Natural Language Inference from Multiple Premises
This work addresses a more complex inference task for natural language processing researchers, but it is incremental as it builds on standard textual entailment.
The paper tackles the problem of textual entailment from multiple premise sentences by introducing a new dataset that reduces trivial lexical inferences and focuses on everyday event knowledge, and finds that this setting is more challenging for existing neural models.
We define a novel textual entailment task that requires inference over multiple premise sentences. We present a new dataset for this task that minimizes trivial lexical inferences, emphasizes knowledge of everyday events, and presents a more challenging setting for textual entailment. We evaluate several strong neural baselines and analyze how the multiple premise task differs from standard textual entailment.