Results 11 to 20 of about 16,739,290 (349)
Right for the Right Reasons: Training Differentiable Models by Constraining their Explanations [PDF]
International Joint Conference on Artificial Intelligence, 2017 Neural networks are among the most accurate supervised learning methods in use today, but their opacity makes them difficult to trust in critical applications, especially when conditions in training differ from those in test.
Doshi-Velez, Finale +2 more
core +2 more sources
Unsupervised Monocular Depth Estimation with Left-Right Consistency [PDF]
Computer Vision and Pattern Recognition, 2016 Learning based methods have shown very promising results for the task of depth estimation in single images. However, most existing approaches treat depth prediction as a supervised regression problem and as a result, require vast quantities of ...
Clément Godard +2 more
semanticscholar +1 more source
Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference [PDF]
Annual Meeting of the Association for Computational Linguistics, 2019 A machine learning system can score well on a given test set by relying on heuristics that are effective for frequent example types but break down in more challenging cases.
R. Thomas McCoy +2 more
semanticscholar +1 more source
European Union Regulations on Algorithmic Decision-Making and a "Right to Explanation" [PDF]
The AI Magazine, 2016 We summarize the potential impact that the European Union’s new General Data Protection Regulation will have on the routine use of machine learning algorithms.
B. Goodman, S. Flaxman
semanticscholar +1 more source
Indian Journal of Plastic Surgery, 2020 Background Hypoplasia of thumb is the second common congenital difference of the thumb, next only to duplication. It may occur as an isolated hand difference or as a part of radial longitudinal deficiency.
G. Balakrishnan +2 more
doaj +1 more source
Gestational diabetes mellitus- right person, right treatment, right time? [PDF]
, 2017 Background: Personalised treatment that is uniquely tailored to an individual’s phenotype has become a key goal of clinical and pharmaceutical development across many, particularly chronic, diseases.
Lindsay, Robert S. +2 more
core +1 more source
Scientific Reports, 2020 Hemispheric asymmetry in hand preference for passive cutaneous perception compared to active haptic perception is not well known. A functional near-infrared spectroscopy was used to evaluate the laterality of cortical facilitation when 31 normal right ...
J. C. Ye +4 more
semanticscholar +1 more source
Epithelial to Mesenchymal Transition History: From Embryonic Development to Cancers
Biomolecules, 2021 Epithelial to mesenchymal transition (EMT) is a process that allows epithelial cells to progressively acquire a reversible mesenchymal phenotype. Here, we recount the main events in the history of EMT.
Camille Lachat +2 more
doaj +1 more source
Making deep neural networks right for the right scientific reasons by interacting with their explanations [PDF]
Nature Machine Intelligence, 2020 Deep neural networks have demonstrated excellent performances in many real-world applications. Unfortunately, they may show Clever Hans-like behaviour (making use of confounding factors within datasets) to achieve high performance.
P. Schramowski +7 more
semanticscholar +1 more source
Archives of Disease in Childhood, 2007 This article provides an overview of different types of rights to aid consideration of, and debate about, children and young people's rights in the context of paediatrics and child health. It demonstrates how children's rights may or may not differ from adult rights and the implications for practice. It shows that applying a children's rights framework
openaire +3 more sources