Timezone: »

Leap-Of-Thought: Teaching Pre-Trained Models to Systematically Reason Over Implicit Knowledge
Alon Talmor · Oyvind Tafjord · Peter Clark · Yoav Goldberg · Jonathan Berant

Wed Dec 09 08:10 AM -- 08:20 AM (PST) @ Orals & Spotlights: Continual/Meta/Misc Learning

To what extent can a neural network systematically reason over symbolic facts? Evidence suggests that large pre-trained language models (LMs) acquire some reasoning capacity, but this ability is difficult to control. Recently, it has been shown that Transformer-based models succeed in consistent reasoning over explicit symbolic facts, under a "closed-world" assumption. However, in an open-domain setup, it is desirable to tap into the vast reservoir of implicit knowledge already encoded in the parameters of pre-trained LMs. In this work, we provide a first demonstration that LMs can be trained to reliably perform systematic reasoning combining both implicit, pre-trained knowledge and explicit natural language statements.
To do this, we describe a procedure for automatically generating datasets that teach a model new reasoning skills, and demonstrate that models learn to effectively perform inference which involves implicit taxonomic and world knowledge, chaining and counting. Finally, we show that "teaching" models to reason generalizes beyond the training distribution: they successfully compose the usage of multiple reasoning skills in single examples. Our work paves a path towards open-domain systems that constantly improve by interacting with users who can instantly correct a model by adding simple natural language statements.

Author Information

Alon Talmor (Allen Institute for AI, Tel Aviv University)
Oyvind Tafjord (Allen Institute for AI)
Peter Clark (Allen Institute for AI)
Yoav Goldberg (Allen Institute for AI, Bar Ilan University)
Jonathan Berant (Tel Aviv University)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors