New method combines RBMs to solve complex combinatorial optimization problems.
problem Solving large-scale combinatorial optimization problems.
method Combining pretrained RBMs to create larger models, using MCMC for solution finding.
result Combined representations provide more accurate solutions for the same sample size.
A new optimization method reduces memory and compute requirements for deep learning.
problem Memory and compute constraints in second-order stochastic optimizers for deep learning.
method Proposes KrAD, a novel factorization to approximate inverse Fisher matrix without inversion, leading to KrADagrad.
result Improves performance over Shampoo for 32-bit precision and comparable/generalization on real datasets.
Minwise hashing is the standard technique in the context of search and databases for efficiently estimating set (e.g., high-dimensional 0/1 vector) similarities. Recently, b-bit minwise hashing was proposed which significantly improves upon the original minwise hashing in practice by storing only the lowest b bits of e…
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.
We propose a principled method for gradient-based regularization of the critic of GAN-like models trained by adversarially optimizing the kernel of a Maximum Mean Discrepancy (MMD). We show that controlling the gradient of the critic is vital to having a sensible loss function, and devise a method to enforce exact, ana…
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…
LSTMs and other RNN variants have shown strong performance on character-level language modeling. These models are typically trained using truncated backpropagation through time, and it is common to assume that their success stems from their ability to remember long-term contexts. In this paper, we show that a deep (64-…
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.
Quantized deep neural networks (QDNNs) are attractive due to their much lower memory storage and faster inference speed than their regular full precision counterparts. To maintain the same performance level especially at low bit-widths, QDNNs must be retrained. Their training involves piecewise constant activation func…
Tensor factorization has become an increasingly popular approach to knowledge graph completion(KGC), which is the task of automatically predicting missing facts in a knowledge graph. However, even with a simple model like CANDECOMP/PARAFAC(CP) tensor decomposition, KGC on existing knowledge graphs is impractical in res…
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.
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.
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.
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.
Reduced precision computation for deep neural networks is one of the key areas addressing the widening compute gap driven by an exponential growth in model size. In recent years, deep learning training has largely migrated to 16-bit precision, with significant gains in performance and energy efficiency. However, attemp…
In this paper, we propose to (seamlessly) integrate b-bit minwise hashing with linear SVM to substantially improve the training (and testing) efficiency using much smaller memory, with essentially no loss of accuracy. Theoretically, we prove that the resemblance matrix, the minwise hashing matrix, and the b-bit minwise…
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.
Paper proposes a 1-bit mean estimation method with near-optimal sample complexity.
problem Distributed mean estimation with 1-bit communication constraints.
method Randomized and sequentially-chosen interval queries to estimate mean.
result Sample complexity bound matches minimax lower bound with logarithmic factors.
A new method for 1-bit matrix completion that is faster and more accurate.
problem Estimating a low-rank matrix from binary observations.
method Majorization-Minimization Gauss-Newton (MMGN) method.
result MMGN outperforms existing methods in accuracy and speed.
Study on numerical reliability of AD for MaxPool in neural nets.
problem Reliability of automatic differentiation for nonsmooth operations like MaxPool.
method Investigation across precision levels and architectures on various datasets.
result Lower norms of nonsmooth Jacobians help maintain stable learning.
Learning-based hashing algorithms are ``hot topics" because they can greatly increase the scale at which existing methods operate. In this paper, we propose a new learning-based hashing method called ``fast supervised discrete hashing" (FSDH) based on ``supervised discrete hashing" (SDH). Regressing the training exampl…
Paper proposes a 1-bit quantization scheme for high-dimensional statistical estimation.
problem High-dimensional statistical estimation with limited data.
method Uniformly dithered 1-bit quantization for sparse covariance matrix estimation, sparse linear regression, and matrix completion.
result Near minimax rates in sub-Gaussian regime and improved rates in heavy-tailed regime.
A new sampling method called Restart improves both speed and quality of generative processes.
problem Balancing speed and quality in generative processes involving differential equations.
method Alternates between adding noise and following ODE, improving both speed and quality.
result Surpasses previous SDE and ODE samplers in both speed and accuracy.
Paper applies FloatSD8 to LSTM networks, reducing complexity and power.
problem Training and inference complexity of LSTM networks.
method Applied FloatSD8 for weights, 8-bit quantization for gradients/activations, reduced arithmetic precision.
result Successfully trained LSTM models with reduced complexity and preserved accuracy.
Y. Nikonorov completes a proof in a geometry paper.
problem Completing a proof in a geometry paper.
method Completing an argument from a previous proof.
result Proof of Theorem 2.5 in JGA 27 (2017) is now complete.
Study groups of order 64 and non-homeomorphic double Kodaira fibrations with same invariants.
problem Investigate finite quotients of braid groups and their quotients.
method Use algebraic and geometric methods to classify groups and construct fibrations.
result Prove that for groups of order 64, if not 32, then order is at least 64, and classify cases where equality holds.
Generative adversarial networks have been successfully applied to inpainting in natural images. However, the current state-of-the-art models have not yet been widely adopted in the medical imaging domain. In this paper, we investigate the performance of three recently published deep learning based inpainting models: co…
Improves bit error tolerance in RRAM-based BNNs without overfitting.
problem Bit errors in RRAM-based BNNs reduce accuracy and overfit to training error rates.
method Proposes straight-through gradient approximation and a novel regularizer.
result Improves BNNs' robustness to bit errors without overfitting.
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.
New scheme optimizes BMI through probabilistic and geometric shaping.
problem Optimizing bit-wise mutual information (BMI) for coded modulation.
method Joint optimization of BMI through probabilistic and geometric shaping.
result Joint optimization enables a continuum of constellation geometries and probability distributions.
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.
Improved SNNs with quantized activations outperform traditional networks.
problem Maintaining SotA accuracy in SNNs with limited bit precision.
method Interpolating between non-spiking and spiking regimes using signal processing tools.
result First hybrid SNN outperforms traditional RNNs in accuracy with reduced bit precision.
Quantum annealer speeds up RBM training for image classification.
problem Training RBM with contrastive divergence (CD) is slow and computationally expensive.
method Used D-Wave 2000Q quantum annealer to calculate model expectation of gradient learning for RBM.
result Quantum training yields similar classification performance to CD but faster.
Convolutional neural networks have recently achieved significant breakthroughs in various image classification tasks. However, they are computationally expensive,which can make their feasible mplementation on embedded and low-power devices difficult. In this paper convolutional neural network binarization is implemente…
PowerGossip compresses model differences for decentralized deep learning with low-rank linear compressors.
problem Communication bottleneck in decentralized deep learning models.
method Low-rank linear compressors applied on model differences using power iteration steps.
result Asymptotically independent of network and compression, faster convergence, and comparable performance to tuned compression algorithms.
BITS for GAPS uses Bayesian methods to improve surrogate model accuracy in complex systems.
problem Improving surrogate model accuracy in complex physical systems with uncertainty.
method Bayesian Information-Theoretic Sampling for hierarchical Gaussian Process Surrogates.
result Increased expected information gain and predictive accuracy by targeting high-uncertainty regions.
Neural network VQ-VAE with WaveNet decodes speech at 1.6 kbps with high quality.
problem Efficiently transmitting and storing speech signals at low bit-rates.
method VQ-VAE and WaveNet architecture for speech coding.
result Speech coding at 1.6 kbps with perceptual quality between MELP and AMR-WB.
While the channel capacity reflects a theoretical upper bound on the achievable information transmission rate in the limit of infinitely many bits, it does not characterise the information transfer of a given encoding routine with finitely many bits. In this note, we characterise the quality of a code (i. e. a given en…
Low bit-width integer weights and activations are very important for efficient inference, especially with respect to lower power consumption. We propose Monte Carlo methods to quantize the weights and activations of pre-trained neural networks without any re-training. By performing importance sampling we obtain quantiz…
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.
The bits-back argument suggests that latent variable models can be turned into lossless compression schemes. Translating the bits-back argument into efficient and practical lossless compression schemes for general latent variable models, however, is still an open problem. Bits-Back with Asymmetric Numeral Systems (BB-A…
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.
New algorithm learns halfspaces over hypercube with random bit flips.
problem Agnostic learning of Boolean halfspaces over discrete domains is computationally hard.
method Smoothed analysis with random bit flips for discrete inputs.
result First efficient algorithm for smoothed agnostic learning of halfspaces over Boolean hypercube.
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.
Paper proposes a log-domain training method to reduce neural network complexity.
problem High computational complexity in training deep neural networks limits real-time training.
method End-to-end training and inference scheme using approximate logarithmic operations in the log-domain.
result 16-bit log-based training achieves within 1% accuracy of floating-point baselines.
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 the description of the instanton Floer homology of a surface times a circle due to Muñoz, we compute the nilpotency degree of the endomorphism u2−64. We then compute the framed instanton homology of a surface times a circle with non-trivial bundle, which is closely related to the kernel of u2−64. We discuss th…