Paper compresses neural network weight-updates for image artifacts removal.
problem Efficiently compressing neural network weight-updates for image artifacts removal.
method Fine-tuning a pre-trained artifact removal network on target data with a compression objective that encourages sparse and quantized weight-updates.
result Achieves reconstruction quality comparable to traditional codecs at comparable bitrates.
Paper introduces compressibility loss for learning sparse neural network weights.
problem Learning highly compressible neural network weights.
method Applying a compressibility loss to minimize the negated sparsity of the signal.
result At critical points, weight vectors are ternary signals with a sparsity directly related to the objective value.
Model compression has been introduced to reduce the required hardware resources while maintaining the model accuracy. Lots of techniques for model compression, such as pruning, quantization, and low-rank approximation, have been suggested along with different inference implementation characteristics. Adopting model com…
A new method compresses deep neural networks by predicting and quantizing weights between layers.
problem Resource constraints in deep neural networks.
method Inter-Layer Weight Prediction (ILWP) and quantization based on Smoothly Varying Weight Hypothesis (SVWH).
result The method achieves higher weight compression rates at the same accuracy level.
WNQ uses weight normalization to reduce quantization error in deep neural networks.
problem High quantization error in deep neural networks due to long-tail distribution of weights.
method Weight normalization to avoid long-tail distribution and reduce quantization error.
result WNQ achieves state-of-the-art performance on CIFAR-100 and ImageNet.
FSNet compresses CNNs by sharing weights in 1D vectors.
problem Compressing deep CNNs while maintaining performance.
method Learning a 1D vector (Filter Summary) to share weights in convolutional filters.
result FSNet reduces parameter space and accelerates computation.
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.
This work compresses heavy-tailed weight matrices for tighter generalization bounds.
problem Empirical evidence linking heavy-tailed weight matrices to test set accuracy but lack of formal relationship with generalization bounds.
method Utilized the compression framework to show that heavy-tailed matrices can be compressed, resulting in sparse weight matrices.
result Demonstrated a non-vacuous generalization bound for compressed networks with heavy-tailed weight matrices.
A simple method compresses neural network weights using entropy penalties.
problem Neural network weight compression for scalability and accuracy.
method Reparameterization with learned entropy penalty for compression.
result Maximized classification accuracy and model compressibility.
This work improves neural network compression by jointly pruning weights and activations.
problem Improving deployment efficiency of deep neural networks.
method Joint regularization technique that simultaneously prunes weights and activations.
result Jointly pruned network (JPnet) achieves significant computation cost reduction.
Weight Squeezing transfers knowledge from large models to smaller ones, improving performance and speed.
problem Transfer learning and model compression for faster and more efficient training.
method Reparameterization of weights from a large model to a smaller one.
result Weight Squeezing outperforms other methods on GLUE benchmark with faster training.
This paper proposes a new weight representation scheme for efficient model compression and performance enhancement.
problem Challenges in achieving performance enhancement on devices due to irregular sparse matrix representations.
method Fine-grained and unstructured pruning method combined with structured weight encryption.
result Achieved high compression ratios and performance on various deep learning models.
The success of deep learning in numerous application domains created the de- sire to run and train them on mobile devices. This however, conflicts with their computationally, memory and energy intense nature, leading to a growing interest in compression. Recent work by Han et al. (2015a) propose a pipeline that involve…
Paper improves MIRACLE for faster, more robust neural network compression.
problem Efficiently compressing neural networks while maintaining performance.
method Introduces Mean-KL parameterization to constrain compression cost.
result Mean-KL parameterization leads to twice as fast convergence and more robust compression.
We compress large neural networks for quick adaptation to specific contexts.
problem How to quickly adapt a pretrained large neural network to specific contexts.
method Propose a Bayesian hypernetwork framework to compress the network and encourage sparsity.
result Generated compressed networks are significantly smaller than baseline methods.
The large memory requirements of deep neural networks limit their deployment and adoption on many devices. Model compression methods effectively reduce the memory requirements of these models, usually through applying transformations such as weight pruning or quantization. In this paper, we present a novel scheme for l…
Model compression has gained a lot of attention due to its ability to reduce hardware resource requirements significantly while maintaining accuracy of DNNs. Model compression is especially useful for memory-intensive recurrent neural networks because smaller memory footprint is crucial not only for reducing storage re…
This paper investigates compression techniques for deep neural networks to reduce their size without sacrificing performance.
problem Compression of large deep neural networks for resource-limited platforms.
method Weight pruning, quantization, and lossless weight matrix representations based on source coding.
result Achieved up to 165 times compression rate while maintaining or improving model performance.
Paper proposes new pruning techniques to significantly reduce DNN weights and improve model compression.
problem High computation and memory storage challenges in deep learning networks.
method Combines filter and column pruning with ADMM algorithm and introduces Network Purification and Unused Path Removal (P-RM) for post-processing.
result Achieved up to 60x compression on ResNet-18 CIFAR-10, demonstrating effectiveness of proposed methods.
DP-Net uses dynamic programming for efficient deep neural network compression.
problem Efficiently compressing deep neural networks while maintaining accuracy.
method Dynamic Programming for optimal weight quantization and clustering-friendly training.
result Achieves up to 77X compression ratio on Wide ResNet with minimal accuracy loss.
We consider a decomposition method for compressive streaming data in the context of online compressive Robust Principle Component Analysis (RPCA). The proposed decomposition solves an n-ℓ1 cluster-weighted minimization to decompose a sequence of frames (or vectors), into sparse and low-rank components, from com…
New compression methods handle biased input sequences for more accurate posterior summaries.
problem Handling biased input sequences for accurate posterior summaries.
method Stein kernel thinning, low-rank SKT, Stein recombination, Stein Cholesky.
result Achieves accurate posterior summaries with biased input sequences.
Differentiable model compression adds noise to parameters during training.
problem Model compression for efficient deployment.
method Adding pseudo quantization noise to model parameters during training.
result Compresses model by more than 8x on ImageNet with 0.3% accuracy loss.
This paper studies how to compress neural networks while maintaining accuracy.
problem Compressing a two-layer neural network with fewer nodes without losing accuracy.
method Using tools from high-dimensional probability, the authors minimize the L_2 loss between the target and compressed networks.
result The error rate of the approximation is shown as a function of input dimension and network size in the mean-field limit.
We extend quantization-aware training to extreme model compression.
problem Maximizing model accuracy with minimal model size.
method Quantize a random subset of weights during training, allowing unbiased gradients through other weights.
result Established new state-of-the-art compromises between accuracy and model size.
We consider the problem of deep neural net compression by quantization: given a large, reference net, we want to quantize its real-valued weights using a codebook with K entries so that the training loss of the quantized net is minimal. The codebook can be optimally learned jointly with the net, or fixed, as for bina…
Bayesian neural networks are compressed using feature and weight pruning based on posterior inclusion probabilities.
problem Efficiently compressing Bayesian neural networks to reduce computation cost and improve generalizability.
method Bayesian model selection principles are applied to obtain posterior inclusion probabilities for pruning and feature selection.
result Pruned models show better generalizability on simulated and real-world data.
DACE estimates covariance from compressed data, improving accuracy.
problem Estimating covariance from large, distributed data.
method Data-aware weighted sampling for unbiased estimation.
result DACE provides more accurate covariance estimation under compression.
Proposes a link between randomness and compression in deep learning.
problem Improving efficiency in deep learning training.
method Introduces a novel tomographic compression framework called Dual Tomographic Compression (DTC).
result Demonstrates high correlation between learning performance and Gibbs entropy over compression ratios.
This paper proposes a new method for efficient data compression using Bayesian neural networks.
problem Efficient compression of data represented as functions mapping coordinates to signal values.
method Overfitting variational Bayesian neural networks to the data and compressing an approximate posterior weight sample using relative entropy coding.
result Our method achieves strong performance on image and audio compression while retaining simplicity.
The current trend of pushing CNNs deeper with convolutions has created a pressing demand to achieve higher compression gains on CNNs where convolutions dominate the computation and parameter amount (e.g., GoogLeNet, ResNet and Wide ResNet). Further, the high energy consumption of convolutions limits its deployment on m…
Theoretical framework for neural network compression using sparsity norms.
problem Understanding and quantifying compressibility and accuracy trade-offs in neural networks.
method Using sparsity-sensitive ℓ_q-norm to characterize compressibility and developing adaptive pruning algorithms.
result Theoretical relationship between network sparsity and compressibility with controlled accuracy degradation.
AutoCompress automatically prunes DNNs to ultra-high compression rates without accuracy loss.
problem Efficiently compressing deep neural networks to reduce storage and computation requirements.
method Automatic hyperparameter determination, ADMM-based structured weight pruning, purification step, heuristic search.
result Achieves ultra-high pruning rates on weights and FLOPs, up to 33x in pruning rate.
This paper finds a new way to compress CNN weights, improving on pruning and quantization.
problem Improving performance and storage efficiency of CNNs.
method Identifying and exploiting repeated patterns in CNN weight tensors, using Huffman coding and block sparse matrix formats.
result Achieved compaction ratios of 1.4x to 3.1x in addition to pruning and quantization.
IMPACT optimizes LLM compression by focusing on activation importance, reducing model size up to 55.4%.
problem Resource constraints in deploying large language models (LLMs).
method IMPACT integrates activation importance into low-rank compression, optimizing for both size and accuracy.
result IMPACT achieves up to 55.4% greater model size reduction while maintaining comparable or better accuracy.
Layer fusion reduces deep neural network layers with minimal loss in accuracy.
problem Model compression to reduce neural network size and computation.
method Fusion of similar layers to reduce model size with minimal performance loss.
result Deep networks can be compressed up to 3.33x with minimal accuracy loss.
Three new efficient algorithms project vectors onto weighted l1 ball.
problem Sparse system identification and feature selection.
method Projected gradient descent algorithms with linear or highly competitive quadratic worst case complexities.
result Efficient tools for machine learning methods like compress sensing and feature selection.
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…
New method corrects quantization errors in LLMs using low-rank matrices.
problem Correcting quantization errors in large language models.
method Introducing low-rank weight matrices to correct quantized activations in LLMs.
result Reduces accuracy gap with original model by more than 50% using low-rank matrices.
PCNN prunes CNN weights efficiently for hardware acceleration.
problem Efficiently compressing CNN models for hardware acceleration.
method PCNN uses a novel Sparsity Pattern Mask (SPM) to encode and prune weights.
result PCNN achieves up to 8.4X compression with minimal accuracy loss.
Unified SVD compression fails in practical tasks, highlighting the importance of per layer activation reconstruction.
problem The failure of a unified SVD compression method in practical tasks like perplexity and accuracy.
method Unified optimization problem for SVD based compression methods, focusing on cross-layer coupling.
result Downstream metrics like perplexity and accuracy degrade severely compared to standard per layer SVD LLM.
We study the use of knowledge distillation to compress the U-net architecture. We show that, while standard distillation is not sufficient to reliably train a compressed U-net, introducing other regularization methods, such as batch normalization and class re-weighting, in knowledge distillation significantly improves …
Condensa programmatically optimizes neural network compression.
problem Finding optimal compression strategies for neural networks.
method Bayesian optimization-based algorithm for automatic sparsity inference.
result Significant memory and runtime improvements for real-world DNNs.
Neural networks compress uninformative input directions, improving test error.
problem Data lie in a high-dimensional space but labels vary along a lower-dimensional manifold.
method One-hidden layer network trained with gradient descent, analyzing weight evolution and compression.
result Compression factor λ ∼ √p improves test error, with β Feature > β Lazy.
Paper proposes a framework and algorithm for model compression in neural networks.
problem Training neural networks with model compression techniques suffers from accuracy loss and convergence issues.
method Holistic framework based on nonconvex optimization, using NN-BCD algorithm with closed-form iteration scheme.
result The proposed algorithm globally converges to a critical point at a rate of O(1/k).
This paper compresses neural networks by permuting and quantizing weights.
problem Efficiently compressing large neural networks for resource-constrained platforms.
method Permuting and quantizing weights, connecting to rate-distortion theory, and using annealed quantization.
result Significant compression with minimal accuracy loss, e.g., 40-70% reduction in gap with uncompressed model.
Compression and computational efficiency in deep learning have become a problem of great significance. In this work, we argue that the most principled and effective way to attack this problem is by adopting a Bayesian point of view, where through sparsity inducing priors we prune large parts of the network. We introduc…
EAST compresses deep ConvNets for tiny memory nodes.
problem Memory constraints in tiny devices for deep ConvNets.
method Encoding-Aware Sparse Training (EAST) with adaptive group pruning and LZ4 weight encoding.
result EAST achieves deep memory compression with lower sparsity and higher accuracy.