Results 1 to 10 of about 642,833 (199)
Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022 Anomaly detection is essential for preventing hazardous outcomes for safety-critical applications like autonomous driving. Given their safety-criticality, these applications benefit from provable bounds on various errors in anomaly detection. To achieve this goal in the semi-supervised setting, we propose to provide Probably Approximately Correct (PAC)
Li, Shuo +4 more
openaire +3 more sources
PAC-NeRF: Physics Augmented Continuum Neural Radiance Fields for Geometry-Agnostic System Identification [PDF]
International Conference on Learning Representations, 2023 Existing approaches to system identification (estimating the physical parameters of an object) from videos assume known object geometries. This precludes their applicability in a vast majority of scenes where object geometries are complex or unknown.
Xuan Li +6 more
semanticscholar +1 more source
Generalization Bounds: Perspectives from Information Theory and PAC-Bayes [PDF]
arXiv.org, 2023 A fundamental question in theoretical machine learning is generalization. Over the past decades, the PAC-Bayesian approach has been established as a flexible framework to address the generalization capabilities of machine learning algorithms, and design ...
Fredrik Hellström +3 more
semanticscholar +1 more source
Online PAC-Bayes Learning [PDF]
Neural Information Processing Systems, 2022 Most PAC-Bayesian bounds hold in the batch learning setting where data is collected at once, prior to inference or prediction. This somewhat departs from many contemporary learning problems where data streams are collected and the algorithms must ...
Maxime Haddouche, Benjamin Guedj
semanticscholar +1 more source
A Theory of PAC Learnability of Partial Concept Classes [PDF]
IEEE Annual Symposium on Foundations of Computer Science, 2021 We extend the classical theory of PAC learning in a way which allows to model a rich variety of practical learning tasks where the data satisfy special properties that ease the learning process.
N. Alon +3 more
semanticscholar +1 more source
Polarization-Adjusted Convolutional (PAC) Codes: Sequential Decoding vs List Decoding [PDF]
IEEE Transactions on Vehicular Technology, 2020 In the Shannon lecture at the 2019 International Symposium on Information Theory (ISIT), Arıkan proposed to employ a one-to-one convolutional transform as a pre-coding step before the polar transform.
Mohammad Rowshan, A. Burg, E. Viterbo
semanticscholar +1 more source
List Decoding of Arıkan’s PAC Codes † [PDF]
International Symposium on Information Theory, 2020 Polar coding gives rise to the first explicit family of codes that provably achieve capacity with efficient encoding and decoding for a wide range of channels.
Hanwen Yao, Arman Fazeli, A. Vardy
semanticscholar +1 more source
BMC Infectious Diseases, 2022 Background HIV-1 DNA persists in infected cells, forming viral reservoirs. Pre-antiretroviral treatment (ART) HIV-1 DNA load was reported to predict ART success in European severely immunocompromised patients.
Desmorys Raoul Moh +9 more
doaj +1 more source
BMC Public Health, 2021 Background Pre-exposure prophylaxis (PrEP) is recommended by the WHO for HIV prevention among female sex workers (FSWs). A study conducted in 2016–2017 in Côte d’Ivoire showed that if PrEP is acceptable, FSWs also have many uncovered sexual health needs.
Valentine Becquet +11 more
doaj +1 more source
AAAI Conference on Artificial Intelligence, 2021 We study PAC learning in the presence of strategic manipulation, where data points may modify their features in certain predefined ways in order to receive a better outcome. We show that the vanilla ERM principle fails to achieve any nontrivial guarantee
Hanrui Zhang, Vincent Conitzer
semanticscholar +1 more source