scholarly journals Response-based outcome predictions and confidence regulate feedback processing and learning

eLife ◽  
2021 ◽  
Vol 10 ◽  
Author(s):  
Romy Frömer ◽  
Matthew R Nassar ◽  
Rasmus Bruckner ◽  
Birgit Stürmer ◽  
Werner Sommer ◽  
...  
Keyword(s):  
Bayesian Inference ◽  
Performance Monitoring ◽  
Inference Model ◽  
Feedback Processing ◽  
Online Evaluation ◽  
Evaluation Of Performance ◽  
New Information ◽  
Current Events ◽  
Control Learning ◽  
Outcome Predictions

Influential theories emphasize the importance of predictions in learning: we learn from feedback to the extent that it is surprising, and thus conveys new information. Here we explore the hypothesis that surprise depends not only on comparing current events to past experience, but also on online evaluation of performance via internal monitoring. Specifically, we propose that people leverage insights from response-based performance monitoring – outcome predictions and confidence – to control learning from feedback. In line with predictions from a Bayesian inference model, we find that people who are better at calibrating their confidence to the precision of their outcome predictions learn more quickly. Further in line with our proposal, EEG signatures of feedback processing are sensitive to the accuracy of, and confidence in, post-response outcome predictions. Taken together, our results suggest that online predictions and confidence serve to calibrate neural error signals to improve the efficiency of learning.

scholarly journals I knew that! Response-based Outcome Predictions and Confidence Regulate Feedback Processing and Learning

2018 ◽  
Author(s):  
R. Frömer ◽  
M.R. Nassar ◽  
R. Bruckner ◽  
B. Stürmer ◽  
W. Sommer ◽  
...  
Keyword(s):  
Bayesian Inference ◽  
Performance Monitoring ◽  
Inference Model ◽  
Feedback Processing ◽  
Online Evaluation ◽  
Evaluation Of Performance ◽  
New Information ◽  
Current Events ◽  
Control Learning ◽  
Outcome Predictions

AbstractInfluential theories emphasize the importance of predictions in learning: we learn from feedback to the extent that it is surprising, and thus conveys new information. Here we explore the hypothesis that surprise depends not only on comparing current events to past experience, but also on online evaluation of performance via internal monitoring. Specifically, we propose that people leverage insights from response-based performance monitoring – outcome predictions and confidence – to control learning from feedback. In line with predictions from a Bayesian inference model, we find that people who are better at calibrating their confidence to the precision of their outcome predictions learn more quickly. Further in line with our proposal, EEG signatures of feedback processing were sensitive to the accuracy of, and confidence in, post-response outcome predictions. Taken together, our results suggest that online predictions and confidence serve to calibrate neural error signals to improve the efficiency of learning.

Impact of the Uncertainties of the RAOs of a Semisubmersible Platform on the Performance of a Motion-Based Wave Inference Method

2019 ◽  
Author(s):  
J. Mas-Soler ◽  
Pedro C. de Mello ◽  
Eduardo A. Tannuri ◽  
Alexandre N. Simos ◽  
A. Souto-Iglesias
Keyword(s):  
Bayesian Inference ◽  
Transfer Functions ◽  
Inference Model ◽  
Sea State ◽  
Statistical Errors ◽  
Regularization Parameters ◽  
Ill Posed ◽  
State Estimations ◽  
The Impact ◽  
Mathematical Process

Abstract Motion based wave inference allows the estimation of the directional sea spectrum from the measured motions of a vessel. Solving the resulting inverse problem is challenging as it is often ill-posed; as a matter of fact, statistical errors of the estimated platform response functions (RAOs) may lead to misleading estimations of the sea states as many noise values are severely amplified in the mathematical process. Hence, in order to obtain reliable estimations of the sea conditions some hypothesis must be included by means of regularization parameters. This work discusses how these errors affect the regularization parameters and the accuracy of the sea state estimations. For this purpose, a statistical quantification of the errors associated to the estimated transfer functions has been included in an expanded Bayesian inference approach. Then, the resulting statistical inference model has been verified by means of a comparison between the outputs of this approach and those obtained without considering the statistical errors in the Bayesian inference. The assessment of the impact on the accuracy of the estimations is based on the results of a dedicated model-scale experimental campaign, which includes more than 150 different test conditions.

Inter-industry wage inequality: persistent differences and turbulent equalisation

2020 ◽  
Vol 44 (4) ◽  
pp. 919-942
Author(s):  
Patrick Mokre ◽  
Miriam Rehm
Keyword(s):  
Economic Theory ◽  
Bayesian Inference ◽  
Wage Inequality ◽  
Fixed Effects ◽  
Wage Differentials ◽  
Wage Determination ◽  
Inference Model ◽  
Hierarchical Bayesian Inference ◽  
Industrial Wage ◽  
Profit Rates

Abstract The empirical stylised fact of persistent inter-industry wage differentials is an enduring challenge to economic theory. This paper applies the classical theory of ‘real competition’ to the turbulent dynamics of these inter-industrial wage differentials. Theoretically, we argue that competitive wage determination can be decomposed into equalising, dispersing and turbulently equalising factors. Empirically, we show graphically and econometrically for 31 US industries in 1987–2016 that wage differentials, like regulating profit rates, are governed by turbulent equalisation. Furthermore, we apply a fixed-effects OLS as well as a hierarchical Bayesian inference model and find that the link between regulating profit rates and wage differentials is positive, significant and robust.

Application of Bayesian Inference to Milling Force Modeling

2014 ◽  
Vol 136 (2) ◽  
Author(s):  
Jaydeep M. Karandikar ◽  
Tony L. Schmitz ◽  
Ali E. Abbas
Keyword(s):  
Bayesian Inference ◽  
Linear Regression ◽  
Probability Distributions ◽  
Linear Regression Method ◽  
Regression To The Mean ◽  
Force Coefficients ◽  
Force Modeling ◽  
New Information ◽  
Mh Algorithm ◽  
High Level

This paper describes the application of Bayesian inference to the identification of force coefficients in milling. Mechanistic cutting force coefficients have been traditionally determined by performing a linear regression to the mean force values measured over a range of feed per tooth values. This linear regression method, however, yields a deterministic result for each coefficient and requires testing at several feed per tooth values to obtain a high level of confidence in the regression analysis. Bayesian inference, on the other hand, provides a systematic and formal way of updating beliefs when new information is available while incorporating uncertainty. In this work, mean force data is used to update the prior probability distributions (initial beliefs) of force coefficients using the Metropolis-Hastings (MH) algorithm Markov chain Monte Carlo (MCMC) approach. Experiments are performed at different radial depths of cut to determine the corresponding force coefficients using both methods and the results are compared.

Overview of Bayesian Statistics

2020 ◽  
pp. 0193841X1989562
Author(s):  
David Rindskopf
Keyword(s):  
Bayesian Inference ◽  
Bayesian Statistics ◽  
Statistical Modeling ◽  
Probability Distributions ◽  
Bayes Theorem ◽  
Special Issue ◽  
Popular Approach ◽  
New Information ◽  
Basic Concepts ◽  
Phd Program

Bayesian statistics is becoming a popular approach to handling complex statistical modeling. This special issue of Evaluation Review features several Bayesian contributions. In this overview, I present the basics of Bayesian inference. Bayesian statistics is based on the principle that parameters have a distribution of beliefs about them that behave exactly like probability distributions. We can use Bayes’ Theorem to update our beliefs about values of the parameters as new information becomes available. Even better, we can make statements that frequentists do not, such as “the probability that an effect is larger than 0 is .93,” and can interpret 95% (e.g.) intervals as people naturally want, that there is a 95% probability that the parameter is in that interval. I illustrate the basic concepts of Bayesian statistics through a simple example of predicting admissions to a PhD program.

A User's Willingness to Adopt a New Information System: The Influence of the Decision-Making Improvements and Performance-Monitoring Dimensions of the System

2001 ◽  
Vol 15 (2) ◽  
pp. 61-79 ◽  
Author(s):  
F. Greg Burton ◽  
Robert A. Leitch ◽  
Brad M. Tuttle
Keyword(s):  
Information System ◽  
Private Information ◽  
Performance Monitoring ◽  
Risk Preference ◽  
Vickrey Auctions ◽  
New Information ◽  
Willingness To Adopt ◽  
Agency Issues ◽  
And Performance ◽  
Economic Considerations

We investigate the influence of economic, risk preference, and agency components on a user's willingness to adopt a new information system. Using Vickrey auctions to elicit users' utility functions, we find that economic considerations are tempered by riskseeking behavior that leads to a lowering of bids for systems with greater economic benefit. With respect to agency issues, we find that subjects value their private information in that they bid less for systems that result in large decreases to their private information. This result is consistent with the prevalence of slack-inducing contracts in the field.

Sign in / Sign up

Export Citation Format

Share Document