
MSD removes dequantization bottleneck in LLM inference by approximating high-precision activations.
problem Dequantization bottleneck in LLM inference on modern AI accelerators.
method MSD decomposes high-precision activations into multiple low-precision components for direct multiplication with quantized weights.
result MSD avoids INT8-to-BF16 weight conversion, reducing dequantization cycles and HBM traffic.
A new method quantizes neural networks to low-precision without STE, improving accuracy.
problem Quantization of neural networks to low-precision without a complete theoretical understanding.
method Alpha-blending (AB) using stochastic gradient descent (SGD) to quantize weights and gradually increase the coefficient α. result Improves top-1 accuracy by 0.9% on 1-bit BinaryNet, 0.82% on 8-bit MobileNet v1, and 2.93% on 4-bit ResNet_50 v1/2 compared to STE.
A new backprop method reduces memory usage for deep neural networks.
problem Memory-intensive backpropagation in deep neural networks.
method Uses approximations of activations in buffers to reduce memory usage.
result Performance close to exact training with 4-bit precision activations.
This paper proposes a learning framework for n-bit quantized neural networks that improves accuracy and speed on FPGAs.
problem Efficiently implementing quantized neural networks on FPGAs to maintain accuracy and speed.
method A novel learning framework for n-bit QNNs, constrained weights, reconstructed gradient function, n-BQ-NN structure, and SVPE array.
result Quantized models achieve almost the same accuracy as full-precision models and outperform typical low-precision QNNs.
New techniques improve 16-bit training accuracy without 32-bit units.
problem Training deep learning models with only 16-bit floating-point units.
method Studied BFloat16 units and applied stochastic rounding and Kahan summation techniques.
result Up to 7% absolute validation accuracy gain in 16-bit-FPU training.
HiLLoC compresses large images losslessly using VAEs.
problem Lossless compression of large color photographs.
method Fully convolutional VAE models trained on ImageNet are applied to lossless compression.
result Achieves state-of-the-art compression for full-size ImageNet images.
Bayesian Bits unifies quantization and pruning through gradient optimization.
problem Joint mixed precision quantization and pruning for efficient neural networks.
method Gradient-based optimization with a novel bit width decomposition and learnable stochastic gates.
result Bayesian Bits achieves better accuracy vs. efficiency trade-off compared to static bit width networks.
Low-bit training framework reduces energy consumption in CNNs.
problem Reducing energy consumption in convolutional neural networks.
method Low-bit training framework using MLS tensor format with dynamic quantization.
result Achieves superior trade-off between accuracy and bit-width.
To make deep neural networks feasible in resource-constrained environments (such as mobile devices), it is beneficial to quantize models by using low-precision weights. One common technique for quantizing neural networks is the straight-through gradient method, which enables back-propagation through the quantization ma…
Paper improves DNN accelerator robustness against bit errors with energy savings.
problem Bit errors in quantized DNN weights reduce energy efficiency.
method Combines robust fixed-point quantization, weight clipping, and random bit error training.
result Significantly improves robustness against random bit errors with high energy savings.
Bit-slice sparsity improves ReRAM-based DNN acceleration.
problem Limited ADC power and area constraints in ReRAM-based DNN accelerators.
method Proposed bit-slice L1 algorithm to induce sparsity during training.
result 2x sparsity improvement compared to previous methods.
Estimating mean from one-bit samples of symmetric log-concave distributions.
problem Estimating the mean of a symmetric log-concave distribution with limited one-bit measurements.
method Analyzes mean squared error in three settings: centralized, adaptive, and distributed, with and without quantization.
result One round of adaptivity is sufficient to achieve optimal mean-square error in the adaptive setting.
Majority bit estimation in noisy random recursive DAGs.
problem Estimating the majority bit in a noisy random recursive DAG.
method Majority rule among nodes, with bit flipping and noisy channel.
result Identification of the threshold for p at which majority rule yields errors. Paper proposes a CNN-based method for estimating intra frame bits and quality.
problem Efficient video delivery and bit allocation in video coding.
method Deep learning approach using CNNs trained on original frames and encoded distortions.
result Accurate estimation of intra frame bits and quality for better bit allocation.
Improves matrix multiplication throughput for asymmetric bit-width operands.
problem Matrix multiplications between asymmetric bit-width operands, especially 8- and 4-bit, are not efficiently handled by existing SIMD instructions.
method Proposes a new SIMD matrix multiplication instruction that uses mixed precision on inputs (8- and 4-bit) and accumulates into 16-bit output, improving throughput.
result Offers 2x improvement in throughput compared to existing symmetric-operand-size instructions, with negligible overflow.
In our previous work we have shown that resistive cross point devices, so called Resistive Processing Unit (RPU) devices, can provide significant power and speed benefits when training deep fully connected networks as well as convolutional neural networks. In this work, we further extend the RPU concept for training re…
Paper proposes a hybrid model-based and data-driven approach for one-bit compressive autoencoding.
problem Designing efficient one-bit compressive autoencoding models for complex systems.
method Hybrid model-based and data-driven methodology for one-bit sparse signal recovery.
result Significant improvement in one-bit compressive autoencoding compared to state-of-the-art algorithms.
Paper proposes a hybrid model-based and data-driven method for one-bit compressive variational autoencoding.
problem Designing efficient one-bit compressive sensing systems.
method Hybrid model-based and data-driven approach for one-bit compressive variational autoencoding.
result Significant improvement in one-bit compressive sensing compared to state-of-the-art methods.
New protocols show 1-bit mean estimation can be order-optimal without interaction.
problem Can 1-bit mean estimation be optimal without interaction?
method Adaptive and non-adaptive threshold and interval queries, with one adaptive transition.
result Arbitrary non-adaptive quantizers can match the adaptive rate, suggesting interaction is not necessary.
Topological data analysis classifies encrypted bits with success.
problem Classifying encrypted data with traditional machine learning methods.
method Persistent homology for generating topological features, machine learning pipeline.
result Successfully classifies encrypted data, outperforming classical models.
Paper offers robust recovery for 1-bit sensing with partial Gaussian circulant matrices.
problem Accurately recovering vectors from 1-bit measurements using structured matrices.
method Correlation-based optimization with randomly signed partial Gaussian circulant matrices and generative models.
result Recovery guarantees match those for i.i.d. Gaussian matrices but with faster computation.
One-bit feedback suffices for a bandit problem's optimal strategy.
problem Optimal strategy for multi-armed bandit problem with limited feedback.
method Coding and decoding schemes for one-bit feedback to mimic full-reward feedback.
result Regret ratio approaches 1 with one-bit feedback.
New algorithm tackles batched stochastic linear bandits with 1-bit communication constraints.
problem Stochastic linear bandits with 1-bit communication constraints.
method Phased-elimination algorithms based on G-optimal designs and 1-bit mean estimation.
result Achieves near-optimal regret bounds for broad scaling regimes.
Implementing large-scale deep neural networks with high computational complexity on low-cost IoT devices may inevitably be constrained by limited computation resource, making the devices hard to respond in real-time. This disjunction makes the state-of-art deep learning algorithms, i.e. CNN (Convolutional Neural Networ…
Moniqua improves SGD convergence with quantized communication.
problem Efficiently communicating in decentralized SGD with limited bandwidth.
method Modulo quantized communication in decentralized SGD.
result Moniqua converges at the same rate as full-precision communication with less bits.
Study 1-bit compressive sensing with generative models, improving recovery accuracy.
problem Accurately recover sparse vectors from binary measurements with generative models.
method Analyzes noiseless and noisy 1-bit measurements with i.i.d.~Gaussian and Lipschitz continuous generative priors, proving sample complexity bounds and stability properties.
result Proves sample complexity bounds and stability properties for 1-bit compressive sensing with generative models.
ADD embeds a 48-bit message into images, achieving high accuracy and speed.
problem Embedding high-fidelity messages into images to detect authenticity and source.
method Two-stage process: linear combination and addition of watermark to image, followed by decoding.
result ADD achieves 100% decoding accuracy for 48-bit watermarking, with minimal performance drop under various distortions.
This letter proposes a dictionary learning algorithm for blind one bit compressed sensing. In the blind one bit compressed sensing framework, the original signal to be reconstructed from one bit linear random measurements is sparse in an unknown domain. In this context, the multiplication of measurement matrix $\Ab$ an…
This paper introduces new methods to improve 1-bit matrix completion by considering cluster effects.
problem Improving 1-bit matrix completion for clustered data.
method Group-Specific 1-bit Matrix Completion (GS1MC) and Cluster Developing Matrix Completion (CDMC).
result GS1MC and CDMC outperform existing methods in synthetic and real-world data.
FleXOR trains fractional quantization for neural networks, improving accuracy and size.
problem Quantization limits to integer bits restricts compression and accuracy.
method Encryption algorithm with XOR gates for fractional bits during inference.
result FleXOR achieves high accuracy with fractional sub-1-bit weights.
Paper studies distributed learning with limited communication bits, achieving optimal error exponents.
problem Distributed hypothesis testing with constant communication bits.
method Geometric approach in distribution spaces, encoding empirical distributions to transmission bits.
result Optimal achievable error exponents and coding schemes for various communication constraints.
WrapNet optimizes inference for low-resolution neural networks by using 8-bit additions.
problem Reducing multiplication complexity in low-resolution neural networks.
method Adapting neural networks to use low-resolution (8-bit) additions in accumulators, with a cyclic activation layer and overflow penalty regularizer.
result Achieves comparable classification accuracy to 32-bit counterparts using low-resolution additions.
The method of random projections has become a standard tool for machine learning, data mining, and search with massive data at Web scale. The effective use of random projections requires efficient coding schemes for quantizing (real-valued) projected data into integers. In this paper, we focus on a simple 2-bit coding …
Kernel Quantization improves CNN compression without sacrificing performance.
problem Efficiently compressing CNN models without significant performance loss.
method Quantizes convolution kernels as the unit, learning a codebook for low-bit indexes.
result Significant compression ratio achieved with minimal accuracy loss.