On-device research index

arXiv research

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.

169,341 papers · 148 categories

Trend · papers per month

count % of corpus

16 results for “MXNet”

A comprehensive study of batch construction strategies for recurrent neural networks in MXNetcs.LG

A new batch construction method for RNNs outperforms existing approaches in MXNet.

problem Improving the efficiency and performance of recurrent neural networks in MXNet.

method Proposes an alternately sorted batch construction strategy for RNNs.

result Alternately sorted batches outperform bucketing and other methods in training time and recognition performance.

batch construction recurrent neural networks MXNet mini-batch training stochastic sorting

2017-05-05abs ↗pdf ↗

Auto-Differentiating Linear Algebracs.MS

MXNet extension enables end-to-end learning of hybrid models.

problem Lack of efficient support for linear algebra primitives in deep learning systems.

method Implementation of matrix decompositions as differentiable operators in MXNet.

result Substantial reductions in implementation complexity and running time for Gaussian process and Bayesian models.

MXNet linear algebra differentiable operators machine learning Gaussian processes

2017-10-24abs ↗pdf ↗

GluonCV and GluonNLP: Deep Learning in Computer Vision and Natural Language Processingcs.LG

GluonCV and GluonNLP simplify deep learning for CV and NLP.

problem Lack of easy-to-use deep learning tools for CV and NLP.

method Developed modular APIs and pre-trained models for rapid prototyping.

result Facilitates rapid prototyping and reproducible research in CV and NLP.

deep learning computer vision natural language processing toolkits reproducible research

2019-07-09abs ↗pdf ↗

GluonTS: Probabilistic Time Series Models in Pythoncs.LG

GluonTS simplifies deep learning for time series tasks.

problem Developing and experimenting with time series models.

method Library for deep learning-based time series modeling.

result Simplified development and experimentation for common tasks.

time series deep learning forecasting anomaly detection

2019-06-12abs ↗pdf ↗

TBD: Benchmarking and Analyzing Deep Neural Network Trainingcs.LG

Proposes a new benchmark for deep neural network training.

problem Limited focus on DNN training efficiency.

method Develops TBD benchmark covering various applications and frameworks.

result Highlights inefficiencies in DNN training across different models and hardware.

Deep Neural Networks Training Benchmark Performance Analysis Memory Profiling Machine Learning Applications

2018-03-16abs ↗pdf ↗

Advbox: a toolbox to generate adversarial examples that fool neural networkscs.LG

Advbox generates adversarial examples to test neural network robustness.

problem Vulnerability of neural networks to adversarial examples.

method Tool to create adversarial examples in multiple frameworks.

result Benchmarking robustness of machine learning models.

adversarial examples neural networks robustness PaddlePaddle PyTorch

2020-01-13abs ↗pdf ↗

QR and LQ Decomposition Matrix Backpropagation Algorithms for Square, Wide, and Deep -- Real or Complex -- Matrices and Their Software Implementationmath.NA

Article presents QR and LQ decomposition algorithms for various matrix sizes and ranks.

problem Solving least squares problems in machine learning and computer vision.

method Developed novel matrix backpropagation algorithms for QR and LQ decompositions of different matrix sizes and ranks.

result Numerical stability and computational efficiency of the proposed methods.

matrix backpropagation QR decomposition LQ decomposition least squares machine learning

2020-09-19abs ↗pdf ↗

Foolbox: A Python toolbox to benchmark the robustness of machine learning modelscs.LG

Foolbox measures how robust machine learning models are against adversarial attacks.

problem Easily fooled by imperceptible input perturbations.

method Generates adversarial perturbations and quantifies model robustness.

result Minimum perturbation needed to create adversarial examples.

machine learning robustness adversarial attacks adversarial perturbations deep learning

2017-07-13abs ↗pdf ↗

Improving the Expressiveness of Deep Learning Frameworks with Recursioncs.LG

Adds recursion to deep learning frameworks for better handling of recursive data structures.

problem Lack of support for recursion in existing deep learning frameworks.

method Complements existing frameworks with recursive execution of dataflow graphs and APIs for recursive definitions.

result Recursive implementation reduces training and inference time by more effectively using resources.

deep learning recursion TensorFlow dataflow parallel computation

2018-09-04abs ↗pdf ↗

Sockeye: A Toolkit for Neural Machine Translationcs.CL

Sockeye is an open-source toolkit for neural machine translation.

problem Improving Neural Machine Translation (NMT) models and techniques.

method Scalable training and inference for three NMT architectures, including attentional, self-attentional, and fully convolutional networks.

result Sockeye achieves competitive BLEU scores across different NMT architectures, including a best score for its transformer implementation.

Neural Machine Translation NMT Toolkit Scalable Training Attentional Recurrent Networks Self-Attentional Transformers

2017-12-15abs ↗pdf ↗

Adversarial Robustness Toolbox v1.0.0cs.LG

ART is a Python library for defending ML models against adversarial threats.

problem Vulnerability of ML models to adversarial examples.

method Pre-processing, adversarial data augmentation, runtime detection.

result Provides tools to certify and verify model robustness.

Adversarial Machine Learning Defensive ML Machine Learning Security Artificial Intelligence Safety Machine Learning Defenses

2018-07-03abs ↗pdf ↗

A Model Parallel Proximal Stochastic Gradient Algorithm for Partially Asynchronous Systemscs.LG

AsyB-ProxSGD parallelizes model updates and stochastic gradient descent for large models and data.

problem Efficiently training large models and handling large datasets in parallel.

method AsyB-ProxSGD: model parallel proximal stochastic gradient algorithm for asynchronous systems.

result Achieves linear speedup with

O(K^{1/4})

number of workers for nonconvex problems.

parallel computing stochastic gradient descent model parallelism asynchronous systems large-scale machine learning

2018-10-19abs ↗pdf ↗

DL2: A Deep Learning-driven Scheduler for Deep Learning Clusterscs.LG

DL2 uses deep learning to optimize resource allocation in DL clusters.

problem Efficient resource scheduling for deep learning clusters is challenging.

method DL2 combines supervised learning and reinforcement learning to dynamically allocate resources.

result DL2 reduces average training completion time by 44.1% compared to fairness scheduler.

deep learning scheduler resource allocation deep learning-driven reinforcement learning

2019-09-13abs ↗pdf ↗

Benanza: Automatic

μ

Benchmark Generation to Compute "Lower-bound" Latency and Inform Optimizations of Deep Learning Models on GPUscs.LG

Benanza speeds up DL model optimization by automatically generating micro-benchmarks and identifying inefficiencies.

problem Slow characterization/optimization cycles for DL models on GPUs.

method Benanza includes a model processor, benchmark generator, database, and analyzer.

result Benanza identifies optimizations in parallel layer execution, cuDNN, framework inefficiency, layer fusion, and Tensor Cores.

Deep Learning Latency Optimization Benchmarking GPU DL Models

2019-11-16abs ↗pdf ↗

DLBricks: Composable Benchmark Generation to Reduce Deep Learning Benchmarking Effort on CPUs (Extended)cs.LG

DLBricks automates DL benchmarking on CPUs, reducing effort and time.

problem Lack of representative and up-to-date DL benchmarks on CPUs.

method Decomposes DL models into runnable networks, leveraging layer repetition and auto-generation.

result Accurately estimates DL model performance and speeds up benchmarking time.

Deep Learning Benchmarking CPU DLBricks Layer Repeatability

2019-11-18abs ↗pdf ↗

Spectral Learning on Matrices and Tensorscs.LG

Spectral learning extends matrix methods to tensors for better latent variable modeling.

problem Limitations of matrix-based spectral methods in capturing non-Gaussian data.

method Extend spectral decomposition to tensor-based methods for higher-order moments.

result Tensor decomposition can identify latent effects missed by matrix methods.

spectral learning matrix decomposition tensor decomposition latent variable models higher-order moments

2020-04-16abs ↗pdf ↗