all AI news
BEARS Make Neuro-Symbolic Models Aware of their Reasoning Shortcuts
Feb. 20, 2024, 5:42 a.m. | Emanuele Marconato, Samuele Bortolotti, Emile van Krieken, Antonio Vergari, Andrea Passerini, Stefano Teso
cs.LG updates on arXiv.org arxiv.org
Abstract: Neuro-Symbolic (NeSy) predictors that conform to symbolic knowledge - encoding, e.g., safety constraints - can be affected by Reasoning Shortcuts (RSs): They learn concepts consistent with the symbolic knowledge by exploiting unintended semantics. RSs compromise reliability and generalization and, as we show in this paper, they are linked to NeSy models being overconfident about the predicted concepts. Unfortunately, the only trustworthy mitigation strategy requires collecting costly dense supervision over the concepts. Rather than attempting to …
abstract arxiv bears concepts consistent constraints cs.ai cs.lg encoding knowledge learn neuro paper reasoning reliability rss safety semantics show type
More from arxiv.org / cs.LG updates on arXiv.org
Jobs in AI, ML, Big Data
Lead Developer (AI)
@ Cere Network | San Francisco, US
Research Engineer
@ Allora Labs | Remote
Ecosystem Manager
@ Allora Labs | Remote
Founding AI Engineer, Agents
@ Occam AI | New York
AI Engineer Intern, Agents
@ Occam AI | US
AI Research Scientist
@ Vara | Berlin, Germany and Remote