PulseAugur
EN
LIVE 21:34:05

New research enhances AI's causal discovery and reasoning capabilities

Researchers are developing new methods to improve causal discovery, the process of inferring cause-and-effect relationships from data. One approach, CauTion, integrates large language models (LLMs) with statistical algorithms to enhance accuracy and robustness, particularly for complex graphs. Another area of focus is grounding AI planning in physical causality, moving beyond simple next-token prediction to understand real-world consequences. Additionally, studies are exploring how to ensure the reliability and consistency of causal inference methods, including those based on foundation models and continuous-time systems, to make them more trustworthy for real-world applications. AI

IMPACT Advances in causal discovery and reasoning promise more reliable and robust AI systems capable of understanding and interacting with the world.

RANK_REASON Multiple arXiv papers published on causal discovery and reasoning.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 33 sources. How we write summaries →

COVERAGE [33]

  1. arXiv cs.LG TIER_1 English(EN) · Roman Plaud, Alexandre Perez-Lebel, Antoine Saillenfest, Thomas Bonald, Marine Le Morvan, Ga\"el Varoquaux, Matthieu Labeau ·

    Tailoring Strictly Proper Scoring Rules for Downstream Tasks: An Application to Causal Inference

    arXiv:2606.03332v1 Announce Type: new Abstract: Probabilistic models are typically trained using task-agnostic objectives like log-loss, which can lead to significant errors in downstream estimation. This disconnect is especially critical in Inverse Probability Weighting (IPW) fo…

  2. arXiv cs.AI TIER_1 English(EN) · Bo Peng, Kaiwen Wu, Sirui Chen, Zhiheng Wang, Yu Qiao, Chaochao Lu ·

    CauTion: Knowing When to Trust LLMs for Ensemble Causal Discovery

    arXiv:2606.03602v1 Announce Type: cross Abstract: Causal discovery from observational data remains challenging due to the fundamental limitations of purely statistical methods, such as statistical distinguishability within equivalence classes and sensitivity to finite sample size…

  3. arXiv cs.LG TIER_1 English(EN) · Tong Zhao, Ce Guo, Wayne Luk, Emil Lupu, Ray Dipojjwal ·

    Learning Temporal Causal Structure via Smooth Differentiable Optimization

    arXiv:2606.03227v1 Announce Type: new Abstract: Causal discovery with instantaneous effects in multivariate time series is challenging, as the instantaneous structure must be acyclic. Prior methods enforce this by either separating instantaneous and lagged estimation into multi-s…

  4. arXiv cs.AI TIER_1 English(EN) · Chaochao Lu ·

    CauTion: Knowing When to Trust LLMs for Ensemble Causal Discovery

    Causal discovery from observational data remains challenging due to the fundamental limitations of purely statistical methods, such as statistical distinguishability within equivalence classes and sensitivity to finite sample sizes. While large language models (LLMs) offer a prom…

  5. Hugging Face Daily Papers TIER_1 English(EN) ·

    Tailoring Strictly Proper Scoring Rules for Downstream Tasks: An Application to Causal Inference

    Probabilistic models are typically trained using task-agnostic objectives like log-loss, which can lead to significant errors in downstream estimation. This disconnect is especially critical in Inverse Probability Weighting (IPW) for causal inference, where propensity score error…

  6. arXiv cs.LG TIER_1 English(EN) · Valentyn Melnychuk, Vahid Balazadeh, Stefan Feuerriegel, Rahul G. Krishnan ·

    Frequentist Consistency of Prior-Data Fitted Networks for Causal 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. However, it is unclear whether PFN-based causal esti…

  7. arXiv cs.LG TIER_1 English(EN) · Gongxu Luo, Boyang Sun, Kun Zhang ·

    On the Recoverability of Causal Relations from Bulk Gene Expression Data

    arXiv:2606.00568v1 Announce Type: new Abstract: Bulk gene expression profiling, which aggregates pooled RNA across cells within a biological sample, remains important in the single-cell era because it is typically less noisy, more sensitive, and more cost-effective than single-ce…

  8. arXiv cs.AI TIER_1 English(EN) · Zheng Lu, Mingqi Gao, Qinlei Xie, Wanqi Zhong, Hanwen Cui, Heng Cao, Zirui Song, Yifan Yang, Chong Luo, Bei Liu, Yiming Li ·

    Token Predictors Are Not Planners: Building Physically Grounded Causal Reasoners

    arXiv:2606.01810v1 Announce Type: new Abstract: Current benchmarks for embodied vision-language planning often favor linguistic next-token prediction over physically grounded next-state reasoning. This rewards models that mimic statistical language priors rather than track causal…

  9. arXiv cs.AI TIER_1 English(EN) · Yuzhe Zhang, Chihui Chen, Lina Yao, Chen Wang ·

    Consistency evaluation of benchmarks used for causal discovery

    arXiv:2606.01789v1 Announce Type: new Abstract: In graphical causal model, causal discovery aims to construct a causal graph based on numerical data and domain knowledge in plain text. However, the evaluation of causal discovery methods remains a challenge in the area as the prog…

  10. arXiv cs.AI TIER_1 English(EN) · Erik Jahn, Dominik Janzing ·

    Evaluating Bivariate Causal Statements Based on Mutual Compatibility

    arXiv:2606.00278v1 Announce Type: new Abstract: For many real-world systems, causal ground truth is difficult to obtain, making claims about causal effects hard to assess. We develop methods for evaluating collections of $\binom{n}{2}$ bivariate causal statements over a set of $n…

  11. arXiv cs.MA (Multiagent) TIER_1 English(EN) · Tim Gebbie ·

    A Simple Hierarchical Causality Primer

    We provide a brief primer for the idea behind formalising hierarchical causality in the context of complex systems. Here actors are not simply agents. Actors instantiate causation classes. Agents implement local dynamics in given levels or organisation in a given system. Hierarch…

  12. arXiv cs.MA (Multiagent) TIER_1 English(EN) · Tim Gebbie ·

    A Simple Hierarchical Causality Primer

    We provide a brief primer for the idea behind formalising hierarchical causality in the context of complex systems. Here actors are not simply agents. Actors instantiate causation classes. Agents implement local dynamics in given levels or organisation in a given system. Hierarch…

  13. arXiv cs.LG TIER_1 English(EN) · Zi-Rong Li, Si-Yang Liu, Tian-Zuo Wang, Han-Jia Ye ·

    TabCausal: Pretraining Across Causal Environments for Tabular Causal Discovery

    arXiv:2605.31156v1 Announce Type: new Abstract: Causal discovery aims to recover directed causal relations from observational and interventional data, providing a basis for mechanistic understanding and reliable decision-making. Causal discovery foundation models (CDFMs) seek to …

  14. arXiv cs.LG TIER_1 English(EN) · Dennis Thumm, Ruben Wiedemann, Ying Chen ·

    Towards Continuous-time Causal Foundation Models

    arXiv:2605.28880v1 Announce Type: new Abstract: Extending discrete-time causal Prior-data Fitted Networks for time series to continuous time invites writing the mechanism as a stochastic differential equation (SDE) -- but if the SDE is integrated \emph{once per observation gap}, …

  15. arXiv cs.AI TIER_1 English(EN) · Zizhen Deng, Jiaru Zhang, Rui Ding, Huang Bojun, Jinzhuo Wang, Qiang Fu, Shi Han, Dongmei Zhang ·

    Test Time Training for Supervised Causal Learning

    arXiv:2605.30015v1 Announce Type: cross Abstract: Supervised Causal Learning (SCL) has shown promise in causal discovery by framing it as a supervised learning problem. However, it suffers from significant out-of-distribution generalization challenges. We reveal three limitations…

  16. Hugging Face Daily Papers TIER_1 English(EN) ·

    Test Time Training for Supervised Causal Learning

    Supervised Causal Learning (SCL) has shown promise in causal discovery by framing it as a supervised learning problem. However, it suffers from significant out-of-distribution generalization challenges. We reveal three limitations of previous SCL practices: a significant performa…

  17. arXiv cs.LG TIER_1 English(EN) · Dongmei Zhang ·

    Test Time Training for Supervised Causal Learning

    Supervised Causal Learning (SCL) has shown promise in causal discovery by framing it as a supervised learning problem. However, it suffers from significant out-of-distribution generalization challenges. We reveal three limitations of previous SCL practices: a significant performa…

  18. arXiv cs.LG TIER_1 English(EN) · Jean Kaddour, Aengus Lynch, Qi Liu, Matt J. Kusner, Ricardo Silva ·

    Causal Machine Learning: A Survey and Open Problems

    arXiv:2206.15475v3 Announce Type: replace Abstract: Causal Machine Learning (CausalML) is an umbrella term for machine learning methods that formalize the data-generation process as a structural causal model (SCM). This perspective enables us to reason about the effects of change…

  19. arXiv cs.LG TIER_1 English(EN) · Biao Ouyang, Tengxue Zhang, Zhihao Zhuang, Yang Shu, Chenjuan Guo, Bin Yang ·

    Time Series Causal Discovery via Context-Conditioned and Causality-Augmented Pretraining

    arXiv:2605.26759v1 Announce Type: new Abstract: Causal discovery from time series is critical for many real-world applications, such as tracing the root causes of anomalies. Existing approaches typically rely on dataset-specific optimization, making it difficult to transfer their…

  20. arXiv stat.ML TIER_1 English(EN) · Matias D. Cattaneo, Jason M. Klusowski, Ruiqi Rae Yu ·

    Accuracy Limits of Causal Trees for Individualized Treatment Effects

    arXiv:2509.11381v3 Announce Type: replace-cross Abstract: Recursive decision trees are widely used to estimate heterogeneous causal treatment effects in experimental and observational studies. These methods are typically implemented using CART-type recursive partitioning, with sp…

  21. arXiv stat.ML TIER_1 English(EN) · Ankur Garg, Michael Stettler, Aaron Schein, Julius von K\"ugelgen ·

    Discrete Causal Representations from Heterogeneous Domains: A Bayesian Approach with Social Survey Applications

    arXiv:2606.06288v1 Announce Type: new Abstract: Causal representation learning aims to infer the high-level latent causal concepts that give rise to observed low-level measurements. This is particularly relevant for heterogeneous data from different environments or domains since …

  22. arXiv stat.ML TIER_1 English(EN) · Amirhossein Zare, Amirhessam Zare, Herlock Rahimi, Reza Salarikia, Mohammad Kashkooli ·

    Causal Longitudinal Prior-Fitted Networks for Counterfactual Outcome Prediction

    arXiv:2606.05797v1 Announce Type: cross Abstract: Longitudinal treatment decisions require predicting potential outcomes under future treatment sequences in the presence of time-varying confounding, heterogeneous patient dynamics, and limited domain-specific data. Existing longit…

  23. arXiv stat.ML TIER_1 English(EN) · Hazhir Aliahmadi, Irina Babayan, Greg van Anders ·

    Causal Atlases from Entropic Inference: Bayesian Networks beyond Optimal DAGs

    arXiv:2606.06440v1 Announce Type: cross Abstract: Data-driven causal relationship identification is pertinent to advancing understanding of complex systems both within and beyond science. Bayesian networks offer a probabilistic method for modelling generic causal relationships vi…

  24. arXiv stat.ML TIER_1 English(EN) · Matteo Tusoni, Giuseppe Masi, Andrea Coletta, Aldo Glielmo, Viviana Arrigoni, Novella Bartolini ·

    Robust Causal Discovery in Real-World Time Series with Power-Laws

    arXiv:2507.12257v4 Announce Type: replace-cross Abstract: Exploring causal relationships in stochastic time series is a challenging yet crucial task with a vast range of applications, including finance, economics, neuroscience, and climate science. Many algorithms for Causal Disc…

  25. arXiv stat.ML TIER_1 English(EN) · Greg van Anders ·

    Causal Atlases from Entropic Inference: Bayesian Networks beyond Optimal DAGs

    Data-driven causal relationship identification is pertinent to advancing understanding of complex systems both within and beyond science. Bayesian networks offer a probabilistic method for modelling generic causal relationships via directed acyclic graphs (DAGs). However, typical…

  26. arXiv stat.ML TIER_1 English(EN) · Julius von Kügelgen ·

    Discrete Causal Representations from Heterogeneous Domains: A Bayesian Approach with Social Survey Applications

    Causal representation learning aims to infer the high-level latent causal concepts that give rise to observed low-level measurements. This is particularly relevant for heterogeneous data from different environments or domains since distribution shifts often arise through sparse, …

  27. arXiv stat.ML TIER_1 English(EN) · Mohammad Kashkooli ·

    Causal Longitudinal Prior-Fitted Networks for Counterfactual Outcome Prediction

    Longitudinal treatment decisions require predicting potential outcomes under future treatment sequences in the presence of time-varying confounding, heterogeneous patient dynamics, and limited domain-specific data. Existing longitudinal causal estimators typically train a new mod…

  28. arXiv stat.ML TIER_1 English(EN) · Masahiro Tanaka ·

    Estimating Bidirectional Causal Effects with Large Scale Online Kernel Learning

    arXiv:2511.05050v3 Announce Type: replace Abstract: In this study, a scalable online kernel learning framework is proposed for estimating bidirectional causal effects in systems characterized by mutual dependence and heteroskedasticity. Traditional causal inference often focuses …

  29. arXiv stat.ML TIER_1 English(EN) · Mohammad Ali Javidian ·

    Transferring Information Across Interventions in Causal Bayesian Optimization

    arXiv:2606.01457v1 Announce Type: cross Abstract: Bayesian optimization is a popular way to optimize expensive systems, where every experiment, simulation, or intervention costs time or money. In its standard form, it treats the variables we control as plain inputs to a black box…

  30. arXiv stat.ML TIER_1 English(EN) · Mohammad Ali Javidian ·

    Transferring Information Across Interventions in Causal Bayesian Optimization

    Bayesian optimization is a popular way to optimize expensive systems, where every experiment, simulation, or intervention costs time or money. In its standard form, it treats the variables we control as plain inputs to a black box and cannot tell apart mere correlation from a rea…

  31. arXiv stat.ML TIER_1 English(EN) · Yoichi Chikahara ·

    Moment Matters: Mean and Variance Causal Graph Discovery from Heteroscedastic Observational Data

    arXiv:2602.23602v2 Announce Type: replace Abstract: Heteroscedasticity -- where the variance of a variable changes with other variables -- is pervasive in real data, and elucidating why it arises from the perspective of statistical moments is crucial in scientific knowledge disco…

  32. arXiv stat.ML TIER_1 English(EN) · Dmitry Zaytsev, Valentina Kuskova, Michael Coppedge ·

    From Causal Discovery to Dynamic Causal Inference in Neural Time Series

    arXiv:2603.20980v2 Announce Type: replace-cross Abstract: Time-varying causal models provide a powerful framework for studying dynamic scientific systems, yet most existing approaches assume that the underlying causal network is known a priori - an assumption rarely satisfied in …

  33. arXiv stat.ML TIER_1 English(EN) · Valentina Kuskova, Dmitry Zaytsev, Michael Coppedge ·

    Beyond Coefficients: Forecast-Necessity Testing for Interpretable Causal Discovery in Nonlinear Time-Series Models

    arXiv:2604.18751v1 Announce Type: cross Abstract: Nonlinear machine-learning models are increasingly used to discover causal relationships in time-series data, yet the interpretation of their outputs remains poorly understood. In particular, causal scores produced by regularized …