- Pedro Domingos: Neuro-symbolic does not work yet [R] https://arxiv.org/abs/2205.11502 31 comments machinelearning
- On the Paradox of Learning to Reason from Data - Language models only learn a facsimile of reasoning based off of inherent statistical features https://arxiv.org/abs/2205.11502 15 comments machinelearning
Linking pages
Would you like to stay up to date with Computer science? Checkout Computer science
Weekly.
Related searches:
Search whole site: site:arxiv.org
Search title: [2205.11502] On the Paradox of Learning to Reason from Data
See how to search.