arXiv:2610. 28314v1 Announce Type: cross Abstract: Prior-data fitted networks (PFNs) have recently gained popularity as an end-to-end approach to amortized Bayesian inference.
arXiv:2603. 12037v2 Announce Type: replace Abstract: Foundation models based on prior-data fitted networks (PFNs) have shown strong empirical performance in causal inference by framing the task as an in-context learning problem.
ProximalFM is a transformer‑based model that uses prior‑data fitted networks (PFNs) to perform Bayesian proximal causal inference under hidden confounding. By training on synthetic data generated from structural causal models with oracle counterfactuals, it amortizes the Bayesian operator inversion into a single forward pass, producing posterior estimates of the conditional average treatment effect (CATE). The approach consistently outperforms prior methods across various proximal regimes, especially when latent confounding is strong and proxy variables are weakly informative, and it requires no dataset‑specific tuning.
The paper investigates how well Prior-Data Fitted Networks (PFNs) approximate Bayesian posterior predictive distributions (PPDs). By decomposing deployment error into prior, context, and amortization components, the authors find that PFNs generally recover the target PPD accurately, but residual amortization errors remain in rare settings, leading to disagreement among independently pretrained models. They demonstrate that Bayesian model averaging can consistently reduce this uncertainty and improve predictions, even matching full-size TabICLv2 models.
SelectPFN is a prior‑data‑fitted network that jointly performs Bayesian variable selection and prediction for sparse nonlinear regression. In a single forward pass it outputs posterior inclusion probabilities for each variable and makes predictions using only sampled subsets, ensuring excluded variables cannot influence the result. The method achieves exact finite‑sample false discovery rate control via column‑permutation equivariance and knockoff statistics, and outperforms lasso, BART, DART, LassoNet, and other sparse neural networks on synthetic and real data with up to 500 predictors.
arXiv:2310.02423v3 Announce Type: replace Abstract: We present a new algorithm for amortized inference in sparse probabilistic graphical models (PGMs), which we call $\Delta$-amortized inference ($\D...
The paper introduces a model‑agnostic inference framework for partially identified causal effects that leverages covariate information without requiring discrete covariates or accurate conditional distribution estimates. Using duality theory for optimal transport, the method delivers uniformly valid inference in randomized experiments, is doubly robust in observational settings, achieves asymptotic unbiasedness when nuisance parameters converge semiparametrically, and allows multiplier‑bootstrap selection of covariates and models while remaining computationally efficient. Empirical applications show the approach consistently narrows identified sets and confidence intervals without imposing extra structural assumptions.
arXiv:2606. 30372v1 Announce Type: new Abstract: Quantitative research across the social and behavioral sciences depends on human subject experiments that are expensive, slow, and subject to sampling bias.
arXiv:2601. 07944v2 Announce Type: replace-cross Abstract: Since the turn of the century, approximate Bayesian inference has steadily evolved as new computational techniques have been incorporated to handle increasingly complex, large-scale predictive problems.
The paper introduces ENFR, a method for estimating the absolute evidence ζ of a data‑generating model in Bayesian inference. It learns Bayes factors relative to a flow‑matched reference and then recovers the absolute evidence, improving over existing flow‑based baselines across multiple applications. The authors also provide a joint architecture for estimating evidences of multiple models and demonstrate its performance and scalability on diverse scientific datasets.
arXiv:2610.18945v1 Announce Type: cross Abstract: Amortized Bayesian inference uses simulations to perform Bayesian estimation when likelihoods are expensive or analytically unavailable. However, lea...
The paper investigates preference elicitation under the Bradley‑Terry‑Luce model, focusing on estimating an unknown partworth vector from pairwise queries that satisfy a joint identifiability condition. It derives minimax lower bounds and shows that the canonical maximum likelihood estimator (MLE) exists, is unique, and achieves near‑optimal error rates once the sample size exceeds a design‑dependent threshold, without requiring compactness constraints or external regularizers. The analysis decomposes the estimation error into a linear stochastic term, a second‑order bias, and a higher‑order remainder, providing a unified non‑asymptotic theory for parametric utility elicitation.
The paper introduces Amortized Factor Inference Networks (AFINs), a single inference network that can generalize across varying priors, likelihoods, latent dimensions, and observation counts. AFINs represent a model as a list of typed prior and likelihood factors and use an encode–merge–decode architecture to produce a Gaussian posterior approximation, with learned parameters independent of latent dimensionality or observation count. Experiments on synthetic and real datasets demonstrate that a single trained AFIN achieves competitive posterior approximations at significantly lower test‑time cost than NUTS and per‑task variational inference, with optional importance sampling further improving accuracy.