Timezone: »
Language models have been shown to perform remarkably well on a wide range of natural language processing tasks. In this paper, we propose a novel system that uses language models to perform multi-step logical reasoning. Our system incorporates explicit planning into its inference procedure, thus able to make more informed reasoning decisions at each step by looking ahead into their future effects. Moreover, we propose a training strategy that safeguards the planning process from being led astray by spurious features. Our full system significantly outperforms other competing methods on multiple standard datasets. When using a T5 model as its core component, our system performs competitively compared to GPT-3 despite having only about 1B parameters (i.e., 175 times smaller than GPT-3). When using GPT-3.5, it significantly outperforms chain-of-thought prompting on the challenging PrOntoQA dataset. We have conducted extensive empirical studies to demonstrate that explicit planning plays a crucial role in the system's performance.
Author Information
Hongyu Zhao (University of Chicago)
Kangrui Wang (University of Chicago)
Mo Yu (IBM Research)
Hongyuan Mei (Toyota Technological Institute at Chicago)
More from the Same Authors
-
2023 : Language Models Can Improve Event Prediction by Few-Shot Abductive Reasoning »
Xiaoming Shi · Siqiao Xue · Kangrui Wang · Fan Zhou · James Zhang · Jun Zhou · Chenhao Tan · Hongyuan Mei -
2023 Poster: Towards Coherent Image Inpainting Using Denoising Diffusion Implicit Models »
Guanhua Zhang · Jiabao Ji · Yang Zhang · Mo Yu · Tommi Jaakkola · Shiyu Chang -
2020 Poster: Neural Datalog Through Time: Informed Temporal Modeling via Logical Specification »
Hongyuan Mei · Guanghui Qin · Minjie Xu · Jason Eisner -
2019 Poster: Imputing Missing Events in Continuous-Time Event Streams »
Hongyuan Mei · Guanghui Qin · Jason Eisner -
2019 Oral: Imputing Missing Events in Continuous-Time Event Streams »
Hongyuan Mei · Guanghui Qin · Jason Eisner