Plug-and-play L-GM-AMP improves CS recovery for any i.i.d. source prior.
problem Efficiently recovering signals from compressed measurements with unknown priors.
method Deep learning with Gaussian-mixture model to approximate source prior, combined with learned denoising.
result L-GM-AMP achieves state-of-the-art performance without prior knowledge of source distribution.
Plug-and-play multimodal controller improves class-conditional image generation.
problem Generating class-conditional images from user-specified labels.
method Introduces a `multimodal controller` to generate multimodal data without additional learning parameters.
result Multimodal controlled generative models produce higher quality class-conditional images and novel modalities.
New convergence rates found for PnP methods using MMSE denoisers.
problem Asymptotic convergence of PnP methods with MMSE denoisers.
method Explicitly represented MMSE denoiser as an upper Moreau envelope, derived sublinear convergence rates.
result First sublinear convergence guarantee for PnP proximal gradient descent with MMSE denoiser.
New method uses CNN for seismic inversion uncertainty quantification.
problem Uncertainty quantification in seismic inversion for noisy data.
method Plug-and-Play Stein Variational Gradient Descent (PnP-SVGD) with CNN denoiser.
result High-resolution, trustworthy posterior samples for subsurface structures.
New framework improves experimental design using integral probability metrics.
problem Challenges in Bayesian Optimal Experimental Design (BOED) with KL divergence.
method Integrates integral probability metrics (IPMs) for stability and flexibility.
result IPM-based designs yield more robust and accurate credible sets.
RNE provides a flexible framework for diffusion models, enabling inference-time control and energy-based training.
problem Insufficient knowledge of marginal densities in diffusion models.
method Introduces Radon-Nikodym Estimator (RNE) to reveal the connection between marginal densities and transition kernels.
result RNE delivers strong results in inference-time control and energy-based diffusion training.
Develops a robust method for image reconstruction from limited data.
problem Inference of unknown images from few measurements, often ill-posed.
method Introduces DPnP, a diffusion plug-and-play method combining likelihood and score-based samplers.
result Establishes performance guarantees for DPnP, demonstrating robustness and efficiency.
Paper tackles Bayesian image restoration in low-photon Poisson imaging problems.
problem Bayesian inference in challenging low-photon Poisson imaging problems.
method Plug-and-play (PnP) Langevin sampling strategies with accelerated methods and mirror sampling.
result Effective PnP Langevin sampling methods for low-photon Poisson imaging problems.
PnP-CM integrates CMs into PnP frameworks for efficient inverse problem solving.
problem Efficiently solving inverse problems with high-quality reconstructions.
method Reinterpreting CMs as proximal operators and integrating them into PnP frameworks.
result PnP-CM achieves high-quality reconstructions in as few as 4 NFEs.
Unified framework denoises data and abstains from uncertain predictions.
problem Data quality and predictive uncertainty in deep neural networks.
method Unified filtering framework leveraging data density.
result Framework outperforms state-of-the-art techniques in denoising and abstaining.
Scalable PnP-ADMM for large-scale imaging problems.
problem Heavy computational and memory requirements of current PnP algorithms.
method Incremental variant of PnP-ADMM with theoretical convergence guarantees.
result Fast convergence and scalability compared to existing PnP algorithms.
Bayesian approach improves rain field reconstruction using CMLs and DMs.
problem Challenges in accurately reconstructing ground-level rainfall from CML path-integrated measurements.
method Bayesian inverse problem with Diffusion Models as priors.
result Improved performance in rainfall estimation compared to existing methods.
New method addresses error bounds for PnP-ULA under mismatched models.
problem Error bounds for PnP-ULA under mismatched measurement and prior models.
method Posterior-L2 pseudometric to quantify error bounds.
result Explicit error bound for PnP-ULA under mismatched posterior distribution.
A new framework for controllable generation of discrete masked models.
problem Efficient controllable generation of discrete data models.
method Plug-and-play framework based on importance sampling.
result Demonstrates versatility across multiple domains, including protein design.
New PnP algorithm converges with relaxed proximal gradient descent.
problem Convergence issues in PnP methods with deep denoisers.
method Relaxed proximal gradient descent for PnP with weakly convex regularization.
result Proposed PnP-αPGD converges for a wider range of regularization parameters. New algorithm improves signal recovery from noisy measurements with theoretical guarantees.
problem Recovering signals from noisy measurements in inverse problems.
method Wasserstein-based projections (WP) replacing analytic regularization with data-driven denoising.
result WP approximates true projection with high probability, providing theoretical guarantees.
SNORE applies denoiser only on images with noise of adequate level for image restoration.
problem Image restoration challenges with iterative algorithms and denoising.
method SNORE framework using stochastic regularization and stochastic gradient descent.
result SNORE is competitive with state-of-the-art methods on deblurring and inpainting tasks.
Novel algorithm accelerates PnP methods for image deblurring and super-resolution.
problem Efficiently solving inverse problems and imaging with provable convergence guarantees.
method Incorporates quasi-Newton steps into provable PnP framework based on proximal denoisers.
result 2--8x faster convergence compared to other provable PnP methods with similar quality.
A new method uses DMs as priors for imaging problems, offering more accurate reconstructions.
problem Accurate probabilistic imaging for complex inverse problems.
method Markov chain Monte Carlo algorithm using DMs as plug-and-play priors for solving Bayesian inverse problems.
result Offers more accurate reconstructions and posterior estimation compared to existing methods.
Paper tackles fairness issues in GNNs by proposing ELEGANT for certification.
problem Fairness issues in GNN predictions due to graph data perturbations.
method Proposes ELEGANT framework for certifying fairness of any GNN without assumptions or re-training.
result The fairness of any GNN backbone is impossible to be corrupted under certain perturbation budgets.
New method uses random decompositions for high-dimensional Bayesian optimization.
problem Learning accurate decompositions for high-dimensional black-box functions.
method Data-independent random tree-based decomposition sampling.
result Random decomposition upper-confidence bound algorithm (RDUCB) yields significant empirical gains.
Paper introduces a new sampling method for Bayesian inference.
problem Efficiently sampling from complex posterior distributions.
method Plug-and-Play split Gibbs sampler using variable splitting and ADMM.
result The method allows for embedding deep generative priors in Bayesian inference.
TAMO optimizes multiple objectives in-context using transformers.
problem Balancing competing objectives in expensive, black-box problems.
method Fully amortized, transformer-based policy for multi-objective optimization.
result Significant speedup in proposal time with improved Pareto quality.
A new method uses compressive autoencoders for image restoration.
problem Efficient regularization of inverse problems in computational imaging.
method Variational Bayes Latent Estimation (VBLE) with compressive autoencoders.
result VBLE achieves similar performance to state-of-the-art PnP methods but faster.
New method improves clustering accuracy in noisy single-cell data.
problem Challenges in clustering single-cell RNA sequencing data due to noise and variability.
method Latent plug-and-play diffusion framework with input-space steering.
result Improved clustering accuracy on synthetic and real-world single-cell data.
Deep neural networks trained using a softmax layer at the top and the cross-entropy loss are ubiquitous tools for image classification. Yet, this does not naturally enforce intra-class similarity nor inter-class margin of the learned deep representations. To simultaneously achieve these two goals, different solutions h…
We introduce deep scale-spaces (DSS), a generalization of convolutional neural networks, exploiting the scale symmetry structure of conventional image recognition tasks. Put plainly, the class of an image is invariant to the scale at which it is viewed. We construct scale equivariant cross-correlations based on a princ…
Bayesian methods promise to fix many shortcomings of deep learning, but they are impractical and rarely match the performance of standard methods, let alone improve them. In this paper, we demonstrate practical training of deep networks with natural-gradient variational inference. By applying techniques such as batch n…
Bayesian approach improves neural network classification accuracy and uncertainty.
problem Overconfidence and lack of uncertainty in softmax for classification tasks.
method Model categorical probability using a random variable with a prior distribution.
result Consistent gains in generalization performance across multiple tasks.
Inspired by the immense success of deep learning, graph neural networks (GNNs) are widely used to learn powerful node representations and have demonstrated promising performance on different graph learning tasks. However, most real-world graphs often come with high-dimensional and sparse node features, rendering the le…
Actor-critic methods can achieve incredible performance on difficult reinforcement learning problems, but they are also prone to instability. This is partly due to the interaction between the actor and critic during learning, e.g., an inaccurate step taken by one of them might adversely affect the other and destabilize…
Unified framework for image restoration using equivariant denoisers.
problem Restoring images with realistic priors and invariant transformations.
method Unified framework named ERED based on equivariant denoisers and stochastic optimization.
result Unified framework ERED converges and improves image restoration.
Unified algorithm for tensor decomposition supports multiple loss functions and models.
problem Efficient tensor decomposition for various models and loss functions.
method Hierarchical combination of ADMM and MM for optimization.
result Wide-range applications can be solved by the proposed algorithm.
Trieste optimizes black-box functions using TensorFlow for efficient decision-making.
problem Optimizing complex, black-box functions efficiently.
method Bayesian optimization and active learning with TensorFlow integration.
result Trieste enables efficient exploration of black-box functions.
Recently the field of inverse problems has seen a growing usage of mathematically only partially understood learned and non-learned priors. Based on first principles, we develop a projectional approach to inverse problems that addresses the incorporation of these priors, while still guaranteeing data consistency. We im…
This paper seeks to answer the question: as the (near-) orthogonality of weights is found to be a favorable property for training deep convolutional neural networks, how can we enforce it in more effective and easy-to-use ways? We develop novel orthogonality regularizations on training deep CNNs, utilizing various adva…
Generative models produce realistic objects in many domains, including text, image, video, and audio synthesis. Most popular models---Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs)---usually employ a standard Gaussian distribution as a prior. Previous works show that the richer family of pri…
CGD improves diffusion models' out-of-distribution generalization.
problem Reliable sampling from high-value regions beyond training data.
method Context-guided diffusion (CGD) using unlabeled data and smoothness constraints.
result Substantial performance gains across various diffusion processes.
New method optimizes sensor placement for stochastic systems efficiently.
problem Optimizing sensor placements for black-box stochastic systems with computational constraints.
method Trains a joint energy-based model on simulation data to learn parameter and solution distributions, allowing efficient sensor placement.
result Demonstrates lower computational cost and more informative sensor locations compared to conventional approaches.
Proposes a robust IV estimator using optimal transport for corrupted or adversarial data.
problem Lack of robustness in traditional IV estimators for corrupted or adversarial data.
method Integrates data-derivative information through optimal transport to address geometric aspects of data.
result Improves robustness against data corruption and adversarial attacks.
A lightweight framework improves convergence and stability of PINNs for complex PDEs.
problem Training instability and reduced accuracy in PINNs for complex PDEs.
method Adaptive curvature correction using secant information to optimize first-order optimizers.
result Consistent improvements in convergence speed, stability, and accuracy over standard optimizers.
Several tasks in machine learning are evaluated using non-differentiable metrics such as mean average precision or Spearman correlation. However, their non-differentiability prevents from using them as objective functions in a learning framework. Surrogate and relaxation methods exist but tend to be specific to a given…
AugMask trains diffusion models on incomplete tabular data by augmenting missing values and applying denoising supervision.
problem Training diffusion models on incomplete tabular data with missing values.
method AugMask uses stochastic augmentation and denoising supervision to adapt diffusion models to incomplete data.
result AugMask enables diffusion-based tabular generators to outperform specialized missing-aware baselines across various datasets and missingness regimes.
A new algorithm for compressing latent representations in deep models.
problem Compressing continuous latent representations in deep models.
method Separates model design and training from quantization; uses adaptive quantization based on posterior uncertainty.
result Image compression with the proposed algorithm outperforms JPEG over a wide range of bit rates.
Value-based reinforcement learning (RL) methods like Q-learning have shown success in a variety of domains. One challenge in applying Q-learning to continuous-action RL problems, however, is the continuous action maximization (max-Q) required for optimal Bellman backup. In this work, we develop CAQL, a (class of) algor…
Improved sampling efficiency for inverse problems using variance-reduced diffusion methods.
problem Efficiently estimating noisy scores in inverse problems.
method Developed a nonparametric self-normalized importance sampling estimator and a state-dependent blending rule.
result Improved sample quality for fixed simulation budgets in synthetic targets and PDE-governed inverse problems.
Paper tackles multi-object reinforcement learning, improving skill extrapolation.
problem Learning robust manipulation tasks in multi-object settings.
method Introduces a linear relation network module to enhance skill generalization.
result Agents can extrapolate and generalize to any new object number, scaling linearly.
Robots rely on sensors to provide them with information about their surroundings. However, high-quality sensors can be extremely expensive and cost-prohibitive. Thus many robotic systems must make due with lower-quality sensors. Here we demonstrate via a case study how modeling a sensor can improve its efficacy when em…