The Good, the Bad, and the Ugly of Markov Boundary for Tabular Prediction
Summary
This paper evaluates the practical effectiveness of Markov boundaries for tabular prediction, finding that while theoretically optimal, current causal discovery methods fail to consistently improve predictive performance due to computational limitations and mismatched optimization goals.
View Cached Full Text
Cached at: 06/01/26, 07:18 AM
Paper page - The Good, the Bad, and the Ugly of Markov Boundary for Tabular Prediction
Source: https://huggingface.co/papers/2605.29411 Published on May 28
·
Submitted byhttps://huggingface.co/Shuwan
Shuon Jun 1
Abstract
Research examines the practical effectiveness of Markov boundaries in tabular prediction, finding that while theoretically optimal, current causal discovery methods fail to consistently improve predictive performance due to computational limitations and mismatched optimization goals.
Under standard graphical assumptions, theMarkov boundaryof a target variable is the smallest set of features that renders every other feature redundant. Once the boundary is observed, the target is conditionally independent of the rest of the table. This is a tempting object fortabular prediction, since it names exactly the columns a model should need. Yet modern regressors are still trained on the full feature set. We ask whether theMarkov boundaryis genuinely useful for prediction on SCM3K, a 3,450-task synthetic SCM benchmark with feature counts from 40 to 1000 and six SCM families, evaluated with six regressors. The answer is more nuanced than the theory suggests. Restricting a regressor to the oracle boundary often improves prediction substantially, and the improvement grows as the feature space becomes larger and sparser. But the natural pipeline of recovering the boundary withcausal discoveryand training on the recovered mask does not deliver. Existing estimators exhaust the compute budget before reaching the regime where the boundary helps most, and even where they run they rarely beat the full feature set. We trace this to three causes. Discovery optimizesstructural recoveryrather than prediction. False negatives and false positives carry sharply asymmetric predictive cost. The exact boundary is only one of many feature sets that beat all features. We then develop what these facts imply for prediction-alignedfeature selectionand for tabular models that learn to use causal structure.
View arXiv pageView PDFAdd to collection
Get this paper in your agent:
hf papers read 2605\.29411
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2605.29411 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2605.29411 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2605.29411 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
Can Tabular In-Context Learners Generalize to Biomolecular Property Prediction?
This paper investigates whether tabular in-context learning models, pretrained on synthetic causal tables, can generalize to predict biomolecular properties from limited labeled data. The authors find that these models are competitive for protein fitness regression but that representation choice is crucial for small-molecule classification.
Estimation, Prediction, and Assortment Optimization for Markov Chain Choice Models with Panel Data
This paper proposes a framework for Markov chain choice models with panel data, including estimation via novel EM algorithms that leverage partial-ordering preference information, personalized choice prediction, and assortment optimization. Experimental results on synthetic data and the sushi dataset show improvements over traditional methods.
Towards Unified and Data-Efficient Prognostics and Health Management with Tabular Foundation Models
This paper proposes a framework for applying tabular foundation models to industrial time series for prognostics and health management, demonstrating strong performance and data efficiency across multiple PHM tasks.
Why Large Language Models Fail at Tabular Prediction
A new arXiv paper systematically tests five hypotheses for why large language models fail at tabular prediction, finding that dimensionality is the decisive factor: LLM accuracy degrades as input dimension grows, unlike classical baselines that stay flat or improve.
Prediction Bottlenecks Don't Discover Causal Structure (But Here's What They Actually Do)
This paper challenges the claim that prediction bottlenecks in models like Mamba recover causal structure, demonstrating through a new benchmark that gains are largely due to confounds and robustness artifacts rather than true causal discovery.