New findings show optimal noise in contrastive learning is not the same as data distribution.
problem The optimal noise distribution in contrastive learning is not the same as the data distribution.
method Empirical and theoretical analysis of contrastive learning methods.
result Deviation from the assumption of equal noise and data distribution leads to better statistical estimators.
NCE and CD are shown to be equivalent ML methods.
problem Estimating unnormalised models without normalisation constant.
method NCE uses proxy criterion, CD uses importance sampling.
result NCE and CD are equivalent ML methods.
Noise-Contrastive Estimation improves efficiency for estimating log-likelihood of complex point processes.
problem Estimating log-likelihood of complex multivariate point processes is computationally expensive.
method Noise-Contrastive Estimation adapted for multivariate point processes, with provable guarantees.
result Our method achieves similar log-likelihood with fewer evaluations and less time.
Unified view on learning unnormalized distributions using NCE.
problem Learning unnormalized distributions across different communities.
method Noise-Contrastive Estimation (NCE) for a unified perspective.
result Established finite-sample convergence rates for exponential families.
Contrastive learning performance doesn't degrade with more negative samples.
problem Theoretical and empirical evidence of negative samples hurting performance in contrastive learning.
method Simple theoretical setting and empirical support on CIFAR-10 and CIFAR-100 datasets.
result Contrastive learning performance does not degrade with the number of negative samples.
New insights into noise distribution for self-supervised learning.
problem Challenges the assumption that optimal noise should match data distribution.
method Turns to Noise-Contrastive Estimation (NCE) to define optimality of noise distribution.
result Optimal noise distribution is different from data distribution, challenging GANs assumption.
Calibration of simplified vine copulas using noise contrastive estimation
problem Modeling complex multivariate dependence structures
method Noise contrastive estimation for calibration
result Improved model accuracy when simplifying assumption is violated
Researchers improve NCE by addressing its flat loss landscape issues.
problem NCE's poor performance due to an ill-behaved loss landscape.
method Introduced eNCE with an exponential loss and normalized gradient descent.
result Proven that landscape issues arise from inappropriate noise distribution.
Many parametric statistical models are not properly normalised and only specified up to an intractable partition function, which renders parameter estimation difficult. Examples of unnormalised models are Gibbs distributions, Markov random fields, and neural network models in unsupervised deep learning. In previous wor…
A new whole-sentence language model - neural trans-dimensional random field language model (neural TRF LM), where sentences are modeled as a collection of random fields, and the potential function is defined by a neural network, has been introduced and successfully trained by noise-contrastive estimation (NCE). In this…
PiNGDA learns beneficial noise for graph augmentation stability.
problem Challenges in generating effective and stable graph augmentations.
method PiNGDA uses positive-incentive noise to scientifically analyze and generate beneficial graph augmentations.
result PiNGDA improves GCL performance by learning beneficial noise on graph topology and attributes.
Current meta-learning approaches focus on learning functional representations of relationships between variables, i.e. on estimating conditional expectations in regression. In many applications, however, we are faced with conditional distributions which cannot be meaningfully summarized using expectation only (due to e…
There are many models, often called unnormalized models, whose normalizing constants are not calculated in closed form. Maximum likelihood estimation is not directly applicable to unnormalized models. Score matching, contrastive divergence method, pseudo-likelihood, Monte Carlo maximum likelihood, and noise contrastive…
Quantum algorithm reduces CVA risk-neutral expectation estimation costs.
problem Reducing Monte Carlo sampling cost for CVA on real quantum hardware.
method Noise-aware quantum workflow combining market calibration, discretisation, and oracle construction.
result CABIQAE achieves lower classical post-processing runtime and more effective error exploitation.
Improved VAEs by training a contrastive prior to match posterior.
problem Prior hole problem in VAEs, leading to poor image generation.
method Introduced a contrastive energy-based prior and trained it using noise contrastive estimation.
result Significant improvement in VAE generative performance on various datasets.
Paper proposes a new loss function for conditional models using soft targets.
problem Improving generalization performance of deep neural networks on supervised classification tasks.
method Introduces a new loss function compatible with soft targets, based on noise contrastive estimation.
result Soft target InfoNCE loss performs on par with cross-entropy baselines and outperforms other losses.
Replicated Softmax model, a well-known undirected topic model, is powerful in extracting semantic representations of documents. Traditional learning strategies such as Contrastive Divergence are very inefficient. This paper provides a novel estimator to speed up the learning based on Noise Contrastive Estimate, extende…
New methods improve Monte Carlo estimation of partition functions.
problem Estimating the normalization constant of complex distributions.
method Annealing through paths of distributions to estimate partition functions.
result Optimal path for estimation is arithmetic, improving efficiency.
Gaussians as noise in NCE lead to exponentially bad conditioning, hindering its efficiency.
problem Exponential conditioning of Hessian in NCE with Gaussian noise.
method Using Gaussian as the noise distribution in NCE.
result Gaussian noise in NCE leads to exponentially bad conditioning of the loss Hessian.
NCVis speeds up data visualization for large datasets.
problem Performance issues in t-SNE for large datasets.
method Noise contrastive estimation for scalable visualization.
result NCVis outperforms state-of-the-art techniques in speed and quality.
MCD reformulates conditional density estimation into binary classification.
problem Conditional density estimation in statistical and machine learning.
method Marginal Contrastive Discrimination, reformulating into marginal and ratio density functions for binary classification.
result Significantly outperforms existing methods on most density models and regression datasets.
Obtaining reliable uncertainty estimates of neural network predictions is a long standing challenge. Bayesian neural networks have been proposed as a solution, but it remains open how to specify their prior. In particular, the common practice of an independent normal prior in weight space imposes relatively weak constr…
Unnormalised latent variable models are a broad and flexible class of statistical models. However, learning their parameters from data is intractable, and few estimation techniques are currently available for such models. To increase the number of techniques in our arsenal, we propose variational noise-contrastive esti…
PCA++ improves robustness to background noise in contrastive learning.
problem Recovering shared signal subspaces from positive pairs in high-dimensional data with structured background noise.
method PCA++ uses hard uniformity-constrained contrastive learning to enforce identity covariance on projected features.
result PCA++ outperforms standard PCA and alignment-only PCA+ in simulations and real-world datasets.
DACL tackles domain-specific contrastive learning by using Mixup noise.
problem Domain-specific contrastive learning methods rely on data augmentation techniques that require domain knowledge.
method DACL uses Mixup noise to create similar and dissimilar examples without domain-specific data augmentation.
result DACL outperforms other domain-agnostic noising methods and combines well with domain-specific methods.
We take steps towards understanding the "posterior collapse (PC)" difficulty in variational autoencoders (VAEs),~i.e. a degenerate optimum in which the latent codes become independent of their corresponding inputs. We rely on calculus of variations and theoretically explore a few popular VAE models, showing that PC alw…
AVICA estimates noise levels for better group ICA source recovery.
problem Estimating shared independent sources from multiple noisy views.
method AVICA models each view as a linear mixture of shared sources with additive noise, optimizing noise levels alongside sources.
result AVICA yields better source estimates than other methods, especially in real-world applications like MEG and fMRI.
We prove a new and general concentration inequality for the excess risk in least-squares regression with random design and heteroscedastic noise. No specific structure is required on the model, except the existence of a suitable function that controls the local suprema of the empirical process. So far, only the case of…
Improved VAE model enhances uncertainty estimation for out-of-distribution samples.
problem VAEs assign higher likelihood to out-of-distribution inputs.
method INCPVAE integrates noise contrastive prior into VAEs for reliable uncertainty estimation.
result INCPVAE outperforms standard VAEs in uncertainty estimation for OOD inputs.
This paper broadens contrastive learning for disentangled representations without strict data distribution assumptions.
problem Learning disentangled representations from data with specific assumptions.
method Extends theoretical guarantees for disentanglement to a broader family of contrastive methods, relaxing data distribution assumptions.
result Identifiability of true latents for four contrastive losses proved without common independence assumptions.
We consider classification in the presence of class-dependent asymmetric label noise with unknown noise probabilities. In this setting, identifiability conditions are known, but additional assumptions were shown to be required for finite sample rates, and so far only the parametric rate has been obtained. Assuming thes…
Adaptive multi-stage density ratio estimation improves learning of latent space EBM.
problem Learning energy-based models in latent space is computationally expensive and challenging.
method Adaptive multi-stage density ratio estimation using NCE to bridge the gap between prior and posterior densities.
result The method enables more expressive prior models and sharpens the latent space EBM.
This paper proposes a method for multi-class classification problems, where the number of classes K is large. The method, referred to as Candidates vs. Noises Estimation (CANE), selects a small subset of candidate classes and samples the remaining classes. We show that CANE is always consistent and computationally effi…
We show that the Bregman divergence provides a rich framework to estimate unnormalized statistical models for continuous or discrete random variables, that is, models which do not integrate or sum to one, respectively. We prove that recent estimation methods such as noise-contrastive estimation, ratio matching, and sco…
Self-supervised model detects phoneme boundaries without annotations.
problem Unsupervised phoneme segmentation without manual annotations.
method Convolutional neural network trained with Noise-Contrastive Estimation.
result Model outperforms baselines on TIMIT and Buckeye corpora.
Paper trains two models to improve synthesis quality and unsupervised learning.
problem Improving synthesis quality and unsupervised learning.
method Joint training of an energy-based model and a flow-based model using adversarial value function.
result Significant improvement in synthesis quality of the flow model.
We present an efficient algorithm for maximum likelihood estimation (MLE) of exponential family models, with a general parametrization of the energy function that includes neural networks. We exploit the primal-dual view of the MLE with a kinetics augmented model to obtain an estimate associated with an adversarial dua…
We develop a general method for estimating a finite mixture of non-normalized models. Here, a non-normalized model is defined to be a parametric distribution with an intractable normalization constant. Existing methods for estimating non-normalized models without computing the normalization constant are not applicable …
SuNCEt accelerates contrastive learning with minimal labeled data.
problem Efficiently learning visual representations with limited labeled data.
method Noise-contrastive estimation and neighbourhood component analysis-based semi-supervised loss.
result SuNCEt achieves semi-supervised learning accuracy with less than half the labeled data.
Trans-dimensional random field language models (TRF LMs) where sentences are modeled as a collection of random fields, have shown close performance with LSTM LMs in speech recognition and are computationally more efficient in inference. However, the training efficiency of neural TRF LMs is not satisfactory, which limit…
Improved visual representation learning with conditional negative sampling.
problem Learning strong unsupervised visual representations using contrastive learning.
method Introduce a family of mutual information estimators that sample negatives conditionally.
result Improves accuracy by 2-5% points on four standard image datasets.
New method tests causal association using noise contrastive backdoor adjustment.
problem Testing causal association in complex settings with many confounders.
method Backdoor-HSIC (bd-HSIC) using HSIC for independence testing.
result Calibrated and powerful for binary and continuous treatments with many confounders.
We consider the problem of estimating the mean and covariance of a distribution from iid samples in Rn, in the presence of an η fraction of malicious noise; this is in contrast to much recent work where the noise itself is assumed to be from a distribution of known type. The agnostic problem includes many…
This paper proposes a self-supervised learning approach for video features that results in significantly improved performance on downstream tasks (such as video classification, captioning and segmentation) compared to existing methods. Our method extends the BERT model for text sequences to the case of sequences of rea…
Estimates intrinsic dimension of data sets robustly to noise.
problem Estimating intrinsic dimension of noisy data sets.
method Quantum Cognition Machine Learning for data representation and spectral gap detection.
result Robust estimation of intrinsic dimension in the presence of Gaussian noise.
Letter analyzes training dynamics of a nonlinear contrastive learning model in high dimensions.
problem Understanding training dynamics of nonlinear contrastive learning models in high-dimensional settings.
method High-dimensional analysis using McKean-Vlasov PDEs and low-dimensional ODEs.
result The model's performance evolves according to specific ODEs, revealing features like feature learnability and noise effects.
We solve matrix denoising with both row and column correlations, setting limits and designing optimal methods.
problem Matrix denoising with doubly heteroscedastic noise (both row and column correlations).
method Established information-theoretic and algorithmic limits, designed a novel spectral estimator with optimality guarantees.
result The novel spectral estimator achieves positive correlation with the signal and Bayes-optimal error under one-sided heteroscedasticity.
Strong inductive biases prevent harmless interpolation in overparameterized models.
problem Understanding the conditions under which overparameterized models can interpolate noise without overfitting.
method Theoretical analysis of high-dimensional kernel regression and deep neural networks, focusing on the role of inductive biases.
result The strength of an estimator's inductive bias determines whether interpolation is harmless or requires fitting noise for good generalization.