Protein Fold Recognition Using Markov Logic Networks

Biba, Marenglen; Ferilli, Stefano; Esposito, Floriana

doi:10.1007/978-1-4419-6800-5_4

Cited by 1 publication

(2 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, some of our techniques also apply to models beyond traditional statistical inference models. For example, when applications [BFE11,PZLR14,MZRA15]. Often, these approaches use different statistical inference models such as factor graphs [WJ08], Markov logic networks [RD06], Bayesian networks [Pea00] as a way to integrate knowledge.…”

Section: Can We Efficiently Reason About the Provable Robustness For ...mentioning

confidence: 99%

“…In this paper, we focus on a specific type of ML pipelines that we call the Sensing-Reasoning pipelines. This specific type of pipelines provides a natural way to combine predictive models based on statistical features (e.g., a neural network) and domain knowledge, which enables it to model a range of applications from both computer science [XGA + 19, DDJ + 14, PD07, ZRC + 17] and domain sciences [BFE11,PZLR14,MZRA15].…”

Section: Sensing-reasoning Pipelines and End-to-end Robustnessmentioning

confidence: 99%

See 1 more Smart Citation

Improving Certified Robustness via Statistical Learning with Logical Reasoning

Yang

Zhao

Pei

et al. 2020

Preprint

View full text Add to dashboard Cite

As machine learning (ML) being applied to many mission-critical scenarios, certifying ML model robustness becomes increasingly important. Many previous works focuses on the robustness of independent ML and ensemble models, and can only certify a very small magnitude of the adversarial perturbation. In this paper, we take a different viewpoint and improve learning robustness by going beyond independent ML and ensemble models. We aim at promoting the generic Sensing-Reasoning machine learning pipeline which contains both the sensing (e.g. deep neural networks) and reasoning (e.g. Markov logic networks (MLN)) components enriched with domain knowledge. Can domain knowledge help improve learning robustness? Can we formally certify the end-to-end robustness of such an ML pipeline?We first theoretically analyze the computational complexity of checking the provable robustness in the reasoning component. We then derive the provable robustness bound for several concrete reasoning components. We show that for reasoning components such as MLN and a specific family of Bayesian networks it is possible to certify the robustness of the whole pipeline even with a large magnitude of perturbation which cannot be certified by existing work. Finally, we conduct extensive real-world experiments on large scale datasets to evaluate the certified robustness for Sensing-Reasoning ML pipelines.

show abstract