Improved uncertainty estimation in neural networks with VBLL.
problem Improving uncertainty estimation in neural networks.
method Deterministic variational formulation for training Bayesian last layer neural networks.
result Improves predictive accuracy, calibration, and out-of-distribution detection.
Improved continual learning method using variational inference and FiLM layers.
problem Training models on new tasks and datasets in an online fashion.
method Generalized Variational Continual Learning (GVCL) with likelihood-tempering and FiLM layers.
result GVCL outperforms existing baselines in both small and large datasets, providing better calibration.
New method improves BLL models for complex datasets.
problem Limited expressive capacity of Gaussian priors in BLL models.
method Combines diffusion techniques and implicit priors for variational learning.
result Enhanced predictive accuracy and uncertainty quantification.
Global inducing points improve Bayesian neural network performance.
problem Improving Bayesian neural network performance.
method Adapting correlated approximate posterior to all layers in a Bayesian neural network and deep Gaussian processes using learned global inducing points.
result State-of-the-art performance on CIFAR-10 (86.7%) without data augmentation or tempering.
Variational Autoencoders are powerful models for unsupervised learning. However deep models with several layers of dependent stochastic variables are difficult to train which limits the improvements obtained using these highly expressive models. We propose a new inference model, the Ladder Variational Autoencoder, that…
A new unpooling layer enhances graph generation in molecular models.
problem Efficient graph generation for complex models like molecules.
method Trainable unpooling layer that enlarges and restructures graphs.
result The unpooling layer improves graph generation in molecular models.
The combination of the re-parameterization trick with the use of variational auto-encoders has caused a sensation in Bayesian deep learning, allowing the training of realistic generative models of images and has considerably increased our ability to use scalable latent variable models. The re-parameterization trick is …
In this note we consider setups in which variational objectives for Bayesian neural networks can be computed in closed form. In particular we focus on single-layer networks in which the activation function is piecewise polynomial (e.g. ReLU). In this case we show that for a Normal likelihood and structured Normal varia…
We provide a theoretical explanation of the role of the number of nodes at each layer in deep neural networks. We prove that the largest variation of a deep neural network with ReLU activation function arises when the layer with the fewest nodes changes its activation pattern. An important implication is that deep neur…
Combines neural networks with variational inference for better uncertainty quantification.
problem Overconfident predictions from traditional neural networks and time-consuming Bayesian optimization.
method VIFO (Variational Inference on the Final-Layer Output) using neural networks to learn mean and variance.
result VIFO provides a good tradeoff in run time and uncertainty quantification, especially for out of distribution data.
Derives formulae for general permutation equivariant layers and presents a second order graph variational encoder.
problem Tackles the limitation of previous equivariant neural networks by considering permutations of matrices.
method Derives formulae for general permutation equivariant layers, including matrix permutations. Presents a second order graph variational encoder.
result Latent distribution of equivariant generative models must be exchangeable.
Deep Transformed Gaussian Processes extend TGPs with variational inference for scalable multi-layer modeling.
problem Flexible modeling of complex data distributions.
method DTGPs are a multi-layer model of TGPs using variational inference for scalability.
result DTGPs achieve good scalability and performance in multiple regression datasets.
New BNN method reduces training time and model size.
problem Overconfident predictions in deep learning models.
method Designing STF-BNN for efficient scaling of BNNs.
result Significantly reduces training time and model size compared to vanilla BNNs.
Paper introduces TSSDMN for modeling dynamic multilayer networks.
problem Capturing temporal and cross-layer dynamics in multilayer networks.
method Tensor State Space Model (TSSDMN) using symmetric Tucker decomposition.
result TSSDMN uniquely captures temporal dynamics within and across layers.
VPP learns joint policies for multi-agent RL through interactions.
problem Learning effective joint policies for multi-agent reinforcement learning.
method VPP integrates variational inference into policy layers for efficient sampling and differentiability.
result VPP outperforms previous methods on large-scale multi-agent tasks.
Bayesian optimization improved with variational last layer training.
problem Bayesian optimization performance on complex input correlations.
method Connecting variational Bayesian last layer training to exact conditioning in Gaussian processes, developing an efficient online training algorithm.
result VBLL networks significantly outperform GPs and match well-tuned GPs on benchmark tasks.
New mass inequalities and proofs for causal variational principles.
problem Proving new mass inequalities for causal variational principles.
method Proved a new inequality for minimizers of causal variational principles and applied it to prove the positive mass theorem.
result Introduced a positive quasilocal mass and proved new mass inequalities.
Adding linear layers to ReLU networks favors functions with low mixed variation.
problem Understanding function space bias in overparameterized neural networks.
method Examined a family of networks with varying depths and same capacity but different representation costs, focusing on the effect of adding linear layers to the input side.
result Adding linear layers to shallow ReLU networks results in a bias towards functions with low mixed variation, which can be well approximated by single- or multi-index models.
EMFs combine deep learning and probabilistic models for better density estimation.
problem Combining domain-specific knowledge with general-purpose deep learning.
method Alternating transformations with structured layers that embed domain-specific inductive biases.
result EMFs induce desirable properties like multimodality and hierarchical coupling.
Based on conservation laws for surface layer integrals for critical points of causal variational principles, it is shown how jet spaces can be endowed with an almost-complex structure. We analyze under which conditions the almost-complex structure can be integrated to a canonical complex structure. Combined with the sc…
We present a new flavor of Variational Autoencoder (VAE) that interpolates seamlessly between unsupervised, semi-supervised and fully supervised learning domains. We show that unlabeled datapoints not only boost unsupervised tasks, but also the classification performance. Vice versa, every label not only improves class…
SFSVI uses Gaussian mixtures to approximate neural network outputs for continual learning.
problem Learning new tasks without forgetting old ones in neural networks.
method Sequential function-space variational inference with Gaussian mixture approximation.
result Gaussian mixture SFSVI outperforms other methods in continual learning.
In this paper we introduce deep Gaussian process (GP) models. Deep GPs are a deep belief network based on Gaussian process mappings. The data is modeled as the output of a multivariate GP. The inputs to that Gaussian process are then governed by another GP. A single layer model is equivalent to a standard GP or the GP …
Gaussian processes (GPs) are a good choice for function approximation as they are flexible, robust to over-fitting, and provide well-calibrated predictive uncertainty. Deep Gaussian processes (DGPs) are multi-layer generalisations of GPs, but inference in these models has proved challenging. Existing approaches to infe…
In this note we present a generative model of natural images consisting of a deep hierarchy of layers of latent random variables, each of which follows a new type of distribution that we call rectified Gaussian. These rectified Gaussian units allow spike-and-slab type sparsity, while retaining the differentiability nec…
Bayesian neural networks improve uncertainty calibration without sacrificing accuracy.
problem Bayesian neural networks struggle with uncertainty calibration and high-dimensional geometry.
method Model uncertainty only in weight directions using a von Mises-Fisher posterior on the unit sphere, deriving a compact KL term.
result A lightweight, dimension-aware variational unit improves calibration without sacrificing accuracy.
Alternative neural network training using monotone variational inequality.
problem Training neural networks efficiently and with guarantees.
method Using monotone variational inequality to solve non-convex problems efficiently.
result Our approach leads to fast convergence and competitive performance compared to traditional methods.
For any ReLU network there is a representation in which the sum of the absolute values of the weights into each node is exactly 1, and the input layer variables are multiplied by a value V coinciding with the total variation of the path weights. Implications are given for Gaussian complexity, Rademacher complexity,…
Proposes a Bayesian approach for automatic node selection in sparse neural networks.
problem Reduces structural complexity and computational speedup in large-scale predictive models.
method Uses spike-and-slab Gaussian priors and variational Bayes approach for node selection.
result Establishes variational posterior consistency and optimal contraction rates for sparse networks.
Variational autoencoders (VAEs), that are built upon deep neural networks have emerged as popular generative models in computer vision. Most of the work towards improving variational autoencoders has focused mainly on making the approximations to the posterior flexible and accurate, leading to tremendous progress. Howe…
Proposes a deep probabilistic multi-view model for multi-view learning.
problem Learning from multiple related views with shared latent structure.
method Probabilistic Canonical Correlation Analysis (CCA) in latent space, deep generative networks, variational inference.
result Efficient variational inference approximates posterior distributions of latent multi-view layer.
Enhances speech emotion recognition by adapting to varying time scales.
problem Robust emotion recognition from speech audio with temporal variations.
method Introduces multi-time-scale (MTS) convolutional layers to CNNs.
result MTS layers improve generalization, especially on smaller datasets.
Simplified DGPs training by fixing inducing inputs to subset of data.
problem Challenging training of deep Gaussian processes.
method Fixed subset of data for inducing inputs, variational sampling.
result Significant reduction in trainable parameters and computation cost without performance degradation.
Study spinorial Yamabe problem on product manifolds, proving spike layer solutions.
problem Existence of solutions on product manifolds with curvature effect.
method Variational methods and blow-up techniques.
result Existence of solutions depending only on one factor and exhibiting spike layers.
We present a variation of the Autoencoder (AE) that explicitly maximizes the mutual information between the input data and the hidden representation. The proposed model, the InfoMax Autoencoder (IMAE), by construction is able to learn a robust representation and good prototypes of the data. IMAE is compared both theore…
p3VAE combines physics and machine learning for robust data representations.
problem Improving machine learning models' robustness to environmental factors of variation.
method Physics-informed variational autoencoder integrating physical knowledge with neural networks.
result p3VAE outperforms competing models in extrapolation and interpretability. Hybrid Bayesian neural networks use function uncertainty for probabilistic inference.
problem Uncertainty in neural network weights is hard to specify and interpret.
method Integrates probabilistic layers with standard deterministic layers for function uncertainty.
result Improves probabilistic inference by encoding function uncertainty.
Develops a flexible deep autoencoding topic model with scalable hybrid Bayesian inference.
problem Flexible and interpretable document analysis models.
method DATM with hybrid Bayesian inference, including topic-layer-adaptive stochastic gradient Riemannian MCMC and Weibull variational encoder.
result Demonstrates scalability and efficacy on big corpora in unsupervised and supervised learning tasks.
Nowadays, deep learning can be employed to a wide ranges of fields including medicine, engineering, etc. In deep learning, Convolutional Neural Network (CNN) is extensively used in the pattern and sequence recognition, video analysis, natural language processing, spam detection, topic categorization, regression analysi…
Two main obstacles preventing the widespread adoption of variational Bayesian neural networks are the high parameter overhead that makes them infeasible on large networks, and the difficulty of implementation, which can be thought of as "programming overhead." MC dropout [Gal and Ghahramani, 2016] is popular because it…
Paper proposes an online learning method with multi-level adaptivity for diverse loss functions.
problem Online learning with unknown types and curvatures of functions.
method Multi-layer online ensemble approach with gradient variations.
result Achieves improved regret bounds for different types of loss functions.
Model predicts anti-cancer drug responses using gene and molecular data.
problem Expensive and time-consuming cancer drug discovery and tailoring.
method Uses variational autoencoders and multi-layer perceptrons to encode gene expression and drug data.
result High average R2 of 0.83 and 0.845 in predicting drug responses for breast and pan-cancer cell lines, respectively. VMoER improves uncertainty quantification in MoE layers for scalable foundation models.
problem Uncertainty quantification in large-scale models like MoE layers.
method Structured Bayesian approach with amortized variational inference over routing logits and temperature parameter inference.
result Improves routing stability, reduces calibration error, and increases AUROC by 12%.
This paper introduces a new member of the family of Variational Autoencoders (VAE) that constrains the rate of information transferred by the latent layer. The latent layer is interpreted as a communication channel, the information rate of which is bound by imposing a pre-set signal-to-noise ratio. The new constraint s…
We introduce a variational Bayesian neural network where the parameters are governed via a probability distribution on random matrices. Specifically, we employ a matrix variate Gaussian \cite{gupta1999matrix} parameter posterior distribution where we explicitly model the covariance among the input and output dimensions…
Dirichlet pruning compresses neural networks by removing unimportant units.
problem Compressing large neural network models without sacrificing performance.
method Assigns Dirichlet distribution over network layers' units and uses variational inference to estimate parameters.
result Achieves state-of-the-art compression performance on larger architectures like VGG and ResNet.
CardiCat generates synthetic data for high-cardinality tabular datasets.
problem Learning complexities of high-cardinality categorical features in tabular data.
method Substitutes one-hot encoding with regularized dual encoder-decoder embedding layers.
result Generates high-quality synthetic data with a smaller parameter space.
The paper analyzes Bayesian neural networks trained with VI, proving a law of large numbers for different schemes.
problem Training Bayesian neural networks with variational inference.
method Analyzes three training schemes: exact estimation, Bayes by Backprop, and Minimal VI.
result All training schemes converge to the same mean-field limit.