New methods speed up training of differentially private deep learning models.
problem Training differentially private deep learning models is slower than non-private models.
method Derive and implement new per-example gradient clipping methods compatible with auto-differentiation.
result Significant training speed-ups (54x - 94x) for various models and architectures.
Proposes a method to train neural networks that solve differential equations faster.
problem Training neural networks that solve differential equations becomes computationally expensive.
method Introduces a differentiable surrogate for numerical solver time cost using higher-order derivatives.
result Trains models that are faster to solve while maintaining nearly the same accuracy.
Paper compares semi-supervised training for differentiable particle filters.
problem Lack of labelled data in real-world applications.
method Compares two semi-supervised training objectives.
result Improved performance in environments with scarce labelled data.
Normalization layers improve the accuracy of Differentially Private training of deep neural networks.
problem Reduced accuracy in deep neural networks with Differentially Private training.
method Proposed a novel method for integrating batch normalization with Differentially Private Stochastic Gradient Descent (DPSGD) without additional privacy loss.
result Training deeper networks with better utility-privacy trade-off is possible.
ES for non-differentiable parameters scales to large models.
problem Learning non-differentiable parameters in large models.
method Hybrid approach combining ES for non-differentiable and gradient-based methods for differentiable parameters.
result Hybrid approach is competitive and allows training sparse models from the start.
Differentially private algorithms protect model explanations from leaking training data.
problem Model explanations can leak training data, compromising privacy.
method Adaptive differentially private gradient descent algorithm to produce accurate, private explanations.
result Privacy amplification and reduction of overall privacy loss on explanation data.
A new differential entropy estimator for neural networks training.
problem Lack of effective differential entropy estimators for neural network training.
method KNIFE: a fully parameterized, differentiable kernel-based estimator of differential entropy.
result KNIFE effectively estimates differential entropy and improves neural network training.
Differential privacy reduces model accuracy more for underrepresented groups.
problem Differential privacy impacts model accuracy differently across groups.
method Training neural networks with differential privacy (DP-SGD).
result DP-SGD reduces accuracy more for underrepresented groups.
Study evaluates federated learning with differential privacy on MIMIC-III, improving model performance with careful parameter tuning.
problem Training machine learning models on privacy-sensitive data sets locked in healthcare facilities.
method Extensive evaluation of federated and differential privacy techniques on MIMIC-III dataset, analyzing various parameters.
result Careful parameter tuning is crucial for federated learning with differential privacy, especially for data distribution and communication strategies.
DPDMs use DP-SGD to generate private synthetic data.
problem Generating private synthetic data from sensitive datasets.
method Introduced DP-SGD for DMs, investigated DM parameterization and sampling, proposed noise multiplicity.
result Achieved state-of-the-art performance in image generation benchmarks.
Neural networks trained with Q-learning converge to optimal control.
problem Optimal control of reinforcement learning with neural networks.
method Analysis of a limit differential equation derived from Q-learning.
result The limit differential equation has a unique stationary solution that is optimal.
Paper explores challenges in training PINNs and loss landscape effects.
problem Challenges in training Physics-Informed Neural Networks (PINNs) due to loss landscape issues.
method Examined gradient-based optimizers Adam, L-BFGS, and their combination Adam+L-BFGS, and introduced NysNewton-CG (NNCG).
result Adam+L-BFGS outperforms other optimizers, and NysNewton-CG significantly improves PINN performance.
DPlis improves privacy in deep learning models by smoothing loss functions.
problem Privacy leakage in deep learning models trained on private data and low model performance.
method DPlis constructs a smooth loss function to favor noise-resilient models.
result DPlis effectively boosts model quality and training stability under privacy constraints.
Novel deep learning approach for fast, differentiable fluid simulations.
problem Challenges in solving incompressible fluid dynamics equations efficiently.
method Physics-constrained training approach for convolutional neural networks.
result Trained models can handle various fluid phenomena and offer fast simulations.
Efficient echo state network with explicit memory performs well on benchmark tasks.
problem Training differentiable neural computers is difficult and time-consuming.
method Echo state network with an explicit memory.
result Echo state network can recognize all regular languages, including those contractive networks cannot.
Differentially private GANs improve image privacy without significant quality loss.
problem Anonymizing image data sets while maintaining image quality.
method Training GANs with differential privacy on MNIST, analyzing privacy-utility trade-offs and explaining optimization methods.
result An increasing privacy budget adds little to generated image quality, revealing a saturated training regime.
Differentiable losses for combinatorial optimization problems in sequence modeling.
problem Mismatch between training and inference objectives in sequence models.
method Gradient descent over linear programs representing combinatorial optimization problems.
result Gradient descent can be applied to combinatorial optimization problems efficiently.
New federated f-differential privacy for collaborative learning.
problem Privacy in federated learning.
method Introducing federated f-differential privacy and proposing a generic private federated learning framework. result Proves federated f-differential privacy provides privacy guarantee on each record of one client's data. Paper improves deep learning privacy with new f-differential privacy framework.
problem Training neural networks on sensitive data while maintaining privacy.
method Introduced and analyzed f-differential privacy for neural networks training. result Improved privacy guarantees for neural networks training without sacrificing accuracy.
Study improves sampling efficiency of diffusion models using RL and PDEs.
problem Training neural stochastic differential equations without access to target samples.
method Proves equivalences between RL methods and PDEs, uses coarse time discretization.
result Improves sample efficiency and reduces computational cost.
Differential ML combines AAD with ML for fast, accurate financial derivatives pricing and risk management.
problem Computational bottlenecks in financial derivatives risk management.
method Novel algorithms using automatic adjoint differentiation (AAD) for training fast, accurate approximations in real-time.
result Convergence guarantees for fast, accurate pricing and risk approximations for arbitrary derivatives instruments.
Introduces a differentiable approximation to the zero-one loss.
problem Incompatibility of zero-one loss with gradient-based optimization.
method Smooth projection onto hypersimplex through constrained optimization.
result Achieves significant improvements in generalization under large-batch training.
Efficiently computes per-example gradients in CNNs for differential privacy.
problem Computing per-example gradients in CNNs for differential privacy.
method Comparison of existing strategies and introduction of a new per-example gradient calculation method.
result The new method is advantageous depending on model architecture and training.
This paper analyzes how differential privacy and data skewness affect membership inference attacks.
problem Membership inference attacks on privately trained models.
method Developed MPLens system to evaluate membership inference vulnerability.
result Membership inference risk is higher with skewed training data and differential privacy has trade-offs.
Large data collections required for the training of neural networks often contain sensitive information such as the medical histories of patients, and the privacy of the training data must be preserved. In this paper, we introduce a dropout technique that provides an elegant Bayesian interpretation to dropout, and show…
Algorithm optimizes millions of hyperparameters efficiently.
problem Training modern network architectures with millions of hyperparameters.
method Combines implicit function theorem with efficient inverse Hessian approximations for gradient-based optimization.
result Jointly tuning weights and hyperparameters is only a few times more costly than standard training.
Attention augments forest for tabular data accuracy.
problem Training tabular data models with high accuracy and efficiency.
method Tree Attention Block (TAB) in differentiable forest framework.
result Attention augmented differentiable forest achieves comparable and sometimes higher accuracy than GBDT models.
Efficiently samples complex distributions using tensor train format.
problem Sampling from high-dimensional complex probability densities efficiently.
method Integrates tensor train format with backward stochastic differential equations (BSDEs) for fast, robust, and accurate sampling.
result Improved efficiency in sampling from challenging target distributions.
New training scheme reduces adversarial examples by increasing decision boundary margin.
problem Vulnerability of neural networks to adversarial examples.
method Differential training using a loss function on feature differences.
result Differential training significantly reduces adversarial examples.
Paper trains language models without memorizing user data.
problem Training production-scale language models with user privacy.
method Federated Learning with Differentially Private Federated Averaging (DP-FedAvg).
result Demonstrated training of a production NWP model with user-level differential privacy.
The paper introduces a new ODE approach to improve Wasserstein GANs.
problem Improving Wasserstein GANs for better training results.
method Derives an ODE representing the gradient flow of Wasserstein-1 loss and proposes a new model W1-FE.
result W1-FE outperforms WGAN in training experiments across various dimensions.
This paper improves deep learning models' accuracy with differential privacy using gradient encoding and denoising.
problem Deep learning models leak sensitive information about their training datasets.
method Gradient encoding to map gradients to a smaller vector space, and denoising for post-processing.
result Our technique achieves better model accuracy with differential privacy guarantees compared to state-of-the-art methods.
DP-CGAN generates private synthetic data and labels.
problem Preserving privacy in synthetic data generation.
method Differentially private conditional GAN (DP-CGAN) with clipping and perturbation.
result DP-CGAN generates visually and empirically promising results on MNIST with low privacy cost.
New method improves training stochastic neural networks with tighter guarantees.
problem Training stochastic neural networks with provable guarantees.
method Developed partially-aggregated estimators and reformulated PAC-Bayesian bounds.
result Derives a differentiable objective leading to tighter generalisation guarantees.
Improves discrete latent representations using differentiable approximation bridges.
problem Improving discrete latent representations in neural networks.
method Training with a differentiable approximation bridge (DAB) neural network.
result Improves state-of-the-art performance in various domains.
Paper improves differential privacy SGD by considering data heterogeneity.
problem Improving differential privacy in machine learning with varying data contributions.
method Introducing influence function to quantify data contributions, designing PIDP-SGD algorithm.
result PIDP-SGD significantly improves machine learning model performance.
The paper discusses how to improve machine learning models using partial differential equations.
problem Improving the performance and generalization of machine learning models.
method The paper reframes implicit regularization techniques in deep learning as explicit gradient regularization using partial differential equations.
result Explicit regularization using PDEs can lead to better model performance and generalization.
Paper introduces a differentially private generative model using gradient flow and sliced Wasserstein distance.
problem Protecting privacy in sensitive training data for generative models.
method Gradient flow in the space of probability measures, Gaussian-smoothed Sliced Wasserstein Distance, and numerical scheme for SDE.
result Demonstrates higher-fidelity data generation at low privacy budget compared to existing methods.
New system defends deep neural networks like ResNet-34.
problem Training deep neural networks like ResNet-34 is challenging.
method Differentiable abstract interpretation and a DSL for training objectives.
result Can defend significantly larger networks than before.
Differentiable model compression adds noise to parameters during training.
problem Model compression for efficient deployment.
method Adding pseudo quantization noise to model parameters during training.
result Compresses model by more than 8x on ImageNet with 0.3% accuracy loss.
Introduces Relational Privacy (RP) to control relation memorization in question answering models.
problem Relation memorization in question answering models can lead to privacy issues.
method Formalizes Relational Privacy (RP) and Differential Relational Privacy (DrP), providing bounds on relation memorization.
result DrP allows effective learning of general properties of underlying concepts while preventing relation memorization.
Improved privacy bounds enhance deep learning training efficiency.
problem Enhancing privacy guarantees in deep learning models.
method Deriving optimal DP parameters using f-divergences. result Significantly reduces the number of iterations needed for training deep learning models.
New method improves privacy of machine learning models using shuffling.
problem Maintaining privacy in machine learning models trained on large datasets.
method Recovering privacy guarantees for cyclic Stochastic Gradient Descent (SGD), a faster alternative to traditional SGD.
result Privacy guarantees for cyclic SGD are competitive with those for sampling with replacement.
Effective Gram matrix predicts deep network generalization.
problem Understanding and predicting deep network generalization.
method Derived a differential equation governing generalization gap, analyzed with effective Gram matrix.
result Effective Gram matrix accurately predicts test loss during training.
Privacy-preserving machine learning methods add randomness, leading to varying predictions.
problem Privacy-preserving machine learning methods add randomness, leading to varying predictions.
method The study analyzes three DP-ensuring algorithms: output perturbation, objective perturbation, and DP-SGD.
result The degree of predictive multiplicity rises as the level of privacy increases, and is unevenly distributed across individuals and demographic groups.
Differentially private learning on real-world data poses challenges for standard machine learning practice: privacy guarantees are difficult to interpret, hyperparameter tuning on private data reduces the privacy budget, and ad-hoc privacy attacks are often required to test model privacy. We introduce three tools to ma…
This work integrates differentiation and integration in Physics-Informed Neural Networks.
problem Solving integro-differential equations and computing integral transforms.
method Augmenting Physics-Informed Neural Networks with automatic integration.
result Solving complex integral transforms and integro-differential equations.
ICON learns differential equation operators from prompts, reducing retraining and improving few-shot learning.
problem Training neural networks to solve differential equations without retraining for new problems.
method In-Context Operator Networks (ICON) that learns operators from prompted data and applies them to new problems.
result ICON can generalize to new operators beyond the training distribution and requires only a few demos.