NRWS improves training of SBNs and HMs using natural gradient.
problem Training Sigmoid Belief Networks and Helmholtz Machines efficiently.
method Exploits block-diagonal structure of Fisher Information Matrices to use natural gradient.
result NRWS and NBiHM achieve better log-likelihood and faster convergence.
Stochastic control-flow models (SCFMs) are a class of generative models that involve branching on choices from discrete random variables. Amortized gradient-based learning of SCFMs is challenging as most approaches targeting discrete variables rely on their continuous relaxations---which can be intractable in SCFMs, as…
A new method for training generative models with sparse supervision.
problem Training deep generative models with sparse and varying supervision.
method Caffeinated Wake-Sleep (CWS) method, combining reweighted wake-sleep and teacher-forcing.
result The CWS method is robust to variable length supervision and performs well on various datasets.
Massively parallel RWS improves inference in complex models.
problem Exponential sample requirement for effective importance weighting.
method Draws K samples of all n latent variables and individually reasons through all combinations. result Significant improvements over standard RWS.
QEM uses parallel importance weighting for fast approximate Bayesian inference.
problem Bayesian inference challenges in large models with many observations and latent variables.
method Expectation Maximization (EM) with massively parallel importance weighting.
result QEM is faster and more scalable than RWS and VI.
NAS-X improves inference and model learning for SLVMs.
problem Challenges in analytic inference and model learning for flexible SLVMs.
method NAS-X combines reweighted wake-sleep and smoothing sequential Monte Carlo.
result NAS-X provides low-bias and low-variance gradient estimates.
New method improves variational inference for better posterior approximation.
problem Challenges in minimizing inclusive KL divergence for amortized variational inference.
method Likelihood-tempered sequential Monte Carlo samplers to estimate inclusive KL gradient.
result SMC-Wake method fits variational distributions more accurately than existing methods.
Paper improves variance control in importance weighted variational bounds.
problem Improving the variance of gradient estimators for IWAE.
method Develops a novel control variate that grows SNR as √K for large K.
result Empirically, the method yields superior variance reduction for generative models.
A new method learns latent variable updates directly, not approximating the posterior.
problem Intractable maximum-likelihood learning for complex latent-variable models.
method Amortised learning using wake-sleep Monte-Carlo strategy.
result Demonstrated effectiveness on various complex models.
New algorithm minimizes inclusive KL for VI, improving accuracy.
problem Improving variational inference accuracy with KL(p||q).
method Markovian score climbing (MSC) using stochastic gradients.
result MSC converges to local optimum of inclusive KL without bias.
The importance weighted autoencoder (IWAE) (Burda et al., 2016) is a popular variational-inference method which achieves a tighter evidence bound (and hence a lower bias) than standard variational autoencoders by optimising a multi-sample objective, i.e. an objective that is expressible as an integral over K>1 Mont…
Deep latent variable models have become a popular model choice due to the scalable learning algorithms introduced by (Kingma & Welling, 2013; Rezende et al., 2014). These approaches maximize a variational lower bound on the intractable log likelihood of the observed data. Burda et al. (2015) introduced a multi-sample v…
Framework for lifelong learning using eigentasks to avoid forgetting and transfer knowledge.
problem Avoiding forgetting and transferring knowledge in lifelong learning.
method Eigentask framework: skills paired with generative models, wake-sleep cycle for learning and consolidation.
result Improved performance in supervised continual learning, evidence of forward knowledge transfer.
Iterative reweighted algorithms, as a class of algorithms for sparse signal recovery, have been found to have better performance than their non-reweighted counterparts. However, for solving the problem of multiple measurement vectors (MMVs), all the existing reweighted algorithms do not account for temporal correlation…
DDVI uses diffusion models for variational inference, improving latent variable model performance.
problem Improving variational inference in latent variable models.
method Introduces diffusion-based variational posteriors trained with a regularized ELBO.
result Outperforms alternative variational posteriors on various benchmarks and a biology task.
A new approach to learning in brain-like networks using adversarial algorithms.
problem Complex inter-dependencies in brain-like networks not compatible with conditional independence assumptions.
method Adversarial algorithm for learning models of perceptual processing.
result The approach can mimic known neural phenomena and yields testable hypotheses.
This paper develops a novel deep recurrent neural network for sequential signal reconstruction.
problem Sequential signal reconstruction from low-dimensional measurements.
method Unfolding a reweighted ℓ1-ℓ1 minimization algorithm to design a deep recurrent neural network. result The proposed reweighted-RNN significantly outperforms existing RNN models in sequential frame reconstruction.
Improved matching for multiple objects using a novel reweighting method.
problem Current multi-object matching methods have limitations and are not robust.
method Proposes a novel iterative reweighting strategy using the graph connection Laplacian.
result Demonstrates superior performance over state-of-the-art methods.
Reweighted l1-algorithms have attracted a lot of attention in the field of applied mathematics. A unified framework of such algorithms has been recently proposed by Zhao and Li. In this paper we construct a few new examples of reweighted l1-methods. These functions are certain concave approximations of the l0-norm func…
Proposes a new method to improve regression models with reweighted samples.
problem Improves regression models' performance under low sample sizes and covariate perturbations.
method Reparametrizes sample weights using a doubly non-negative matrix and solves the reweighted estimate efficiently.
result Adversarial reweighting strategy delivers promising results on various datasets.
Improved learning to reweight using deep interactions between student and teacher models.
problem Limitation of existing learning to reweight methods in utilizing student model's internal states.
method Proposes an algorithm that uses the student model's internal states to the teacher model, which returns adaptive weights to enhance student model training.
result Significant improvement over previous methods in image classification and neural machine translation experiments.
Defense against user shilling attacks in collaborative filtering using edge reweighting.
problem Vulnerability of collaborative filtering to profile injection attacks.
method Adversarial robustness based edge reweighting to attenuate non-robust edges.
result Effective defense against various types of attacks demonstrated through experiments.
We analyze a reweighted version of the Kikuchi approximation for estimating the log partition function of a product distribution defined over a region graph. We establish sufficient conditions for the concavity of our reweighted objective function in terms of weight assignments in the Kikuchi expansion, and show that a…
CurveRL optimizes large model reasoning by reweighting prompts based on their rank and density.
problem Improving large language model reasoning through context reweighting.
method Formulated prompt reweighting as a functional derivative, proposing CurveRL based on quantile coordinate transform.
result CurveRL consistently outperforms existing methods across multiple benchmarks.
Unified analysis of reweighted least-squares algorithms for linear models.
problem Recovering unknown signals from linear measurements using reweighted least squares.
method Unified asymptotic analysis of IRLS, lin-RFM, and alternating minimization algorithms.
result The algorithms can achieve favorable performance in a few iterations with appropriate reweighting.
Federated learning has a variety of applications in multiple domains by utilizing private training data stored on different devices. However, the aggregation process in federated learning is highly vulnerable to adversarial attacks so that the global model may behave abnormally under attacks. To tackle this challenge, …
New reweighted losses improve diffusion model training and image quality.
problem Training and improving diffusion models for image generation.
method Constructing a cascade of time-dependent variational lower bounds.
result Significant improvements in pixel-space image modeling quality.
Domain adaptation provides a powerful set of model training techniques given domain-specific training data and supplemental data with unknown relevance. The techniques are useful when users need to develop models with data from varying sources, of varying quality, or from different time ranges. We build CrossTrainer, a…
Reweighting improves GAN accuracy without sacrificing statistical power.
problem Improving the fidelity of generative models.
method Post-hoc reweighting function applied to generated examples.
result Weighted GAN examples significantly improve accuracy.
New algorithm improves online binary classification with constant time complexity.
problem Online binary classification with rebalancing.
method Non-iteratively reweighted recursive least-squares.
result Exacts converges to batch formulation and outperforms existing algorithms.
Recently, the paradigm of unfolding iterative algorithms into finite-length feed-forward neural networks has achieved a great success in the area of sparse recovery. Benefit from available training data, the learned networks have achieved state-of-the-art performance in respect of both speed and accuracy. However, the …
Resampling outperforms reweighting for correcting biased data in machine learning models.
problem Correcting sampling bias in machine learning models trained on biased data sets.
method Compared resampling and reweighting techniques, focusing on their performance with stochastic gradient algorithms.
result Resampling outperforms reweighting when combined with stochastic gradient algorithms.
Toy model study shows resampling/reweighting can improve feature learning in imbalanced classification.
problem Improving feature learning in imbalanced classification problems.
method High-dimensional toy model with replica method, class-wise resampling/reweighting, and simplified model.
result No resampling/reweighting can sometimes give best feature learning performance.
Machine learning tools are commonly used in modern high energy physics (HEP) experiments. Different models, such as boosted decision trees (BDT) and artificial neural networks (ANN), are widely used in analyses and even in the software triggers. In most cases, these are classification models used to select the "signal"…
Precise scientific analysis in collider-based particle physics is possible because of complex simulations that connect fundamental theories to observable quantities. The significant computational cost of these programs limits the scope, precision, and accuracy of Standard Model measurements and searches for new phenome…
Algorithm samples polygons of fixed edge lengths in any dimension.
problem Sampling random closed polygons with fixed edge lengths in any dimension.
method Weighted edge vectors on unit sphere, Möbius transformation, reweighting factors.
result Algorithm samples polygons according to standard probability measures efficiently.
The paper improves sample reweighting methods for adapting to covariate shifts.
problem Improving accuracy in reproducing kernel Hilbert spaces when data distributions differ.
method Combining known error bounds for reweighted kernel regression in RKHS to show reduced sample size needed for accuracy.
result Under weak smoothness conditions, fewer samples are needed for the same accuracy as standard supervised learning.
Principal component analysis (PCA) is often used to reduce the dimension of data by selecting a few orthonormal vectors that explain most of the variance structure of the data. L1 PCA uses the L1 norm to measure error, whereas the conventional PCA uses the L2 norm. For the L1 PCA problem minimizing the fitting error of…
A bilevel optimization method for reweighting data leads to suboptimal results.
problem Reweighting data for better model performance on a different distribution.
method Formalizes data reweighting as a bilevel optimization problem and analyzes the limitations of classical solvers.
result The final data weights are often very sparse, indicating suboptimal solutions.
Optimizes sample reweighting to match laws under covariate shift using Wasserstein distance.
problem Matching laws of samples with different distributions under covariate shift.
method Minimizes Wasserstein distance between empirical measures of samples using Nearest Neighbors weights.
result Consistent reweighting leads to asymptotic convergence of empirical measures.
Q-MMR evaluates policies using reweighted rewards and moment matching.
problem Off-policy evaluation in finite-horizon MDPs.
method Q-MMR learns scalar weights for data points via a moment matching objective against a value-function discriminator class.
result Data-dependent finite-sample guarantee with a dimension-free error bound.
Method reweights instances and classes to improve robustness in noisy data.
problem Improving deep learning performance in the presence of label noise.
method Formulates constrained optimization problems to assign importance weights to instances and class labels.
result Significant performance gains observed in benchmark datasets with label noise.
We introduce the implicit processes (IPs), a stochastic process that places implicitly defined multivariate distributions over any finite collections of random variables. IPs are therefore highly flexible implicit priors over functions, with examples including data simulators, Bayesian neural networks and non-linear tr…
ARL improves fairness without protected features, showing AUC improvements for worst-case groups.
problem Training fairness in ML without known protected features.
method Adversarially Reweighted Learning (ARL) using non-protected features and task labels.
result ARL improves Rawlsian Max-Min fairness with notable AUC improvements for worst-case groups.
Kernel methods can learn hierarchical polynomials efficiently.
problem Learning hierarchical structure from data.
method Iteratively reweighting kernel machines using derivatives.
result Efficient learning of hierarchical polynomials.
Method reweights auxiliary tasks to reduce data need for main task.
problem Limited labeled data for supervised learning.
method Formulates weighted likelihood function as surrogate prior, minimizing divergence to true prior.
result Effective use of limited labeled data with auxiliary tasks, improving performance.
Iteratively reweighted ℓ1 algorithm is a popular algorithm for solving a large class of optimization problems whose objective is the sum of a Lipschitz differentiable loss function and a possibly nonconvex sparsity inducing regularizer. In this paper, motivated by the success of extrapolation techniques in accele…
Ribbon: Scalable Approximation and Robust Uncertainty Quantification
problem Reliably quantifying predictive uncertainty for complex models
method Ribbon, a scalable approximation to Dirichlet-reweighted bootstrap uncertainty
result Asymptotically equivalent to a flat-prior Laplace approximation under correct likelihood specification, recovers robust sandwich covariance under misspecification