Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

PathReasoner: Modeling Reasoning Path with Equivalent Extension for Logical Question Answering

About

Logical reasoning task has attracted great interest since it was proposed. Faced with such a task, current competitive models, even large language models (e.g., ChatGPT and PaLM 2), still perform badly. Previous promising LMs struggle in logical consistency modeling and logical structure perception. To this end, we model the logical reasoning task by transforming each logical sample into reasoning paths and propose an architecture \textbf{PathReasoner}. It addresses the task from the views of both data and model. To expand the diversity of the logical samples, we propose an atom extension strategy supported by equivalent logical formulas, to form new reasoning paths. From the model perspective, we design a stack of transformer-style blocks. In particular, we propose a path-attention module to joint model in-atom and cross-atom relations with the high-order diffusion strategy. Experiments show that PathReasoner achieves competitive performances on two logical reasoning benchmarks and great generalization abilities.

Fangzhi Xu, Qika Lin, Tianzhe Zhao, Jiawei Han, Jun Liu• 2024

Related benchmarks

TaskDatasetResultRank
Logical reasoningLogiQA (test)
Accuracy45.01
92
Logical reasoningReClor (test)
Accuracy64.1
87
Logical reasoningLogiQA (val)
Accuracy43.16
50
Logical reasoningReClor (test-e)
Accuracy80.91
23
Logical reasoningReClor (test-H)
Accuracy50.89
23
Dialogue-based Multiple-choice Question AnsweringDREAM (test)
Accuracy86.84
21
Logical reasoningReClor (val)
Accuracy (ReClor val)70.4
15
Logical reasoningZsLR zero-shot v1
Accuracy (All Test Cases)52.7
7
Logical reasoningZsLR zero-shot v4
Accuracy (Test-All)0.566
7
Logical reasoningZsLR zero-shot v5
Accuracy (Test-All)57.2
7
Showing 10 of 16 rows

Other info

Follow for update