This paper distills Bayesian posterior expectations for deep neural networks.
problem Improving deep neural network performance and uncertainty quantification.
method Develops a framework for distilling expectations from Bayesian posterior distributions using Monte Carlo samples.
result The framework successfully distills posterior predictive distribution and expected entropy.
The paper extends entropy maximization to multiscale settings and applies it to neural networks.
problem Achieving optimal risk bounds in neural networks using multiscale entropy.
method Generalizing maximum entropy to multiscale settings and applying it to neural networks.
result The multiscale Gibbs posterior can achieve a smaller excess risk than the single-scale Gibbs posterior in a teacher-student scenario.
Study on optimal information acquisition in Kyle model with entropy cost.
problem Optimal information acquisition in Kyle model with entropy cost.
method Continuous signals are optimal, and any signal with a logit posterior distribution yields the same ex-ante value.
result Posterior expected payoff becomes normally distributed as information acquisition cost increases.
New entropy-based objective for sparse coding improves learning.
problem Sparse coding with probabilistic priors and non-Gaussian observables.
method Derive a solely entropy-based learning objective for sparse coding parameters.
result Fully analytical ELBO objective for sparse coding with non-trivial posterior approximations.
New decision-theoretic characterization separates belief and decision posteriors.
problem Understanding the conditions under which loss-based updating coincides with Bayesian updating.
method Decision-theoretic approach to distinguish belief and decision posteriors.
result Generalized Bayes coincides with ordinary Bayesian updating only if the loss is proportional to negative log-likelihood.
Infinite mixture models are commonly used for clustering. One can sample from the posterior of mixture assignments by Monte Carlo methods or find its maximum a posteriori solution by optimization. However, in some problems the posterior is diffuse and it is hard to interpret the sampled partitionings. In this paper, we…
Entropy regularization improves interpretability of probabilistic clustering models.
problem Bayesian nonparametric mixture models often produce unbalanced cluster frequencies.
method Interpreting the posterior as penalized likelihood, entropy regularization reduces sparsely-populated clusters.
result The proposed entropy-regularized estimator enhances interpretability without sacrificing computational convenience.
We show that Entropy-SGD (Chaudhari et al., 2017), when viewed as a learning algorithm, optimizes a PAC-Bayes bound on the risk of a Gibbs (posterior) classifier, i.e., a randomized classifier obtained by a risk-sensitive perturbation of the weights of a learned classifier. Entropy-SGD works by optimizing the bound's p…
Detecting and recovering labels in binomial logistic mixtures is challenging due to an information gap.
problem Detecting and recovering labels in binomial logistic mixtures
method Propose two feasibility-aware inference procedures
result Avoid misleading component selections and improve label probability calibration
New approach to portfolio optimization shows entropy regularization is ineffective.
problem Entropy regularization in mean-variance portfolio optimization under drift uncertainty.
method Combining Bayesian filtering and stochastic policy optimization.
result Entropy regularization does not accelerate learning about unknown drift.
Bayesian Markowitz portfolio problem shows entropy regularization is ineffective.
problem Entropy regularization in Bayesian Markowitz portfolio optimization.
method Combines continuous-time Bayesian filtering with stochastic policy optimization.
result Entropy regularization does not accelerate learning of unknown drift.
This paper improves Bayesian inference for predictive models with limited data.
problem Effective uncertainty quantification for training predictive models with limited data.
method Entropy-regularized gradient estimators to approximate the Bayesian posterior.
result The method generates diverse samples from the posterior distribution efficiently.
One of the core problems in variational inference is a choice of approximate posterior distribution. It is crucial to trade-off between efficient inference with simple families as mean-field models and accuracy of inference. We propose a variant of a greedy approximation of the posterior distribution with tractable bas…
Conservation laws improve diffusion model training by optimizing likelihood.
problem Training diffusion models with denoising objectives.
method Developed conservation laws based on GEXIT functions for memoryless noise processes.
result Unified characterization of diffusion model likelihood, reducing training to learning marginal posteriors.
This paper proposes a new method for efficient data compression using Bayesian neural networks.
problem Efficient compression of data represented as functions mapping coordinates to signal values.
method Overfitting variational Bayesian neural networks to the data and compressing an approximate posterior weight sample using relative entropy coding.
result Our method achieves strong performance on image and audio compression while retaining simplicity.
Unified approach for selecting summary statistics in ABC.
problem Efficient inference from large datasets in likelihood-free methods.
method Characterizing and unifying three classes of summary statistics, minimizing expected posterior entropy.
result EPE-minimizing summaries lead to competitive posterior inference.
Information-theoretic quantities, such as conditional entropy and mutual information, are critical data summaries for quantifying uncertainty. Current widely used approaches for computing such quantities rely on nearest neighbor methods and exhibit both strong performance and theoretical guarantees in certain simple sc…
SymCircuit learns PC structure via entropy-regularized RL, improving inference efficiency and accuracy.
problem Greedy algorithms in PC structure learning lead to suboptimal solutions.
method Entropy-regularized reinforcement learning to train a learned generative policy for PC structure inference.
result SymCircuit learns the optimal policy as a tempered Bayesian posterior, improving inference efficiency and accuracy.
Improves GP models with known bounds for sampling and optimization.
problem Functions with known upper and lower bounds.
method Transforms GP models with bounds for posterior sampling and BO.
result Bounded entropy search (BES) selects points satisfying constraints.
We propose a max-pooling based loss function for training Long Short-Term Memory (LSTM) networks for small-footprint keyword spotting (KWS), with low CPU, memory, and latency requirements. The max-pooling loss training can be further guided by initializing with a cross-entropy loss trained network. A posterior smoothin…
VAR-GPs solve continual learning by updating posteriors sequentially.
problem Catastrophic forgetting in sequential learning tasks.
method Sparse inducing point approximations and auto-regressive variational distribution.
result VAR-GPs prevent catastrophic forgetting and outperform baselines.
Many inference problems involving questions of optimality ask for the maximum or the minimum of a finite set of unknown quantities. This technical report derives the first two posterior moments of the maximum of two correlated Gaussian variables and the first two posterior moments of the two generating variables (corre…
LES optimizes designs by sampling descent sequences, achieving strong sample efficiency.
problem Optimizing large, complex design spaces is infeasible and unnecessary.
method LES uses Bayesian optimization to target solutions reachable by iterative optimizers.
result LES achieves strong sample efficiency compared to existing methods.
Proposes an active RBI framework using Rényi information measures for more informed decision-making.
problem Optimal latent variable estimates in real-time settings with streaming noisy observations.
method Unified inference and query selection steps through Rényi entropy and α-divergence; new objective called Momentum for exploration.
result Analytically demonstrates superior performance compared to conventional methods like mutual information.
PAC-Bayes bounds for Gibbs posteriors derived via singular learning theory.
problem Generalization bounds for overparameterized models with data-dependent priors.
method Explicit non-asymptotic PAC-Bayes bounds using singular learning theory.
result Explicit posterior-averaged risk bounds for overparameterized models.
Synthesizes sensor likelihoods to enforce accuracy constraints in uncertain systems.
problem Designing sensing architectures for systems with uncertain or unavailable sensor models and accuracy requirements.
method Inverts the design flow, synthesizing measurement likelihoods that minimize Kullback-Leibler divergence from the prior while enforcing an accuracy bound.
result The method synthesizes a maximum-entropy posterior and induced likelihood, accommodating various discrepancy metrics.
A new method, REC, compresses images by encoding their latent representations efficiently.
problem Efficiently compressing single images with latent representations.
method Relative Entropy Coding (REC) that directly encodes latent representations with codelength close to relative entropy.
result REC is more efficient for single image compression compared to previous methods and is competitive for lossy compression.
In this paper we study the probabilistic properties of the posteriors in a speech recognition system that uses a deep neural network (DNN) for acoustic modeling. We do this by reducing Kaldi's DNN shared pdf-id posteriors to phone likelihoods, and using test set forced alignments to evaluate these using a calibration s…
Proposes a method to sample from flat basins of posterior distributions in Bayesian deep learning.
problem Sampling from multi-modal posterior distributions leads to overfitting due to trapping in bad modes.
method Introduces an auxiliary guiding variable to bias MCMC sampling towards flat basins of the energy landscape.
result The method converges faster and outperforms existing methods in sampling from flat basins of the posterior.
PPT optimizes transformer behavior by steering its latent posterior using prior samples.
problem Eliciting desired behavior from transformers without backpropagation.
method Posterior Prefix Tuning (PPT) uses predictive Monte Carlo (PMC) samples and importance sampling to optimize the latent posterior.
result PPT optimizes transformer behavior without backpropagation, achieving high utility across different utility functions.
This paper optimizes trading strategies to minimize risk and maximize profit while accounting for market uncertainty.
problem Optimizing trading strategies to minimize risk and maximize profit while accounting for market uncertainty.
method Relative entropy-regularized robust optimal control problem, modeled as a stochastic differential game.
result Analytical expressions for optimal strategy and trajectory are derived under specific assumptions.
New method optimizes Bayesian optimization for high-dimensional posterior samples.
problem Difficult inner-loop optimization of posterior sample paths in Bayesian optimization.
method Global rootfinding approach with carefully selected starting points.
result The method discovers the global optimum most of the time with just one starting point per set.
We introduce the Mutual Information Machine (MIM), a probabilistic auto-encoder for learning joint distributions over observations and latent variables. MIM reflects three design principles: 1) low divergence, to encourage the encoder and decoder to learn consistent factorizations of the same underlying distribution; 2…
Proposes a new method to measure epistemic uncertainty in Bayesian neural networks.
problem Measuring epistemic uncertainty in Bayesian neural networks for out-of-distribution detection.
method Proposes measuring disagreement between logits and their pre-softmax counterparts as an epistemic uncertainty measure.
result Proposed epistemic uncertainty scores outperform mutual information and equal predictive entropy performance.
New method reduces GP bandit complexity while maintaining good performance.
problem Computational burden in Bayesian optimization with Gaussian processes.
method Information thresholding to compress GP posterior and reduce complexity.
result Sublinear regret bounds with sublinear posterior complexity.
We analyze the problem of learning a single user's preferences in an active learning setting, sequentially and adaptively querying the user over a finite time horizon. Learning is conducted via choice-based queries, where the user selects her preferred option among a small subset of offered alternatives. These queries …
A new measure helps compute suboptimality in entropy-regularized methods.
problem Computing suboptimality in entropy-regularized variational objectives when unnormalised densities are unavailable.
method Introduced 'kernel gradient discrepancy' (KGD) to compute suboptimality explicitly.
result KGD characterizes kernel Stein discrepancy (KSD) in the standard Bayesian context and measures variational gradient size.
We develop a new method for regularising neural networks. We learn a probability distribution over the activations of all layers of the model and then insert imputed values into the network during training. We obtain a posterior for an arbitrary subset of activations conditioned on the remainder. This is a generalisati…
A new CoVaR framework integrates expert views using entropy pooling.
problem Risk assessment and spillover effects from diverse expert views.
method Entropy pooling method to integrate expert views and compute general CoVaR.
result General CoVaR shows linear relationships with expectations and differences in expectations, and nonlinear dependencies with variance, quantiles, and correlation.
New summary measures reveal geometric structure in weighted measures on manifolds.
problem Lack of geometric information in standard weight-only summaries.
method Heat-kernel entropy profiles, tracking nonuniformity across scales.
result Geometric effective sample size discounts nearby or duplicate particles.
New method recovers clean data from corrupted samples.
problem Recovering clean data from corrupted samples with uncertainty.
method Probabilistic Tomographic Auto-Encoder method that derives reduced entropy condition approximate inference.
result Superior performance in imputation and de-noising compared to existing methods.
Transformers mimic Bayesian reasoning in controlled settings, revealing geometric mechanisms.
problem Verifying if transformers perform Bayesian reasoning rigorously in natural data.
method Constructing Bayesian wind tunnels with known posteriors and proving memorization impossibility.
result Transformers achieve 10−3-10−4 bit accuracy in Bayesian posteriors, while MLPs fail. This study proposes hidden state curiosity to enhance RL models' resilience against noise.
problem Curiosity traps in RL models distract agents from discovering novel experiences.
method Proposed hidden state curiosity based on the Free Energy Principle to reward agents for KL divergence between predictive priors and posteriors.
result Agents with hidden state curiosity are more resilient against curiosity traps compared to those with prediction error curiosity.
Word2vec (Mikolov et al., 2013) has proven to be successful in natural language processing by capturing the semantic relationships between different words. Built on top of single-word embeddings, paragraph vectors (Le and Mikolov, 2014) find fixed-length representations for pieces of text with arbitrary lengths, such a…
We study large-scale kernel methods for acoustic modeling and compare to DNNs on performance metrics related to both acoustic modeling and recognition. Measuring perplexity and frame-level classification accuracy, kernel-based acoustic models are as effective as their DNN counterparts. However, on token-error-rates DNN…
We present the mixture-of-parents maximum entropy Markov model (MoP-MEMM), a class of directed graphical models extending MEMMs. The MoP-MEMM allows tractable incorporation of long-range dependencies between nodes by restricting the conditional distribution of each node to be a mixture of distributions given the parent…
Generative Cross-Entropy improves classification with fewer labels.
problem Limited sample efficiency of cross-entropy loss in data-scarce scenarios.
method Proposes Generative Cross-Entropy (GenCE), a new loss function that incorporates generative principles into a standard discriminative network.
result Generative Cross-Entropy outperforms traditional cross-entropy loss across various datasets and conditions.
We consider the problem of group testing with sum observations and noiseless answers, in which we aim to locate multiple objects by querying the number of objects in each of a sequence of chosen sets. We study a probabilistic setting with entropy loss, in which we assume a joint Bayesian prior density on the locations …