Examples
Runnable notebooks demonstrating causalatee end to end, grouped by what they cover.
Fine-tuning Task Models
-
Causality Detection
Fine-tune a transformer for the binary "does this sentence express a causal relation?" classification task.
-
Causal Candidate Extraction
Fine-tune a span-extraction model to find every cause/effect entity span in a sentence.
-
Causality Identification
Fine-tune a relation classifier: given two already-marked entity spans, does a causal relation hold between them?
Baselines
-
SDP Causality Extraction
Extract causal relations from a syntactic dependency parse's shortest path between two entities — no fine-tuning required.
Causal Graphs & Mining
-
CauseNet and CGF
Load a CauseNet sample, save it as CGF, and load it back memory-mapped, without materializing it in Python objects.
-
Mining a Causal Graph
Turn a corpus of raw documents into an aggregated causal graph with
causalatee.mining's streaming pipeline.