Prior-Data Fitted Networks
PulseAugur coverage of Prior-Data Fitted Networks — every cluster mentioning Prior-Data Fitted Networks across labs, papers, and developer communities, ranked by signal.
- 2026-05-11 research_milestone A new paper introduces a method for uncertainty quantification in Prior-Data Fitted Networks. source
2 day(s) with sentiment data
-
New Prior-Data Fitted Networks Enhance Bayesian Variable Selection
Researchers have introduced Prior-data fitted networks (PFNs), a novel approach for Bayesian variable selection in statistical modeling. Two specific PFN variants, CD-PFN and PC-PFN, are detailed for identifying conditi…
-
New GAS-PFN model advances zero-shot traffic forecasting
Researchers have developed a new model called the Graph-Aware Spatiotemporal Prior-Data Fitted Network (GAS-PFN) designed for zero-shot traffic forecasting. This model integrates road topology into its training process,…
-
New Debiased PFNs Enhance Bayesian Inference for Semi-Parametric Targets
This paper introduces Debiased Prior-Data Fitted Networks (PFNs) to improve amortized Bayesian inference for semi-parametric targets. The authors establish theoretical foundations, showing that idealized PFNs can have s…
-
New research applies synthetic pretraining to graph-based relational models
Researchers have explored the application of Prior-Data Fitted Networks (PFNs) to graph-based relational foundation models, specifically the Griffin model. Unlike transformer-based models, PFNs have seen limited explora…
-
New TabClustPFN model advances tabular data clustering
Researchers have introduced TabClustPFN, a novel Prior-data Fitted Network designed for clustering tabular data. This new method addresses key challenges in tabular clustering, such as the absence of supervision, unknow…
-
SelectPFN advances Bayesian variable selection for sparse nonlinear regression
Researchers have developed SelectPFN, a novel Prior-Data Fitted Network (PFN) capable of performing Bayesian variable selection and prediction simultaneously for sparse nonlinear regression. This new model can identify …
-
New research reveals spectral structures within Prior-Data Fitted Networks
Two new research papers explore the internal mechanisms of Prior-Data Fitted Networks (PFNs), a type of tabular foundation model. The first paper provides mechanistic evidence that PFNs capture spectral structures relat…
-
New method improves feature acquisition policy evaluation for tabular models
Researchers have developed a new method for evaluating active feature acquisition policies using tabular foundation models. This approach addresses biases that arise from imbalanced offline data coverage, which can inco…
-
OFAG: A Unified Foundation Model for Attributed Graph Clustering
Researchers have developed OFAG, a novel foundation model designed for attributed graph clustering. This model aims to provide a single, adaptable solution that can be applied to diverse attributed graphs without requir…
-
New IQS-BO method streamlines Bayesian Optimization with learned query selection
Researchers have introduced IQS-BO, a novel approach to Bayesian Optimization that significantly reduces computational costs. Unlike traditional methods requiring repeated surrogate model refitting, IQS-BO leverages Pri…
-
LLM-generated causal priors boost inference model performance
Researchers have developed a new framework for selecting causal priors in machine learning models, specifically for amortized causal inference tasks. This framework, called closed-loop prior selection, uses large langua…
-
New study finds advanced GFMs only slightly outperform GNNs on node prediction tasks
A recent study re-evaluated nine Graph Foundation Models (GFMs) for node property prediction tasks, a common application in Graph ML used for areas like fraud detection and recommendation systems. The research found tha…
-
New $\alpha$-PFN method speeds up Bayesian optimization with learned approximations
Researchers have developed a novel method called $\alpha$-PFN to accelerate entropy search (ES) acquisition functions used in Bayesian optimization. This approach utilizes Prior-data Fitted Networks (PFNs) to learn appr…
-
Paper: Transformers can learn distributions in-context
A new paper explores the theoretical capabilities of transformers in learning distributions within context, specifically focusing on Bayesian prediction tasks. Researchers demonstrate how transformers can implement grad…
-
New methods tackle class imbalance in tabular AI models
Researchers have adapted classical techniques to address class imbalance in Prior-Data Fitted Networks (PFNs) for tabular classification. They found that thresholding performs exceptionally well due to PFNs' calibration…
-
New method enhances uncertainty quantification for PFNs
Researchers have developed a new method for uncertainty quantification in Prior-Data Fitted Networks (PFNs), which are advanced models for tabular data prediction. This novel approach, based on martingale posteriors, pr…