Integration: Entailment Checker

Haystack node for checking the entailment between a statement and a list of Documents

Stefano Fiorucci

Live Demo: Fact Checking 🎸 Rocks!   Generic badge

How it works

Entailment Checker Node

  • The node takes a list of Documents (commonly returned by a Retriever) and a statement as input.
  • Using a Natural Language Inference model, the text entailment between each text passage/Document (premise) and the statement (hypothesis) is computed. For every text passage, we get 3 scores (summing to 1): entailment, contradiction and neutral.
  • The text entailment scores are aggregated using a weighted average. The weight is the relevance score of each passage returned by the Retriever, if available. It expresses the similarity between the text passage and the statement. Now we have a summary score, so it is possible to tell if the passages confirm, are neutral or disprove the user statement.
  • Empirical consideration: if in the first N documents (N<K), there is strong evidence of entailment/contradiction (partial aggregate scores > threshold), it is better not to consider the less relevant other (K-N) documents.


pip install haystack-entailment-checker


Basic example

from haystack import Document
from haystack_entailment_checker import EntailmentChecker

ec = EntailmentChecker(
        model_name_or_path = "microsoft/deberta-v2-xlarge-mnli",
        use_gpu = False,
        entailment_contradiction_threshold = 0.5)

doc = Document("My cat is lazy")

print("My cat is very active", [doc]))
# ({'documents': [...],
# 'aggregate_entailment_info': {'contradiction': 1.0, 'neutral': 0.0, 'entailment': 0.0}}, ...)

Fact-checking pipeline (Retriever + EntailmentChecker)

from haystack import Document, Pipeline
from haystack.nodes import BM25Retriever
from haystack.document_stores import InMemoryDocumentStore
from haystack_entailment_checker import EntailmentChecker

# the knowledge base can consist of many documents
docs = [...]
ds = InMemoryDocumentStore(use_bm25=True)

retriever = BM25Retriever(document_store=ds)
ec = EntailmentChecker()

pipe = Pipeline()
pipe.add_node(component=retriever, name="Retriever", inputs=["Query"])
pipe.add_node(component=ec, name="EntailmentChecker", inputs=["Retriever"])"YOUR STATEMENT TO CHECK")