aiqbench analysis
This research introduces a novel geometric framework for causal discovery that handles latent variables using Lie bracket diagnostics to screen candidate directed effects before downstream scoring. The proposed hybrid pipeline, combining BRIDGE and Spectral Kernel Flow Matching methods with local BIC scoring, achieves high accuracy on complex nonlinear DAGs while providing explicit guarantees for data screening retention.
Latent Confounded Causal Discovery via Lie Bracket Geometry
arXiv:2606.19610v2 Announce Type: replace-cross Abstract: We study causal discovery from observational and interventional regimes when latent variables may affect the measured system. Our first algorithm, BRIDGE (Bracket Residuals for Interventional Discovery and Geometric Estimation), combines a density-ratio or transport engine with a high-recall geometric screen and passes the retained arrows to a score-based or differentiable discovery method. The main formulation and experiments use known single-node intervention targets; in that regime the screen is designed to retain candidate directed effects, while a downstream learner determines the final graph or equivalence-class representation. Our second algorithm, Spectral Kernel Flow Matching (SKFM), amortizes the response fields, summarizes residual nonclosure by a spectral visible-footprint subspace, and applies an order-dependent graph extractor. Direct extraction succeeds on calibrated chains and selected motifs, but is unstable on harder random DAGs when the order must be learned. On ten-node nonlinear random DAGs, the more reliable hybrid role of the geometry is as a candidate generator: calibrated SKFM/Bridge fields followed by local BIC scoring achieve mean directed $F_1\simeq0.86$. Sachs protein signaling provides a real-data stress test and supports a diagnostic, not fully identified, interpretation. The contribution is therefore a practical interventional screening pipeline, explicit guarantees for screen retention and residual-footprint rank under stated assumptions, and a falsifiable account of the boundary between geometric diagnostics and causal identification.
Abstract:We study causal discovery from observational and interventional regimes when latent variables may affect the measured system. Our first algorithm, BRIDGE (Bracket Residuals for Interventional Discovery and Geometric Estimation), combines a density-ratio or transport engine with a high-recall geometric screen and passes the retained arrows to a score-based or differentiable discovery method. The main formulation and experiments use known single-node intervention targets; in that regime the screen is designed to retain candidate directed effects, while a downstream learner determines the final graph or equivalence-class representation. Our second algorithm, Spectral Kernel Flow Matching (SKFM), amortizes the response fields, summarizes residual nonclosure by a spectral visible-footprint subspace, and applies an order-dependent graph extractor. Direct extraction succeeds on calibrated chains and selected motifs, but is unstable on harder random DAGs when the order must be learned. On ten-node nonlinear random DAGs, the more reliable hybrid role of the geometry is as a candidate generator: calibrated SKFM/Bridge fields followed by local BIC scoring achieve mean directed $F_1\simeq0.86$. Sachs protein signaling provides a real-data stress test and supports a diagnostic, not fully identified, interpretation. The contribution is therefore a practical interventional screening pipeline, explicit guarantees for screen retention and residual-footprint rank under stated assumptions, and a falsifiable account of the boundary between geometric diagnostics and causal identification.
Revised to remove unsupported Kan-extension interpretations and clarify scope. Lie brackets are now nonspecific closure diagnostics, not certificates of latent confounding. Theorems are narrowed to conditional screen retention, residual-footprint rank, and order-dependent acyclicity. BRIDGE/SKFM algorithms and experiments remain intact
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2606.19610 [cs.LG]
(or arXiv:2606.19610v2 [cs.LG] for this version)
Full text under the publisher's open licence · 265 words
Related coverage
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · Oct 8, 2026
FoundCause: Causal Discovery with Latent Confounders from Observational Data
arXiv · Jun 17, 2026
PIRAMID: Progress and Plans
LessWrong · Jul 27, 2026
Provable Guarantees and Efficient Learning of Structural Equation Models with Latent Confounders
arXiv · Sep 17, 2026