Complex reasoning over text requires understanding and chaining together
free-form predicates and logical connectives. Prior work has largely tried to
do this either symbolically or with black-box transformers. We present a middle
ground between these two extremes: a compositional model reminiscent of neural
module networks that can perform chained logical reasoning. This model first
finds relevant sentences in the context and then chains them together using
neural modules. Our model gives significant performance improvements (up to
29\% relative error reduction when comfibined with a reranker) on ROPES, a
recently introduced complex reasoning dataset.

本文介绍了一种类似神经模块网络的组合模型，旨在对文本进行复杂的逻辑推理，通过寻找文章中相关的句子，并使用神经模块将它们链接起来，模型在 ROPES 上取得了显著的性能改进。

段落推理的多步推理

Multi-Step Inference for Reasoning Over Paragraphs

Logical connectives and their implications on the meaning of a natural
language sentence are a fundamental aspect of understanding. In this paper, we
investigate whether visual question answering (VQA) systems trained to answer a
question about an image, are able to answer the logical composition of multiple
such questions. When put under this \textit{Lens of Logic}, state-of-the-art
VQA models have difficulty in correctly answering these logically composed
questions. We construct an augmentation of the VQA dataset as a benchmark, with
questions containing logical compositions and linguistic transformations
(negation, disjunction, conjunction, and antonyms). We propose our {Lens of
Logic (LOL)} model which uses question-attention and logic-attention to
understand logical connectives in the question, and a novel
Fr\'echet-Compatibility Loss, which ensures that the answers of the component
questions and the composed question are consistent with the inferred logical
operation. Our model shows substantial improvement in learning logical
compositions while retaining performance on VQA. We suggest this work as a move
towards robustness by embedding logical connectives in visual understanding.

本文研究了视觉问答系统是否能够回答多个经过逻辑组合的问题，并构建了一个逻辑复合和语言转换（否定、析取、合取和反义词）的 VQA 基准库。其中提出了一种使用问题注意力和逻辑注意力的 “Lens of Logic（LOL）” 模型，并采用新颖的 Frechet-Compatibility Loss 来保证组成问题的回答与推断出的逻辑操作一致。该模型在学习逻辑组合时表现出显著的提升，同时保持 VQA 的性能，从而将逻辑连接词嵌入到视觉理解中实现了鲁棒性的提高。

VQA-LOL: 逻辑视角下的视觉问答

VQA-LOL: Visual Question Answering under the Lens of Logic

In this paper we present an extension of Peirce's existential graphs to
provide a diagrammatic representation of expressions in Quantified Equilibrium
Logic (QEL). Using this formalisation, logical connectives are replaced by
encircled regions (circles and squares) and quantified variables are
represented as "identity" lines. Although the expressive power is equivalent to
that of QEL, the new representation can be useful for illustrative or
educational purposes.

本文提出一种 Peirce 存在图的扩展，以提供 Quantified Equilibrium Logic（QEL）表达式的图表表示。使用这种形式化方法，逻辑连接词被圆圈和正方形等区域所代替，量化变量则用 “身份” 线表示。虽然表达能力和 QEL 一样，但新的表示方法可用于说明或教育目的。