Improving importance estimation in covariate shift for providing accurate prediction error

2022 ◽  
pp. 116376
Author(s):  
Laura Fdez-Díaz ◽  
Sara Glez-Tomillo ◽  
Elena Montañés ◽  
José Ramón Quevedo
2020 ◽  
Vol 11 (4) ◽  
pp. 955-975
Author(s):  
Oliver Lukitsch

AbstractOrthodox neurocognitive accounts of the bodily sense of agency suggest that the experience of agency arises when action-effects are anticipated accurately. In this paper, I argue that while successful anticipation is crucial for the sense of agency, the role of unsuccessful prediction has been neglected, and that inefficacy and uncertainty are no less central to the sense of agency. I will argue that this is reflected in the phenomenology of agency, which can be characterized both as the experience of (1) efficacy and (2) effort. Specifically, the “sense of efficacy” refers to the perceptual experience of an action unfolding as anticipated. The “sense of effort”, in contrast, arises when an action has an uncertain trajectory, feels difficult, and demands the exertion of control. In this case, actions do not unfold as anticipated and require continuing adaptation if they are to be efficacious. I propose that, taken individually, the experience of efficacy and effort are insufficient for the sense of agency and that these experiences can even disrupt the sense of agency when they occur in isolation from each other. I further argue that a fully-fledged sense of agency depends on the temporally extensive process of prediction error-cancelation. This way, a comparator account can accommodate both the role of accurate prediction and prediction error and thus efficacy and effort.


Accurate prediction of a nonlinear system from limited data requires sensitivity to the variation of the system’s properties in state space. Two aspects of this variability are examined, throwing new light on the ‘limits of predictability’ as well as individual predictions. A prediction scheme which embraces the variability both of dynamics and geometry is outlined and illustrated. The paper concludes with a discussion of residual predictability, proposing a simple test to detect systematic prediction error, which indicates that further improvement in prediction accuracy is possible.


2020 ◽  
Vol 149 (9) ◽  
pp. 1755-1766 ◽  
Author(s):  
William J. Villano ◽  
A. Ross Otto ◽  
C. E. Chiemeka Ezie ◽  
Roderick Gillis ◽  
Aaron S. Heller

Author(s):  
K.S. Klen ◽  
◽  
M.K. Yaremenko ◽  
V.Ya. Zhuykov ◽  
◽  
...  

The article analyzes the influence of wind speed prediction error on the size of the controlled operation zone of the storage. The equation for calculating the power at the output of the wind generator according to the known values of wind speed is given. It is shown that when the wind speed prediction error reaches a value of 20%, the controlled operation zone of the storage disappears. The necessity of comparing prediction methods with different data discreteness to ensure the minimum possible prediction error and determining the influence of data discreteness on the error is substantiated. The equations of the "predictor-corrector" scheme for the Adams, Heming, and Milne methods are given. Newton's second interpolation formula for interpolation/extrapolation is given at the end of the data table. The average relative error of MARE was used to assess the accuracy of the prediction. It is shown that the prediction error is smaller when using data with less discreteness. It is shown that when using the Adams method with a prediction horizon of up to 30 min, within ± 34% of the average energy value, the drive can be controlled or discharged in a controlled manner. References 13, figures 2, tables 3.


2017 ◽  
Author(s):  
Mohammad Atif Faiz Afzal ◽  
Chong Cheng ◽  
Johannes Hachmann

Organic materials with a high index of refraction (RI) are attracting considerable interest due to their potential application in optic and optoelectronic devices. However, most of these applications require an RI value of 1.7 or larger, while typical carbon-based polymers only exhibit values in the range of 1.3–1.5. This paper introduces an efficient computational protocol for the accurate prediction of RI values in polymers to facilitate in silico studies that an guide the discovery and design of next-generation high-RI materials. Our protocol is based on the Lorentz-Lorenz equation and is parametrized by the polarizability and number density values of a given candidate compound. In the proposed scheme, we compute the former using first-principles electronic structure theory and the latter using an approximation based on van der Waals volumes. The critical parameter in the number density approximation is the packing fraction of the bulk polymer, for which we have devised a machine learning model. We demonstrate the performance of the proposed RI protocol by testing its predictions against the experimentally known RI values of 112 optical polymers. Our approach to combine first-principles and data modeling emerges as both a successful and highly economical path to determining the RI values for a wide range of organic polymers.


Sign in / Sign up

Export Citation Format

Share Document