A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
Memory split advantage: thinner networks outperform a single wide network.
NPAS trains neural networks with a fixed parameter budget, improving performance and compactness.
We investigate how to train kernel approximation methods that generalize well under a memory budget. Building on recent theoretical work, we define a measure of kernel approximation error which we find to be more predictive of the empirical generalization performance of kernel approximation methods than conventional me…
A new algorithm reduces memory usage for deep learning models.
Memory bandwidth bottleneck is a major challenges in processing machine learning (ML) algorithms. In-memory acceleration has potential to address this problem; however, it needs to address two challenges. First, in-memory accelerator should be general enough to support a large set of different ML algorithms. Second, it…
We propose a novel active learning framework for activity recognition using wearable sensors. Our work is unique in that it takes physical and cognitive limitations of the oracle into account when selecting sensor data to be annotated by the oracle. Our approach is inspired by human-beings' limited capacity to respond …
Recomputation algorithms collectively refer to a family of methods that aims to reduce the memory consumption of the backpropagation by selectively discarding the intermediate results of the forward propagation and recomputing the discarded results as needed. In this paper, we will propose a novel and efficient recompu…
Kernel-based methods enjoy powerful generalization capabilities in handling a variety of learning tasks. When such methods are provided with sufficient training data, broadly-applicable classes of nonlinear functions can be approximated with desired accuracy. Nevertheless, inherent to the nonparametric nature of kernel…
We introduce a new sequential Monte Carlo algorithm we call the particle cascade. The particle cascade is an asynchronous, anytime alternative to traditional particle filtering algorithms. It uses no barrier synchronizations which leads to improved particle throughput and memory efficiency. It is an anytime algorithm i…
We study the problem of identifying correlations in multivariate data, under information constraints: Either on the amount of memory that can be used by the algorithm, or the amount of communication when the data is distributed across several machines. We prove a tight trade-off between the memory/communication complex…
CogScale benchmarks AI architectures for sequential processing.
Proposes a CL technique to improve accuracy and reduce forgetting.
New algorithms for constrained online optimization with memory and predictions.
POET enables large neural network training on tiny devices with reduced energy.
Spatiotemporal sequence prediction is an important problem in deep learning. We study next-frame(s) video prediction using a deep-learning-based predictive coding framework that uses convolutional, long short-term memory (convLSTM) modules. We introduce a novel reduced-gate convolutional LSTM(rgcLSTM) architecture that…
In this paper, we consider a dynamic asset pricing model in an approximate fractional economy to address empirical regularities related to both investor protection and past information. Our newly developed model features not only in terms with a controlling shareholder who diverts a fraction of the output, but also goo…
We introduce a new sub-linear space sketch---the Weight-Median Sketch---for learning compressed linear classifiers over data streams while supporting the efficient recovery of large-magnitude weights in the model. This enables memory-limited execution of several statistical analyses over streams, including online featu…
TASO optimizes CNN models for memory-constrained devices.
This work investigates power laws in deep neural network ensembles and predicts their performance.
SketchOGD improves memory efficiency for continual learning.
Improved deep learning model deployment on tiny MCUs with mixed-precision quantization.
Time series forecasting is an important problem across many domains, including predictions of solar plant energy output, electricity consumption, and traffic jam situation. In this paper, we propose to tackle such forecasting problem with Transformer [1]. Although impressed by its performance in our preliminary study, …
RAD estimates gradients with less memory, faster than small batch sizes.
HawkesLLM models text generation with temporal influence, improving semantic alignment under limited memory.
New constraints on space and adaptivity in bandits force more batches and memory use.
We introduce interacting particle Markov chain Monte Carlo (iPMCMC), a PMCMC method based on an interacting pool of standard and conditional sequential Monte Carlo samplers. Like related methods, iPMCMC is a Markov chain Monte Carlo sampler on an extended space. We present empirical results that show significant improv…
Model compression has emerged as an important area of research for deploying deep learning models on Internet-of-Things (IoT). However, for extremely memory-constrained scenarios, even the compressed models cannot fit within the memory of a single device and, as a result, must be distributed across multiple devices. Th…
Moniqua improves SGD convergence with quantized communication.
The Long-Short-Term-Memory Recurrent Neural Networks (LSTM RNNs) are a popular class of machine learning models for analyzing sequential data. Their training on modern GPUs, however, is limited by the GPU memory capacity. Our profiling results of the LSTM RNN-based Neural Machine Translation (NMT) model reveal that fea…
This work optimizes quantization of linear models to reduce memory usage.
We introduce a DNN training technique that learns only a fraction of the full parameter set without incurring an accuracy penalty. To do this, our algorithm constrains the total number of weights updated during backpropagation to those with the highest total gradients. The remaining weights are not tracked, and their i…
Machine learning (ML) problems are often posed as highly nonlinear and nonconvex unconstrained optimization problems. Methods for solving ML problems based on stochastic gradient descent are easily scaled for very large problems but may involve fine-tuning many hyper-parameters. Quasi-Newton approaches based on the lim…
End-to-end training of automated speech recognition (ASR) systems requires massive data and compute resources. We explore transfer learning based on model adaptation as an approach for training ASR models under constrained GPU memory, throughput and training data. We conduct several systematic experiments adapting a Wa…
New method SF-AdamW trains large models without decay phases or memory overhead.
Max-norm regularizer has been extensively studied in the last decade as it promotes an effective low-rank estimation for the underlying data. However, such max-norm regularized problems are typically formulated and solved in a batch manner, which prevents it from processing big data due to possible memory budget. In th…
We propose doubly nested network(DNNet) where all neurons represent their own sub-models that solve the same task. Every sub-model is nested both layer-wise and channel-wise. While nesting sub-models layer-wise is straight-forward with deep-supervision as proposed in \cite{xie2015holistically}, channel-wise nesting has…
We propose a streaming submodular maximization algorithm "stream clipper" that performs as well as the offline greedy algorithm on document/video summarization in practice. It adds elements from a stream either to a solution set or to an extra buffer based on two adaptive thresholds, and improves by a final…
Recurrent neural networks can be difficult to train on long sequence data due to the well-known vanishing gradient problem. Some architectures incorporate methods to reduce RNN state updates, therefore allowing the network to preserve memory over long temporal intervals. To address these problems of convergence, this p…
Paper proposes efficient BNN inference flow to reduce computation and memory costs.
A new method for active learning works well across all label budgets.
Optimizer memory affects learning rate sensitivity in shuffle order, impacting fine-tuning noise.
Georgia needs a new budget code to manage fiscal policies effectively.
We present a dual subspace ascent algorithm for support vector machine training that respects a budget constraint limiting the number of support vectors. Budget methods are effective for reducing the training time of kernel SVM while retaining high accuracy. To date, budget training is available only for primal (SGD-ba…
Efficiently simulates risk budgeting portfolios using novel algorithms.
A new algorithm ThreeSieves maximizes submodular functions efficiently in streaming data.
Bridges uplift modeling and sequential decision-making with online budget allocation.
DSA efficiently allocates sparsity across layers for budgeted pruning.
MPC outperforms reactive budgeting in non-stationary return environments.