A simple model explains inference scaling in neural models.
problem Understanding how model performance improves with repeated inference attempts.
method A statistical ansatz based on memorization to study inference scaling laws.
result Inference loss exhibits a power law decay with increasing trials.
Natural image statistics exhibit hierarchical dependencies across multiple scales. Representing such prior knowledge in non-factorial latent tree models can boost performance of image denoising, inpainting, deconvolution or reconstruction substantially, beyond standard factorial "sparse" methodology. We derive a large …
New method infers causal factors from large-scale data without full graph reconstruction.
problem Inferring causal variables from large-scale systems without full causal graph reconstruction.
method Supervised learning on simulated data using a neural network and subsampled-ensemble inference.
result Efficiently identifies causal relationships in large-scale gene regulatory networks.
This work shows how approximate reward models can significantly improve inference-time scaling.
problem Improving the efficiency of inference for large language models.
method Identifying the Bellman error of approximate reward models and using Sequential Monte Carlo (SMC) for inference.
result Approximate reward models can reduce computational complexity from exponential to polynomial in T. The paper addresses the relevance problem in statistical inference.
problem The relevance problem in statistical inference from large-scale data.
method Not specified in the abstract, likely involves statistical methods and analysis of large-scale data.
result The relevance problem is a long-neglected topic in statistical inference.
Guarantees convergence for black-box variational inference without modifications.
problem Convergence guarantees for black-box variational inference.
method Analysis of log-smooth posterior densities, location-scale variational family, and convergence rates of algorithm design choices.
result Proximal stochastic gradient descent fixes suboptimal convergence rates and achieves strongest known guarantees.
New scaling laws optimize model size, training, and inference for better performance.
problem Trade-off between model size and inference cost in modern LLMs.
method Train-to-Test (T2) scaling laws that jointly optimize model size, training tokens, and inference samples. result Optimal pretraining decisions shift into overtraining regime, leading to stronger performance.
Bayesian model tackles high-dimensional inverse problems efficiently.
problem Estimating spatially-varying parameters in expensive models.
method Multiscale Bayesian inference with deep generative models and MCMC.
result Efficient estimation of global and local parameter features.
The paper calibrates uncertainty in dropout variational inference models.
problem Miscalibration of model uncertainty in dropout variational inference.
method Logit scaling methods are extended to recalibrate model uncertainty.
result Logit scaling reduces miscalibration, improving reliability of predictions.
New method for online inference using SGD with random scaling.
problem Efficient online inference for SGD parameters.
method Asymptotic pivotal statistics via random scaling.
result Robust and efficient online inference without resampling.
The paper proposes an efficient method to scale Bayesian inference for mixed multinomial logit models to very large datasets.
problem Efficiency in Bayesian inference for mixed multinomial logit models on large datasets.
method Amortized Variational Inference with stochastic backpropagation, automatic differentiation, and GPU acceleration.
result The proposed method achieves significant computational speedups over traditional methods for large datasets.
DriftLite improves inference quality of diffusion models without retraining.
problem Adapting pre-trained diffusion models to new target distributions without retraining.
method Lightweight, training-free particle-based approach that steers inference dynamics with optimal stability control.
result Consistently reduces variance and improves sample quality over existing methods.
Automated synthesis of inductive invariants is an important problem in software verification. Once all the invariants have been specified, software verification reduces to checking of verification conditions. Although static analyses to infer invariants have been studied for over forty years, recent years have seen a f…
Improves diffusion model performance and efficiency through classical search.
problem Tackles inference-time control in diffusion models.
method Proposes a framework combining local and global search for efficient navigation.
result Significant gains in performance and efficiency across various domains.
This paper proposes a new method for learning covers of geometric datasets to improve topological inference and visualization.
problem Improving topological inference and visualization of large-scale geometric datasets.
method Proposes a method for learning topologically-faithful covers of geometric datasets using optimization.
result Simplicial complexes obtained from learned covers outperform standard methods in terms of size and representation of large-scale topology.
Study counterfactuals in cyclic systems with shifts and scales.
problem Counterfactual inference in cyclic systems with shifts and scales.
method Shift-scale interventions in cyclic SCMs.
result Valid inference in cyclic systems with shifts and scales.
A-NeSI scales approximate inference for probabilistic neurosymbolic learning.
problem Combining neural networks with symbolic reasoning for scalable inference.
method A-NeSI: a new framework for PNL using neural networks for approximate inference.
result A-NeSI achieves scalable approximate inference without semantic changes.
New method for online inference of constrained optimization problems.
problem Online inference of constrained stochastic optimization problems.
method Random scaling of Sketched Stochastic Sequential Quadratic Programming (SSQP).
result Asymptotically valid confidence intervals and matrix-free computation.
Model uncertainty obtained by variational Bayesian inference with Monte Carlo dropout is prone to miscalibration. The uncertainty does not represent the model error well. In this paper, temperature scaling is extended to dropout variational inference to calibrate model uncertainty. Expected uncertainty calibration erro…
Study how depth affects inference in deep Bayesian neural networks.
problem Understanding how depth impacts inference in overparameterized linear Bayesian neural networks.
method Interpreting finite deep linear Bayesian neural networks as scale mixtures of Gaussian process predictors.
result Advances analytical understanding of how depth affects inference in a simple class of Bayesian neural networks.
This paper addresses metaconsistency in Bayesian inference for metastable systems.
problem Inference for metastable systems may not be consistent, but can be metaconsistent over large but finite time intervals.
method Introduces metaconsistency in a Bayesian framework, discusses its relation to spectral properties of model dynamics.
result Metaconsistency can be exploited to infer sub-systems efficiently from larger systems.
Paper presents a fast, private MH algorithm for large-scale Bayesian inference.
problem Privacy-preserving Bayesian inference for large-scale data.
method Developed a novel DP-MH algorithm using minibatches.
result First exact and fast DP MH algorithm with privacy, scalability, and efficiency trade-offs.
New method for scalable inference in large-scale regression models with complex error structures.
problem Challenges in statistical inference for large-scale regression models with dependent errors.
method Generalized Method of Wavelet Moments with Exogenous variables (GMWMX).
result Statistical validity and scalability of GMWMX for linear models with complex error structures.
Proposes a method to predict time series data using scale information.
problem Sequential data exhibits different behavior at different scales.
method Scale-based inference using k-nearest neighbors and sequential prediction. result Asymptotically, the algorithm produces the best prediction error.
New algorithm improves inference-time alignment without reward hacking.
problem Improving quality of responses from language models with limited compute.
method Inference-time alignment, focusing on extttInferenceTimePessimism algorithm. result Optimal performance and scaling-monotonicity of extttInferenceTimePessimism. Proposes two-stage robust and sparse distributed inference for large-scale data.
problem Statistical inference in large-scale, high-dimensional, and outlier-contaminated data.
method Two-stage approach: model selection with robust Lasso, fusion of local selections, and bootstrap methods for inference.
result Robust and computationally efficient inference procedures for variable selection, confidence intervals, and standard deviation approximations.
Celeste is a procedure for inferring astronomical catalogs that attains state-of-the-art scientific results. To date, Celeste has been scaled to at most hundreds of megabytes of astronomical images: Bayesian posterior inference is notoriously demanding computationally. In this paper, we report on a scalable, parallel v…
BayesDLL offers a PyTorch library for Bayesian deep learning with large models.
problem Bayesian inference for large-scale deep networks.
method Variational inference, MC-dropout, stochastic-gradient MCMC, Laplace approximation.
result BayesDLL can handle Vision Transformers and pre-trained model weights as priors.
Best-of-Majority improves inference performance in Pass@k settings.
problem Inference in difficult tasks often underperforms with single-shot selection methods.
method Combining majority voting and Best-of-N, Best-of-Majority restricts candidates to high-frequency responses.
result Best-of-Majority achieves minimax optimal regret and outperforms other methods.
New method scales inference for deep discrete models to thousands of states.
problem Traditional inference limits model capacity for deep discrete models.
method Randomized Dynamic Programming (RDP) algorithms for scaling inference.
result RDP achieves low bias and variance while reducing computation.
New method for Bayesian neural networks with unbounded weights.
problem Posterior inference for Bayesian neural networks with unbounded weights.
method Conditionally Gaussian representation for efficient posterior inference.
result Interpretable and computationally efficient procedure for posterior inference.
Learning shrinks hard tail, improving inference performance.
problem Improving inference performance in neural networks.
method Latent Instance Difficulty (LID) model analyzing fine-tuning of neural networks.
result Training-dependent inference scaling, with βexteff growing with sample size before saturating. This paper speeds up inference in large hierarchical models.
problem Inference in large hierarchical models is slow and difficult.
method Amortized variational inference with shared parameters.
result Amortized inference is as accurate as a full-rank Gaussian but much faster.
Method scales Bayesian inference to large datasets and robustifies against outliers.
problem Scalability and robustness to outliers in Bayesian inference.
method Variational inference with β-divergence and Riemannian coresets. result Efficiently constructs cleansed data summaries robust to outliers.
ADASAP accelerates GP inference for large datasets.
problem Scaling Gaussian process inference to large datasets.
method Approximate sketch-and-project algorithm for solving linear systems.
result Sketch-and-project rapidly converges to true posterior mean.
New algorithm speeds up large-scale statistical inference.
problem Efficiently solving large-scale mean-field variational inference problems.
method Developed a novel primal-dual algorithm (PD-VI) and a block-preconditioned extension (P2D-VI) for mean-field variational inference. result PD-VI and P2D-VI achieve faster convergence and better solution quality compared to existing methods. We present a novel method for frequentist statistical inference in M-estimation problems, based on stochastic gradient descent (SGD) with a fixed step size: we demonstrate that the average of such SGD sequences can be used for statistical inference, after proper scaling. An intuitive analysis using the Ornstein-Uhlen…
The paper introduces structured variational families to improve scalability in black-box variational inference.
problem Scalability issues in black-box variational inference, especially for large datasets and hierarchical models.
method Developed structured variational families that achieve better iteration complexity of O(N) compared to full-rank families.
result Structured variational families can achieve better scaling with respect to dataset size N, improving iteration complexity from O(N^2) to O(N).
PAVI speeds up VI for large-scale studies by sharing parameterization across i.i.d. variables.
problem Challenges in Bayesian inference for large population studies with many latent parameters.
method Designing plate-amortized variational inference (PAVI) to share parameterization across i.i.d. variables.
result Significant speedup in training large-scale hierarchical variational distributions.
A new method improves SVI for high-dimensional, poorly-conditioned distributions.
problem Challenges in existing SVI methods for high-dimensional, poorly-conditioned distributions.
method Trust-region optimization approach leveraging conditional independences and second-order information.
result Superior numerical performance and better scalability in high-dimensional distributions.
For large scale on-line inference problems the update strategy is critical for performance. We derive an adaptive scan Gibbs sampler that optimizes the update frequency by selecting an optimum mini-batch size. We demonstrate performance of our adaptive batch-size Gibbs sampler by comparing it against the collapsed Gibb…
Large-scale Gaussian process inference has long faced practical challenges due to time and space complexity that is superlinear in dataset size. While sparse variational Gaussian process models are capable of learning from large-scale data, standard strategies for sparsifying the model can prevent the approximation of …
New asymptotic e-values improve inference by eliminating data-dependent scaling inefficiency.
problem Data-dependent scaling inefficiency in existing asymptotic e-values.
method Drawing on Bentkus's near-optimal concentration inequalities, introduce Bentkus-type asymptotic e-values.
result Bentkus-type asymptotic e-values consistently deliver sharper inference than existing alternatives.
New method estimates bidirectional causal effects in large-scale systems.
problem Estimating bidirectional causal effects in systems with mutual dependence and heteroskedasticity.
method Heteroskedasticity-based identification with online kernel learning and random Fourier features.
result Superior accuracy and stability compared to single equation and polynomial approximations.
Attack graphs provide compact representations of the attack paths that an attacker can follow to compromise network resources by analysing network vulnerabilities and topology. These representations are a powerful tool for security risk assessment. Bayesian inference on attack graphs enables the estimation of the risk …
New method uses hyperbolic space for faster phylogenetic tree inference.
problem Inefficient Euclidean-based phylogenetic inference in high dimensions.
method Developed novel hyperbolic extensions of sequential search algorithms and variational inference methods.
result Improved speed, scalability and performance in phylogenetic inference.
In reinforcement learning (RL), agents often operate in partially observed and uncertain environments. Model-based RL suggests that this is best achieved by learning and exploiting a probabilistic model of the world. 'Active inference' is an emerging normative framework in cognitive and computational neuroscience that …
KFT improves tensor forecasting by incorporating side information.
problem Tensor factorization weaknesses in latent factors.
method Kernel Fried Tensor (KFT) with variational inference.
result Superior performance over LightGBM and FFM.