Deep VIB is a neural-network parameterization of the information bottleneck objective trained via variational inference and the reparameterization trick, yielding improved generalization and adversarial robustness.
hub Mixed citations
TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
Mixed citation behavior. Most common role is method (64%).
53
Pith papers citing it
Method
64% of classified citations
abstract
TensorFlow is an interface for expressing machine learning algorithms, and an implementation for executing such algorithms. A computation expressed using TensorFlow can be executed with little or no change on a wide variety of heterogeneous systems, ranging from mobile devices such as phones and tablets up to large-scale distributed systems of hundreds of machines and thousands of computational devices such as GPU cards. The system is flexible and can be used to express a wide variety of algorithms, including training and inference algorithms for deep neural network models, and it has been used for conducting research and for deploying machine learning systems into production across more than a dozen areas of computer science and other fields, including speech recognition, computer vision, robotics, information retrieval, natural language processing, geographic information extraction, and computational drug discovery. This paper describes the TensorFlow interface and an implementation of that interface that we have built at Google. The TensorFlow API and a reference implementation were released as an open-source package under the Apache 2.0 license in November, 2015 and are available at www.tensorflow.org.
hub tools
citation-role summary
method 7
background 3
other 1
citation-polarity summary
representative citing papers
Floating-point neural networks with automatic differentiation can represent arbitrary floating-point functions and their gradients under mild conditions.
A noisy top-k gated mixture-of-experts layer between LSTMs scales neural networks to 137B parameters with sub-linear compute, beating SOTA on language modeling and machine translation.
Real NVP uses affine coupling layers to create invertible transformations that support exact density estimation, sampling, and latent inference without approximations.
RLGT is a modular reinforcement learning framework for extremal graph theory that handles undirected, directed, looped, and multi-colored graphs to facilitate future research.
IRNet uses per-layer residual shortcuts in fully connected networks to achieve better prediction accuracy and training convergence than prior ML methods on OQMD and Materials Project datasets for material properties.
Reinforcement learning agents solve complex tasks without access to the reward function by training a reward predictor from human comparisons of trajectory segments, requiring feedback on less than 1% of interactions.
SMART transfers knowledge in multi-task linear regression via spectral subspace similarity assumptions, achieving near-minimax Frobenius error rates while requiring only a fitted source model.
Kinetics is a new video dataset of 400 human actions with over 160000 ten-second clips collected from YouTube, accompanied by baseline action-classification results from neural networks.
Hypernetworks generate weights for a main network, allowing LSTMs to use non-shared weights and achieve near state-of-the-art results on sequence modeling tasks while using fewer parameters overall.
Fractional OAM charge ℓ=1.5 induces an optimal 67.5° GKP lattice rotation that reduces error rate 23.9× with <0.2% loss in Fisher information and yields 41% higher metrological capacity.
A parametric autoencoder with non-negativity and softmax constraints learns interpretable latent chemical components and couples them to kinetics and heat release for improved reduced-order modeling of decomposition.
Dorito enables diffraction-limited image reconstruction from JWST AMI observations by deconvolving images or Fourier observables using maximum entropy and total variation regularization.
ShinkaEvolve improves sample efficiency in LLM-driven program evolution via parent sampling, code novelty rejection-sampling, and bandit LLM ensemble selection, achieving new SOTA circle packing with 150 samples and gains on math reasoning and competitive programming tasks.
A conditional GAN reconstructs shape, size, and brightness distributions of simulated fast-rotating stars from intensity interferometry power spectra obtained with 6- and 9-telescope arrays.
Electroweak diboson plus high-mass dijet production observed at 7.4 sigma with signal strength 1.28, plus first semileptonic-channel limits on S02, T0 and M0 Wilson coefficients.
CSEN is a compact convolutional neural network trained to estimate sparse support sets directly from measurements, claiming state-of-the-art accuracy at lower computational cost than iterative methods.
A multi-delay sinc network jointly aligns speech signals with delayed continuous emotion labels and predicts arousal/valence, claiming state-of-the-art speech-only results on RECOLA and SEWA.
An iterative audio-visual approach for speaker diarisation in real-world meetings that enrolls speaker models via correspondence and outperforms prior methods on the AMI corpus.
An FPGA-based neural-network decoder achieves 550 ns deterministic closed-loop latency for real-time distance-3 surface code error correction on a superconducting processor, matching offline decoding performance.
An amortized variational framework jointly targets the posterior and posterior-predictive distributions via a KL upper bound and moment regularization, yielding more accurate predictions at lower online cost than two-stage variational inference.
Entropic optimal transport yields a clustering loss with the same global optimum as log-likelihood but a better-behaved optimization surface, outperforming standard EM in experiments.
Alikhanov-XfPINNs integrates accelerated Alikhanov discretization on nonuniform time grids with physics-informed neural networks to solve general nonlinear fractional PDEs for both forward and inverse problems with improved efficiency and handling of initial singularities.
TCL delivers 16.8x faster tuning on CPU and 12.48x on GPU with modestly lower inference latency by combining RDU active sampling, a lightweight Mamba cost model, and cross-platform continual knowledge distillation.
citing papers explorer
-
Deep Variational Information Bottleneck
Deep VIB is a neural-network parameterization of the information bottleneck objective trained via variational inference and the reparameterization trick, yielding improved generalization and adversarial robustness.
-
Floating-Point Networks with Automatic Differentiation Can Represent Almost All Floating-Point Functions and Their Gradients
Floating-point neural networks with automatic differentiation can represent arbitrary floating-point functions and their gradients under mild conditions.
-
Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
A noisy top-k gated mixture-of-experts layer between LSTMs scales neural networks to 137B parameters with sub-linear compute, beating SOTA on language modeling and machine translation.
-
Density estimation using Real NVP
Real NVP uses affine coupling layers to create invertible transformations that support exact density estimation, sampling, and latent inference without approximations.
-
RLGT: A reinforcement learning framework for extremal graph theory
RLGT is a modular reinforcement learning framework for extremal graph theory that handles undirected, directed, looped, and multi-colored graphs to facilitate future research.
-
IRNet: A General Purpose Deep Residual Regression Framework for Materials Discovery
IRNet uses per-layer residual shortcuts in fully connected networks to achieve better prediction accuracy and training convergence than prior ML methods on OQMD and Materials Project datasets for material properties.
-
Deep reinforcement learning from human preferences
Reinforcement learning agents solve complex tasks without access to the reward function by training a reward predictor from human comparisons of trajectory segments, requiring feedback on less than 1% of interactions.
-
SMART: A Spectral Transfer Approach to Multi-Task Learning
SMART transfers knowledge in multi-task linear regression via spectral subspace similarity assumptions, achieving near-minimax Frobenius error rates while requiring only a fitted source model.
-
The Kinetics Human Action Video Dataset
Kinetics is a new video dataset of 400 human actions with over 160000 ten-second clips collected from YouTube, accompanied by baseline action-classification results from neural networks.
-
HyperNetworks
Hypernetworks generate weights for a main network, allowing LSTMs to use non-shared weights and achieve near state-of-the-art results on sequence modeling tasks while using fewer parameters overall.
-
OAM-Induced Lattice Rotation Reveals a Fractional Optimum in Fault-Tolerant GKP Quantum Sensing
Fractional OAM charge ℓ=1.5 induces an optimal 67.5° GKP lattice rotation that reduces error rate 23.9× with <0.2% loss in Fisher information and yields 41% higher metrological capacity.
-
A Data-Driven Parametric Reduced-Order Chemical Kinetics Model Derived from Atomistic Simulations
A parametric autoencoder with non-negativity and softmax constraints learns interpretable latent chemical components and couples them to kinetics and heat release for improved reduced-order modeling of decomposition.
-
Image reconstruction with the JWST Interferometer
Dorito enables diffraction-limited image reconstruction from JWST AMI observations by deconvolving images or Fourier observables using maximum entropy and total variation regularization.
-
ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution
ShinkaEvolve improves sample efficiency in LLM-driven program evolution via parent sampling, code novelty rejection-sampling, and bandit LLM ensemble selection, achieving new SOTA circle packing with 150 samples and gains on math reasoning and competitive programming tasks.
-
Generative AI for image reconstruction in Intensity Interferometry: a first attempt
A conditional GAN reconstructs shape, size, and brightness distributions of simulated fast-rotating stars from intensity interferometry power spectra obtained with 6- and 9-telescope arrays.
-
Electroweak diboson production in association with a high-mass dijet system in semileptonic final states from $pp$ collisions at $\sqrt{s} = 13$ TeV with the ATLAS detector
Electroweak diboson plus high-mass dijet production observed at 7.4 sigma with signal strength 1.28, plus first semileptonic-channel limits on S02, T0 and M0 Wilson coefficients.
-
Convolutional Sparse Support Estimator Network (CSEN) From energy efficient support estimation to learning-aided Compressive Sensing
CSEN is a compact convolutional neural network trained to estimate sparse support sets directly from measurements, claiming state-of-the-art accuracy at lower computational cost than iterative methods.
-
Jointly Aligning and Predicting Continuous Emotion Annotations
A multi-delay sinc network jointly aligns speech signals with delayed continuous emotion labels and predicts arousal/valence, claiming state-of-the-art speech-only results on RECOLA and SEWA.
-
Who said that?: Audio-visual speaker diarisation of real-world meetings
An iterative audio-visual approach for speaker diarisation in real-world meetings that enrolls speaker models via correspondence and outperforms prior methods on the AMI corpus.
-
Real-time Surface-Code Error Correction Using an FPGA-based Neural-Network Decoder
An FPGA-based neural-network decoder achieves 550 ns deterministic closed-loop latency for real-time distance-3 surface code error correction on a superconducting processor, matching offline decoding performance.
-
Amortized Variational Inference for Joint Posterior and Predictive Distributions in Bayesian Uncertainty Quantification
An amortized variational framework jointly targets the posterior and posterior-predictive distributions via a KL upper bound and moment regularization, yielding more accurate predictions at lower online cost than two-stage variational inference.
-
On Model-Based Clustering With Entropic Optimal Transport
Entropic optimal transport yields a clustering loss with the same global optimum as log-likelihood but a better-behaved optimization surface, outperforming standard EM in experiments.
-
Alikhanov-XfPINNs: Adaptive Physics-Informed Learning for Nonlinear Fractional PDEs on Nonuniform Meshes
Alikhanov-XfPINNs integrates accelerated Alikhanov discretization on nonuniform time grids with physics-informed neural networks to solve general nonlinear fractional PDEs for both forward and inverse problems with improved efficiency and handling of initial singularities.
-
TCL: Enabling Fast and Efficient Cross-Hardware Tensor Program Optimization via Continual Learning
TCL delivers 16.8x faster tuning on CPU and 12.48x on GPU with modestly lower inference latency by combining RDU active sampling, a lightweight Mamba cost model, and cross-platform continual knowledge distillation.
-
MONAI: An open-source framework for deep learning in healthcare
MONAI is a community-supported PyTorch framework that extends deep learning to medical data with domain-specific architectures, transforms, and deployment tools.
-
Rethinking Atrous Convolution for Semantic Image Segmentation
DeepLabv3 improves semantic segmentation by capturing multi-scale context with cascaded or parallel atrous convolutions and adding global context to ASPP, achieving better results on PASCAL VOC 2012 without DenseCRF post-processing.
-
Free surfaces in turbulence -- A unified framework from water surfaces to elastic solids
Linear theory predicts regimes for deformable surfaces in turbulence where the interface is enslaved by flow or shows intrinsic dynamics; simulations of air-water and rubber match predictions without wave turbulence.
-
Rethinking Expert Trajectory Utilization in LLM Post-training for Mathematical Reasoning
Sequential SFT followed by RL, guided by the Plasticity-Ceiling Framework, achieves higher performance ceilings in LLM mathematical reasoning than synchronized methods by optimizing data scale and transition timing.
-
Neural network-based deconvolution for GeV-Scale Gamma-Ray Spectroscopy
A denoising autoencoder followed by a U-Net reconstructs incident gamma spectra from measured positron spectra in a Monte Carlo-optimized spectrometer for multi-MeV to GeV energies.
-
High-precision measurement of the W boson mass with the CMS experiment
CMS measures the W boson mass as 80360.2 ± 9.9 MeV from 2016 data, consistent with the Standard Model prediction.
-
Usenix'23 Extended Version: Smart Learning to Find Dumb Contracts
DLVA trains neural networks on bytecode to match Slither source labels at 92.7% accuracy and 0.2 seconds per contract while outperforming nine other tools at 99.7% average accuracy.
-
Adversarial optimization for joint registration and segmentation in prostate CT radiotherapy
An end-to-end 3D adversarial network estimates deformation vector fields to align CT images and propagate segmentations, showing improved performance and speed over elastix for prostate radiotherapy.
-
Lit2Vec: A Reproducible Workflow for Building a Legally Screened Chemistry Corpus from S2ORC for Downstream Retrieval and Text Mining
Lit2Vec delivers a documented, reproducible pipeline that extracts and annotates a large licensed chemistry paper corpus from S2ORC with paragraph embeddings and subfield labels.
-
SecureAFL: Secure Asynchronous Federated Learning
SecureAFL secures asynchronous federated learning against poisoning attacks by detecting anomalous updates, estimating missing client contributions, and using Byzantine-robust aggregation.
-
Harnessing AI for Inverse Partial Differential Equation Problems: Past, Present, and Prospects
A survey organizing AI methods for inverse PDE problems into inverse problems, inverse design, and control categories, covering applications and future challenges like physics-informed models and uncertainty quantification.
-
Automated Big Data Quality Assessment using Knowledge Graph Embeddings
Knowledge graph embeddings predict missing connections to generate context-specific data quality assessment plans, tested on a radiation sensor dataset.
-
Advance Warning Methodologies for COVID-19 using Chest X-Ray Images
Introduces the Early-QaTa-COV19 dataset and reports that CSEN reaches over 97% sensitivity and over 95.5% specificity for early COVID-19 detection from X-rays.
-
Convolutional Sparse Support Estimator Based Covid-19 Recognition from X-ray Images
Introduces CSEN, a non-iterative network bridging sparse representation and deep learning, for Covid-19 detection from X-ray images with limited training data.
-
Automatically Learning Construction Injury Precursors from Text
Standard NLP classifiers can surface valid injury precursors from raw construction safety reports.
-
Deformable Registration Using Average Geometric Transformations for Brain MR Images
The method augments VoxelMorph with Jacobian and curl channels plus an average-transformation atlas and reports higher Dice scores and more valid Jacobians than the original VoxelMorph on ADNI and MRBrainS18 data.
-
An Actor-Critic-Attention Mechanism for Deep Reinforcement Learning in Multi-view Environments
An attention-augmented actor-critic agent learns to dynamically weight multiple environment views by importance and outperforms baselines on TORCS and three other 3D simulators under noise and partial observability.
-
Improving Semantic Segmentation via Dilated Affinity
Dilated affinity is jointly predicted with segmentation labels to strengthen features and support efficient label propagation refinement on benchmark datasets.
-
Autoencoding sensory substitution
Deep recurrent autoencoders convert images to shortened audio signals that incorporate hearing models, enabling above-chance hand posture discrimination and object reaching after a few hours of training instead of months.
-
Knowledge-incorporating ESIM models for Response Selection in Retrieval-based Dialog Systems
K-ESIM and T-ESIM extend ESIM by incorporating domain knowledge and similar-dialog information, yielding preliminary accuracy gains on Ubuntu and Advising datasets for next-utterance selection.
-
ELKPPNet: An Edge-aware Neural Network with Large Kernel Pyramid Pooling for Learning Discriminative Features in Semantic Segmentation
ELKPPNet combines a balanced encoder-decoder, large kernel spatial pyramid pooling for multi-scale fusion, and an edge-aware loss to claim superior semantic segmentation performance on Cityscapes, CamVid, and NYUDv2 versus prior methods.
-
Brain MR Image Segmentation in Small Dataset with Adversarial Defense and Task Reorganization
The method reaches 84.46% Dice score on brain MR segmentation of gray matter, white matter and major regions using only seven training subjects via adversarial defense and hierarchical task reorganization.
-
A numerical study into neural network surrogate model performance for uncertainty propagation
Numerical study comparing feedforward NN and DeepONet with data-driven and physics-informed losses on stochastic heat equation, highlighting larger errors at distribution tails due to extrapolation.
-
A Pedagogical Framework for Physics-Informed Machine Learning: From Classical Pendulum to Quantum Anharmonic Oscillator Using PyTorch on Modern GPU Hardware
A pedagogical framework implements and benchmarks ANN, CNN, LSTM, and PINN models for classical pendulum and quantum oscillator systems with reported MAEs and GPU speedups.
-
Array Programming with NumPy
NumPy provides array programming tools that form the foundation of the scientific Python ecosystem and enable data analysis across many disciplines.
-
Short-term Electric Load Forecasting Using TensorFlow and Deep Auto-Encoders
A TensorFlow-based deep auto-encoder model is proposed for short-term electric load forecasting and claimed to outperform traditional neural networks in accuracy and stability.