Natural Language Inference from Multiple Premises
نویسندگان
چکیده
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.
منابع مشابه
Natural Solution to FraCaS Entailment Problems
Reasoning over several premises is not a common feature of RTE systems as it usually requires deep semantic analysis. On the other hand, FraCaS is a collection of entailment problems consisting of multiple premises and covering semantically challenging phenomena. We employ the tableau theorem prover for natural language to solve the FraCaS problems in a natural way. The expressiveness of a type...
متن کاملDefeasible AceRules: A Prototype
The paper adapts the syntax of the AceRules system, an existing controlled natural language (CNL), and pipes the resulting output rules to a defeasible inference engine. With this, we can represent and reason with knowledge bases (KB) with strict and defeasible rules. For strict rules, the consequent of a strict rule must hold if the premises hold; for a defeasible rule, the consequent of a def...
متن کاملThe Probability of Conditionals : The Psychological
The two main psychological theories of the ordinary conditional were designed to account for inferences made from assumptions, but few premises in everyday life can be simply assumed true. Useful premises usually have a probability that is less than certainty. But what is the probability of the ordinary conditional and how is it determined? We argue that people use a two stage Ramsey test that ...
متن کاملFiltering semantics for counterfactuals: Bridging causal models
I argue that classical counterfactual semantics in the style of Stalnaker, Lewis, and Kratzer validates an inference pattern that is disconfirmed in natural language. The solution is to alter the algorithm we use to handle inconsistency in premise sets: rather than checking all maximally consistent fragments of a premise sets, as in Krazter’s semantics, we selectively remove some of the premise...
متن کاملNew Foundations for Imperative Logic Iv: Natural Deduction
I present sound and complete natural deduction systems for (quantified modal) imperative logic, in five steps. (1) Syntax: I introduce imperative formal languages by using the imperative operator ‘!’; e.g., if ‘A’ is a declarative sentence, ‘!A’ (“let A be true’’) is an imperative sentence. (2) Semantics: I introduce interpretations of imperative formal languages, and (building on previous work...
متن کامل