Causal structure helps computers understand unnamed variables better
How Causality Bridges the Semantic Gap
Machine LearningArtificial IntelligenceComputation and Language
Summary
Sometimes computer systems have data with variables that are measured but not clearly labeled, making it hard to know what they really mean. The authors propose using the cause-and-effect relationships between variables, rather than just general human knowledge, to figure out what these unlabeled or hidden variables represent. They build a system called CausalBridge that discovers how variables influence each other and uses that map to assign meaningful names, even for hidden variables. This method works better than older techniques that rely only on associations and can quickly identify variable meanings even when much of the system is unknown.
What this means in practice
- •For robotics engineers: Use causal discovery to automatically name sensor and control variables when designing or debugging robotic systems.
- •For data integration teams: Automatically align and name variables from different datasets by inferring causal relations, improving data merging without extensive documentation.$Commercial implications: Enables new data integration software that reduces manual labeling efforts for large-scale enterprise datasets.
Authors
Shuhao Zhang, Xuran Zhou, Han Guo, Pengtao Xie, Yujia Zheng
Abstract
Numerical measurements capture how a system behaves, but often leave the meanings of its variables unspecified. Some variables are measured but never labeled, and others are never measured at all. Existing methods assign semantics to such variables by consulting general human knowledge, but this inherits its biases where that knowledge exists and offers nothing where it does not. We bridge this gap between measurements and their meanings with causal structure instead, reading a variable's semantics from how it acts on other variables. We formalize this as structure-constrained semantic alignment, in which the embedding of each unnamed variable is solved under the dependence relations implied by the causal graph, with the embeddings of a few known names as anchors. Accordingly, we build CausalBridge, a framework that discovers the causal graph from the measurements, latent variables included, solves for the embeddings under those relations, and expresses them as names through a language model. The causal structure reflects the mechanism that generated the measurements and is recovered from the measurements alone, which may make it the one source of information free of bias from human knowledge. We evaluate CausalBridge on five questionnaires and three robotics scenarios, with 20 to 90% of the variable names masked. It recovers the semantics of observed and latent variables more accurately than existing methods that rely on association, and its lead widens as less of the system is documented. The graph it discovers names variables as accurately as the documented one, and a new system is named in minutes and at a fraction of the cost of sampling methods. Once the semantic gap is bridged faithfully, machines can understand the world and take actions causally.