From 498db6380ad21f1007e4d0b0eb173d5a0f0e7ffb Mon Sep 17 00:00:00 2001 From: aver Date: Sun, 25 Feb 2024 15:57:55 +0000 Subject: [PATCH] typo --- _publications/marconato2024bears.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/_publications/marconato2024bears.md b/_publications/marconato2024bears.md index 5aaff30..9b984c8 100644 --- a/_publications/marconato2024bears.md +++ b/_publications/marconato2024bears.md @@ -11,7 +11,7 @@ paperurl: "https://arxiv.org/abs/2402.12240" pdf: "https://arxiv.org/pdf/2402.12240.pdf" venue: "arXiv 2024" code: "https://github.com/samuelebortolotti/bears" -excerpt: "NeSy model can suffer from reasoning shortcuts, and to make them shortcut-aware, we sprinkle a pinch of Bayes to quantify the uncertainty over the extracted concepts, showing it is correlated to the presence of reasoning shortcuts." +excerpt: "NeSy models can suffer from reasoning shortcuts, and to make them shortcut-aware, we sprinkle a pinch of Bayes to quantify the uncertainty over the extracted concepts, showing it is correlated to the presence of reasoning shortcuts." 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 avoid RSs altogether, we propose to ensure NeSy models are aware of the semantic ambiguity of the concepts they learn, thus enabling their users to identify and distrust low-quality concepts. Starting from three simple desiderata, we derive bears (BE Aware of Reasoning Shortcuts), an ensembling technique that calibrates the model's concept-level confidence without compromising prediction accuracy, thus encouraging NeSy architectures to be uncertain about concepts affected by RSs. We show empirically that bears improves RS-awareness of several state-of-the-art NeSy models, and also facilitates acquiring informative dense annotations for mitigation purposes." supplemental: bibtex: "@article{marconato2024bears,