This paper highlights new opportunities for designing large-scale machine learning systems as a consequence of blurring traditional boundaries that have allowed algorithm designers and application-level practitioners to stay -- for the most part -- oblivious to the details of the underlying hardware-level implementatio…
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.
169,291 papers · 148 categories
Trend · papers per month
48 results for “machine learning hardware”
New method bounds hardware noise without assumptions.
problem Estimating hardware noise without assumptions.
method Machine Learning and Conformal Prediction.
result Theoretical upper bounds of fidelity.
Deploying Customized Data Representation and Approximate Computing in Machine Learning Applicationscs.LG
Lop enables efficient machine learning models through customizable data representations and approximate computing.
problem Efficient deployment of machine learning models with reduced computational and storage complexity.
method Integrates low-precision representation and approximate arithmetic operations into machine learning frameworks and hardware.
result Lop facilitates high-level simulations and hardware realization of machine learning models with customizable data representations and approximate arithmetic operations.
Hardware Trojans can fool neural networks, affecting security.
problem Security vulnerabilities in hardware optimized neural networks.
method Developed a framework for inserting malicious hardware Trojans in neural network classifiers.
result Trojans can fool neural networks with minimal impact, affecting 0.03% of neurons.
Survey of QML applications on near-term quantum devices.
problem Achieving quantum advantage on real-world applications.
method Analysis of supervised and unsupervised techniques, including encoding, ansatz structure, error mitigation, and gradient methods.
result Current QML implementations on quantum hardware face limitations but show potential for real-world applications.
FPGA speeds up ICA by orders of magnitude.
problem Slow convergence of adaptive ICA algorithms.
method Equivariant adaptive separation via independence algorithm.
result FPGA implementation improves clock frequency and throughput.
Introducing ReQuEST: an Open Platform for Reproducible and Quality-Efficient Systems-ML Tournamentsstat.ML
ReQuEST platform aims to simplify ML system evaluation across diverse hardware and software.
problem Complexity and time-consuming evaluation of machine learning systems across frameworks and hardware.
method Open co-design tournament platform with online public scoreboard, leveraging Collective Knowledge and ACM artifact methodology.
result Validation and sharing of reproducible machine learning system implementations.
Quantum hardware accelerates training of Boltzmann machines, improving sampling and learning.
problem Training fully visible Boltzmann machines with high-energy barriers.
method Benchmarked quantum annealing hardware for training Boltzmann machines, comparing quantum and classical distributions.
result Quantum hardware can improve training of Boltzmann machines, especially for hard problems.
The paper tackles hardware efficiency in DL models, predicting and optimizing for latency and energy cost.
problem Predicting and optimizing hardware efficiency for DL models during inference.
method Develops predictive models and hardware-aware optimization techniques.
result Predictive models and optimization techniques can significantly improve hardware efficiency in DL applications.
This paper uses machine learning to assist automation engineers in decision making.
problem Imperfect decision making by automation engineers leads to multiple iterations and increased time for software development.
method Defined challenges and proposed solutions using machine learning for automation engineering, including code classification, finding similar code snippets, and hardware selection.
result Paragraph embedding techniques achieved an F1-score of 72% for classifying automation using code snippets, and autoencoder models for hardware recommendation achieved p@3 and p@5 of 0.79 and 0.95, respectively.
Improved machine translation with INT8 hardware using a novel training method.
problem Training accurate machine translation models with limited hardware precision.
method Convert all Transformer matrix multiplications to 8-bit integer (INT8) without sacrificing accuracy.
result INT8 Transformer models achieve BLEU scores 99.3% to 100% relative to FP32 models.
RandNLA uses randomness for matrix problems in machine learning.
problem Matrix problems in machine learning.
method Randomized Numerical Linear Algebra.
result New challenges in RandNLA due to hardware trends and advances in ML.
Myelin framework trains deep learning models on private data.
problem Training deep learning models on private data without revealing the data.
method Combines privacy-preserving techniques like differential privacy and trusted hardware enclaves.
result Establishes a baseline for fully private machine learning performance.
The thesis explores how to integrate machine learning with hardware constraints.
problem Designing efficient neural networks for real-time processing with hardware limitations.
method Developed a library for training and converting sparse quantized neural networks to hardware.
result Demonstrated how to design and optimize neural networks for FPGA-based hardware.
This paper characterizes deep learning models in Facebook's data centers and suggests optimizations.
problem Improving performance of deep learning models in data centers.
method Detailed characterizations, high performance optimizations, co-design suggestions.
result Need for better co-design of algorithms, numerics, and computing platforms.
SynergicLearning combines NN and HD models for high accuracy and efficiency.
problem Combining neural networks and hyperdimensional learning for improved accuracy and efficiency.
method Hybrid model combining NN feature extraction and HD classification, parameterized hardware implementation.
result Improves accuracy by at least 10% compared to HD learning models and 1.60x power efficiency.
CAP-BM learns complex-valued data's amplitude and phase distributions.
problem Learning from complex-valued data with amplitude variation.
method Complex Amplitude-Phase Boltzmann machine (CAP-BM) with Gibbs sampling.
result Necessity of amplitude-amplitude coupling term in CAP-BM.
Hardware-accelerated RBM solves large combinatorial problems and integer factorization.
problem Solving large combinatorial optimization and integer factorization problems.
method Logically synthesized RBM architecture, hardware acceleration, and efficient training methods.
result Hardware-accelerated RBM factorizes 16-bit numbers with 10000x speed and 32x power improvements.
NASCaps automates CapsNet design for better accuracy and hardware efficiency.
problem Designing Capsule Networks is laborious and inefficient.
method Automated Neural Architecture Search (NAS) with Genetic Algorithm optimization.
result Jointly optimizes network accuracy and hardware efficiency.
Machine learning improves PMD compensation in multiplexed systems.
problem Improving performance in multiplexed systems with PMD.
method Model-based machine learning parameterizing the Manakov-PMD equation.
result Performance close to PMD-free case achieved with hardware-friendly DBP and PMD compensation.
Paper tackles noisy neural networks and proposes a method to enhance their robustness.
problem Noisy neural networks struggle with random continuous noise in weights.
method Knowledge distillation combined with noise injection during training.
result Models achieve up to twice greater noise tolerance.
XSP profiles ML models across hardware and software stacks.
problem Challenges in profiling ML model performance across different layers of the stack.
method XSP uses distributed tracing to aggregate data from various sources and introduces a leveled, iterative measurement approach.
result XSP provides insights into ML model execution not easily discernible otherwise.
MLSys aims to bridge ML and systems research.
problem Designing ML systems for real-world deployment is challenging.
method Foster a new conference and research community.
result MLSys conference focuses on intersection of systems and ML.
Sparse evolutionary Deep Learning with over one million artificial neurons on commodity hardwarecs.NE
Sparse neural networks trained on commodity hardware achieve high accuracy for high-dimensional data.
problem Training and deploying modern ANNs on commodity hardware due to model size and data volume challenges.
method Introducing a technique to train truly sparse neural networks with fixed parameter count throughout training.
result Achieved higher accuracy than traditional methods and trained models with over 1 million neurons on a typical laptop.
Quantum reservoir computing improves volatility forecasting.
problem Forecasting realized volatility in finance.
method Quantum reservoir computing with Ising Hamiltonian and feature selection.
result Quantum reservoir computing outperforms benchmarks in volatility forecasting.
AMPNet accelerates deep learning training on interconnected devices.
problem Training efficiency for dynamic neural networks with complex control flow.
method Asynchronous model-parallel training algorithm.
result AMP training achieves similar accuracy but faster convergence and shorter training times.
Paper proposes a new ML approach using only additions and thresholding.
problem Energy efficiency and reduced complexity for IoT ML devices.
method Margin-Propagation (MP) network for inference and learning without MVMs.
result MP-based classifiers achieve comparable results to traditional ML methods with energy savings.
New method for efficient sketching of gradients and Hessians.
problem Memory constraints in training machine learning models.
method A novel framework for scalable gradient and HVP sketching tailored for modern hardware.
result Theoretical guarantees and practical applications in training data attribution and Hessian spectrum analysis.
Pipelined Backpropagation trains large models without batches efficiently.
problem Training large models efficiently on hardware with limited batch sizes.
method Fine-grained Pipelined Backpropagation with Spike Compensation and Linear Weight Prediction.
result Fine-grained Pipelined Backpropagation with a batch size of one matches the accuracy of SGD for multiple networks.
RNNs learn device models from input/output data.
problem Learning complex device models from limited data.
method Empirical study using RNNs to model six different devices.
result RNNs can generate functional software-only models of hardware devices.
Paper finds optimal mini-batch size for SGD to speed up learning.
problem Optimizing mini-batch size for faster SGD convergence.
method Empirical inverse law and theoretical bound on mini-batch SGD training.
result An accurate model for predicting training time and identifying implications for algorithm and hardware.
Paper proposes efficient INT4 quantization for neural networks.
problem Efficient inference on limited hardware resources for large neural networks.
method Formalizes quantization as MMSE problem, optimizes constrained MSE at each layer, partitions parameters, uses multiple quantized tensors.
result 4-bit integer (INT4) quantization yields state-of-the-art results with minimal accuracy loss.
New neuromorphic hardware learns MNIST digits efficiently.
problem Limited scalability and in-hardware learning in existing neuromorphic hardware.
method Low-cost scalable NoC-based SNN architecture with in-hardware STDP learning.
result Demonstrated learning capability of the hardware architecture.
This paper reviews quantum machine learning from NISQ to fault tolerance.
problem The challenges and opportunities in quantum machine learning.
method Comprehensive review of quantum machine learning concepts.
result Coverage of NISQ and fault-tolerant quantum computing approaches.
Framework learns to optimize tensor programs for various hardware.
problem Manual optimization of tensor operators for deep learning limits applicability and increases engineering costs.
method Learning-based statistical cost models guide tensor operator implementations over billions of variants.
result Framework delivers performance competitive with hand-tuned libraries across multiple hardware targets.
On the effects of biased quantum random numbers on the initialization of artificial neural networksquant-ph
Study finds no significant difference in neural network weights with quantum random numbers.
problem Effects of biased quantum random numbers on neural network initialization.
method Empirical study using quantum hardware and classical pseudo-random numbers.
result No statistically significant difference found between quantum random numbers and other types.
Quantum Machine Learningquant-ph
Quantum machine learning aims to leverage quantum computing for better pattern recognition.
problem Finding patterns in data efficiently using classical computers.
method Developing and implementing quantum software to outperform classical machine learning.
result Quantum computers can potentially outperform classical ones in machine learning tasks.
Paper examines power consumption in neural networks using various activation functions.
problem Power consumption in machine learning models.
method Examines power consumption for different activation functions.
result Substantial differences in power consumption exist between activation functions.
A hybrid framework reduces ML complexity on edge devices.
problem Limited memory and energy on edge devices.
method Compressed data collection and tailored deep learning network.
result Significant reduction in computational complexity and memory.
MLModelScope streamlines ML/DL model evaluation and benchmarking.
problem Challenges in evaluating and benchmarking ML/DL models.
method Open-source, framework/hardware agnostic platform with distributed design.
result Demonstrates the impact of model evaluation pipelines and HW/SW choices.
TonY simplifies distributed ML job management.
problem Managing distributed ML jobs is complex and resource-intensive.
method TonY is an open-source orchestrator for distributed ML jobs.
result TonY simplifies distributed ML job management.
New deep learning techniques improve HEP model performance and speed.
problem Improving classification model performance and training speed in HEP.
method Advanced deep learning techniques including data augmentation, learning rate scheduling, ensembling, and new architectures.
result Achieved equal performance to Kaggle winners with faster training and hardware compatibility.
CoCoPIE shows AI can run on regular devices without special hardware.
problem Real-time AI on mobile devices without special hardware.
method Compression-compiler co-design using CoCoPIE framework.
result Real-time AI applications on off-the-shelf mobile devices.
Slalom improves neural network performance in TEEs by delegating work to untrusted hardware.
problem Efficiently executing neural networks in secure hardware environments while maintaining performance and privacy.
method Slalom efficiently partitions DNN computations between trusted and untrusted devices, using matrix multiplication outsourcing.
result 6x to 20x increase in throughput for verifiable inference, and 4x to 11x for verifiable and private inference.
MLPerf benchmarks ML inference systems across diverse hardware.
problem Challenges in evaluating ML inference systems due to variety and complexity.
method MLPerf prescribes a set of rules and best practices for comparability.
result First call for submissions yielded over 600 reproducible measurements.
Explaining Quantum Circuits with Shapley Values: Towards Explainable Quantum Machine Learningquant-ph
Quantum circuits explained using Shapley values for better understanding.
problem Improving the explainability of quantum machine learning circuits.
method Applying Shapley values to quantify gate importance in quantum circuits.
result Quantum circuits can be explained by their gate importance, enhancing understanding and interpretability.
Benchopt automates machine learning benchmarking across languages and hardware.
problem Limited transparency and tedious re-implementation work in machine learning validation.
method A collaborative framework for automating, reproducing, and publishing optimization benchmarks.
result Demonstrates practical findings that highlight the importance of details in machine learning validation.
Quantum models learn unitary actions on entangled states from product states.
problem Generalization to out-of-distribution data in quantum machine learning.
method Proved out-of-distribution generalization for learning unitary actions.
result Learned unitary actions on entangled states from product states.