arXiv research

PowerGossip compresses model differences for decentralized deep learning with low-rank linear compressors.

This paper extends financial theory to measure learnable market structure under computational constraints.

For fast and energy-efficient deployment of trained deep neural networks on resource-constrained embedded hardware, each learned weight parameter should ideally be represented and stored using a single bit. Error-rates usually increase when this requirement is imposed. Here, we report large improvements in error rates …

Moniqua improves SGD convergence with quantized communication.

We present an overview of techniques for quantizing convolutional neural networks for inference with integer weights and activations. Per-channel quantization of weights and per-layer quantization of activations to 8-bits of precision post-training produces classification accuracies within 2% of floating point networks…

New research suggests privileged information doesn't improve model performance.

Two potential bottlenecks on the expressiveness of recurrent neural networks (RNNs) are their ability to store information about the task in their parameters, and to store information about the input history in their units. We show experimentally that all common RNN architectures achieve nearly the same per-task and pe…

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…

Paper develops an efficient mean estimator for 1-bit communication constraints.

New scheme optimizes BMI through probabilistic and geometric shaping.

Paper optimizes KWS models using NAS and quantization for limited resources.

It is of fundamental importance to find algorithms obtaining optimal performance for learning of statistical models in distributed and communication limited systems. Aiming at characterizing the optimal strategies, we consider learning of Gaussian Processes (GPs) in distributed systems as a pivotal example. We first ad…

New algorithm tackles batched stochastic linear bandits with 1-bit communication constraints.

We generated a dataset of 200 GB with 10^9 features, to test our recent b-bit minwise hashing algorithms for training very large-scale logistic regression and SVM. The results confirm our prior work that, compared with the VW hashing algorithm (which has the same variance as random projections), b-bit minwise hashing i…

Binary neural networks (BNN) have been studied extensively since they run dramatically faster at lower memory and power consumption than floating-point networks, thanks to the efficiency of bit operations. However, contemporary BNNs whose weights and activations are both single bits suffer from severe accuracy degradat…

In recent work, Cheu et al. (Eurocrypt 2019) proposed a protocol for $n$-party real summation in the shuffle model of differential privacy with $O_{ε, δ}(1)$ error and $Θ(ε\sqrt{n})$ one-bit messages per party. In contrast, every local model protocol for real summation must incur error $Ω(1/\sqrt{n})$, and there exist …

We consider the problem of estimating the mean of a symmetric log-concave distribution under the constraint that only a single bit per sample from this distribution is available to the estimator. We study the mean squared error as a function of the sample size (and hence the number of bits). We consider three settings:…

FleXOR trains fractional quantization for neural networks, improving accuracy and size.

Differentiable model compression adds noise to parameters during training.

Our work specifies the fundamental cost of using secure aggregation in federated learning.

Fine-tuning large pre-trained models is an effective transfer mechanism in NLP. However, in the presence of many downstream tasks, fine-tuning is parameter inefficient: an entire new model is required for every task. As an alternative, we propose transfer with adapter modules. Adapter modules yield a compact and extens…

Study quantile reward identification with 1-bit feedback constraints.

Improves matrix multiplication throughput for asymmetric bit-width operands.

Despite the recent success of deep transfer learning approaches in NLP, there is a lack of quantitative studies demonstrating the gains these models offer in low-shot text classification tasks over existing paradigms. Deep transfer learning approaches such as BERT and ULMFiT demonstrate that they can beat state-of-the-…

Learning with few samples is a major challenge for parameter-rich models like deep networks. In contrast, people learn complex new concepts even from very few examples, suggesting that the sample complexity of learning can often be reduced. Many approaches to few-shot learning build on transferring a representation fro…

Extends hyperparameter transfer across model sizes and modules, improving training speed.

Study explores efficient continual learning of diffusion models.

Distributed stochastic gradient descent is an important subroutine in distributed learning. A setting of particular interest is when the clients are mobile devices, where two important concerns are communication efficiency and the privacy of the clients. Several recent works have focused on reducing the communication c…

Enhances LLM quantization with MDBF, improving perplexity and accuracy.

New MIMO constellation design for noncoherent communications reduces hardware complexity.

In this work, a classification method for SSVEP-based BCI is proposed. The classification method uses features extracted by traditional SSVEP-based BCI methods and finds optimal discrimination thresholds for each feature to classify the targets. Optimising the thresholds is formalised as a maximisation task of a perfor…

This paper introduces a new member of the family of Variational Autoencoders (VAE) that constrains the rate of information transferred by the latent layer. The latent layer is interpreted as a communication channel, the information rate of which is bound by imposing a pre-set signal-to-noise ratio. The new constraint s…

Stochastic gradient descent (SGD) is one of the most widely used optimization methods for parallel and distributed processing of large datasets. One of the key limitations of distributed SGD is the need to regularly communicate the gradients between different computation nodes. To reduce this communication bottleneck, …

SHVC improves image compression with fewer parameters.

Autoregressive generative models consistently achieve the best results in density estimation tasks involving high dimensional data, such as images or audio. They pose density estimation as a sequence modeling task, where a recurrent neural network (RNN) models the conditional distribution over the next element conditio…

New method improves accuracy of quantized neural networks.

DVAO predicts volumetric ambient occlusion for real-time volume rendering.

Optimization results are one method for understanding neural computation from Nature's perspective and for defining the physical limits on neuron-like engineering. Earlier work looks at individual properties or performance criteria and occasionally a combination of two, such as energy and information. Here we make use …

In this work, a deep learning-based quantization scheme for log-likelihood ratio (L-value) storage is introduced. We analyze the dependency between the average magnitude of different L-values from the same quadrature amplitude modulation (QAM) symbol and show they follow a consistent ordering. Based on this we design a…

Improved 2-bit covariance estimator with reduced operator norm error and no tuning needed.

New constraints on space and adaptivity in bandits force more batches and memory use.

Paper proves existence of a universal codebook for low-precision quantization.

Conventional hardware-friendly quantization methods, such as fixed-point or integer, tend to perform poorly at very low word sizes as their shrinking dynamic ranges cannot adequately capture the wide data distributions commonly seen in sequence transduction models. We present AdaptivFloat, a floating-point inspired num…

We consider the problem of transfer learning in an online setting. Different tasks are presented sequentially and processed by a within-task algorithm. We propose a lifelong learning strategy which refines the underlying data representation used by the within-task algorithm, thereby transferring information from one ta…

This paper resolves BIHT convergence, showing normalization is not necessary in noiseless settings but crucial for robustness.

We consider the problem of estimating the arithmetic average of a finite collection of real vectors stored in a distributed fashion across several compute nodes subject to a communication budget constraint. Our analysis does not rely on any statistical assumptions about the source of the vectors. This problem arises as…

Transferring knowledge across a sequence of reinforcement-learning tasks is challenging, and has a number of important applications. Though there is encouraging empirical evidence that transfer can improve performance in subsequent reinforcement-learning tasks, there has been very little theoretical analysis. In this p…

BITS for GAPS uses Bayesian methods to improve surrogate model accuracy in complex systems.