{"id":"544a1566-646a-4b28-874f-a22b982a498f","arxiv_id":"2504.12085","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"PLACID identifies causal DAGs and direct effects under unobserved confounding with possibly invalid instruments, using surrogate IVs and distance-correlation-based ARG recovery in a partially linear structural equation model.","lead":"PLACID is a new algorithm for learning causal networks among observed variables when hidden confounders exist and some instrumental variables act on multiple targets, using distance-correlation tests and moment conditions in a partially linear model. A generalist should read it because it offers geneticists a route to gene regulatory network inference without assuming linear or perfectly valid instruments.","discovery_kind":"extension","skeptic_critique":null,"referee_report":null,"author_rebuttal":null,"desk_editor":null,"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-16T12:40:09.544298+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}