The Web in the Spider: Associative Learning Theory

2018 ◽  
pp. 11-36
Author(s):  
Vsevolod Kapatsinski
Keyword(s):  
Learning Theory ◽  
Hebbian Learning ◽  
Ideal Observer ◽  
Learning Mechanisms ◽  
Prior Beliefs ◽  
Belief Updating ◽  
Learning Mechanism ◽  
Systems Perspective ◽  
Biological Learning ◽  
The Brain

This chapter provides an overview of basic learning mechanisms proposed within associationist learning theory: error-driven learning, Hebbian learning, and chunking. It takes the complementary learning systems perspective, which is contrasted with a Bayesian perspective in which the learner is an ‘ideal observer’. The discussion focuses on two issues. First, what is a learning mechanism? It is argued that two brain areas implement two different learning mechanisms if they would learn different things from the same input. The available data from neuroscience suggests that the brain contains multiple learning mechanisms in this sense but each learning mechanism is domain-general in applying to many different types of input. Second, what are the sources of bias that influence what a learner acquires from a certain experience? Bayesian theorists have distinguished between inductive bias implemented in prior beliefs and channel bias implemented in the translation from input to intake and output to behaviour. Given the intake and prior beliefs, belief updating in Bayesian models is unbiased, following Bayes Theorem. However, biased belief updating may be another source of bias in biological learning mechanisms.

scholarly journals Electroencephalographic Correlates of Temporal Bayesian Belief Updating and Surprise

2020 ◽  
Author(s):  
Antonino Visalli ◽  
Mariagrazia Capizzi ◽  
Ettore Ambrosini ◽  
Bruno Kopp ◽  
Antonino Vallesi
Keyword(s):  
Event Related Potentials ◽  
Prior Beliefs ◽  
Belief Updating ◽  
Cognitive Domains ◽  
Bayesian Brain ◽  
Eeg Data ◽  
Electrode Space ◽  
Related Potentials ◽  
Brain Data ◽  
The Brain

The brain predicts the timing of forthcoming events to optimize responses to them. Temporal predictions have been formalized in terms of the hazard function, which integrates prior beliefs on the likely timing of stimulus occurrence with information conveyed by the passage of time. However, how the human brain updates prior temporal beliefs is still elusive. Here we investigated electroencephalographic (EEG) signatures associated with Bayes-optimal updating of temporal beliefs. Given that updating usually occurs in response to surprising events, we sought to disentangle EEG correlates of updating from those associated with surprise. Twenty-six participants performed a temporal foreperiod task, which comprised a subset of surprising events not eliciting updating. EEG data were analyzed through a regression-based massive approach in the electrode and source space. Distinct late positive, centro-parietally distributed, event-related potentials (ERPs) were associated with surprise and belief updating in the electrode space. While surprise modulated the commonly observed P3b, updating was associated with a later and more sustained P3b-like waveform deflection. Results from source analyses revealed that surprise encoding comprises neural activity in the cingulo-opercular network (CON). These data provide evidence that temporal predictions are computed in a Bayesian manner, and that this is reflected in P3 modulations, akin to other cognitive domains. Overall, our study revealed that analyzing P3 modulations provides an important window into the Bayesian brain. Data and scripts are shared on OSF: https://osf.io/ckqa5/?view_only=f711e6f878784d4ab94f4b14b31eef46

scholarly journals Cognitive map plasticity and imitation strategies to improve individual and social behaviors of autonomous agents

2010 ◽  
Vol 1 (1) ◽  
Author(s):  
Philippe Laroque ◽  
Nathalie Gaussier ◽  
Nicolas Cuperlier ◽  
Mathias Quoy ◽  
Philippe Gaussier
Keyword(s):  
Autonomous Agents ◽  
Hebbian Learning ◽  
Social Behaviors ◽  
Place Cells ◽  
Cognitive Map ◽  
Multi Agent System ◽  
Learning Mechanism ◽  
Global Performance ◽  
Multi Agent ◽  
The Brain

AbstractStarting from neurobiological hypotheses on the existence of place cells (PC) in the brain, the aim of this article is to show how little assumptions at both individual and social levels can lead to the emergence of non-trivial global behaviors in a multi-agent system (MAS). In particular, we show that adding a simple, hebbian learning mechanism on a cognitive map allows autonomous, situated agents to adapt themselves in a dynamically changing environment, and that even using simple agent-following strategies (driven either by similarities in the agent movement, or by individual marks - “signatures” - in agents) can dramatically improve the global performance of the MAS, in terms of survival rate of the agents. Moreover, we show that analogies can be made between such a MAS and the emergence of certain social behaviors.

Pedagogical Agents in Online Learning

2011 ◽  
pp. 699-704
Author(s):  
Kevin Carmody ◽  
Zane Berge
Keyword(s):  
Learning Theory ◽  
Cognitive Learning ◽  
Learning Theories ◽  
Pedagogical Agents ◽  
Learning Mechanisms ◽  
Learning Mechanism ◽  
Instructional Role ◽  
Online Classrooms ◽  
Processing Information ◽  
E Learning

Lack of personalization and individualized attention are common issues facing distance education designers and instructors. This is a particularly important deficiency as research has shown that personalization can increase learning greatly in comparison to nonpersonalized, information to student, linear instruction (Clark & Mayer, 2003). Advocates of personalization cite cognitive learning theory as the basis for such an approach; when humans communicate with one another they are continuously processing information, either assimilating or disregarding data and forming an understanding of the information in context of the environment and of the person with whom they are interacting. This is a natural learning mechanism that cognitive learning theories state is the foundation for all deep and lasting instruction (Hein, 1991). Through an engagement of the natural learning mechanisms, or cognitive structures, an individual should be capable of learning efficiently and form a more thorough understanding of a topic. Personalization of text through the use of informal speech and the inclusion of virtual coaches known as pedagogical agents are used as personalizing devices. These are particularly relevant options in the design of nonmoderated e-learning, as personalization is meant to fill the void where the instructor once stood. There are exclusions however, as pedagogical agents have been used in “traditional” online classrooms as well. This article focuses on the use of pedagogical agents in e-learning that: -Provides information on pedagogical agents strengths and weaknesses -Provides research relevant to pedagogical agents instructional role -Provides examples of current use -Discusses possibilities of future implementation.

scholarly journals Indexing sensory plasticity: Evidence for distinct Predictive Coding and Hebbian Learning mechanisms in the cerebral cortex

2017 ◽  
Author(s):  
Meg J Spriggs ◽  
Rachael L Sumner ◽  
Rebecca L McMillan ◽  
Rosalyn J Moran ◽  
Ian J Kirk ◽  
...  
Keyword(s):  
Hebbian Learning ◽  
Predictive Coding ◽  
Task Demands ◽  
Learning Mechanisms ◽  
Sensory Plasticity ◽  
Hebbian Plasticity ◽  
Underlying Network ◽  
Generative Mechanisms ◽  
The Brain ◽  
Hebbian Mechanisms

The Roving Mismatch Negativity (MMN), and Visual LTP paradigms are widely used as independent measures of sensory plasticity. However, the paradigms are built upon fundamentally different (and seemingly opposing) models of perceptual learning; namely, Predictive Coding (MMN) and Hebbian plasticity (LTP). The aims of the current study were to 1) compare the generative mechanisms of the MMN and visual LTP, therefore assessing whether Predictive Coding and Hebbian mechanisms co-occur in the brain, and 2) assess whether the paradigms identify similar group differences in plasticity. Forty participants were split into two groups based on the BDNF Val66Met polymorphism and were presented with both paradigms. Consistent with Predictive Coding and Hebbian predictions, Dynamic Causal Modelling revealed that the generation of the MMN modulates forward and backward connections in the underlying network, while visual LTP only modulates forward connections. Genetic differences were identified in the ERPs for both paradigms, but were only apparent in backward connections of the MMN network. These results suggest that both Predictive Coding and Hebbian mechanisms are utilized by the brain under different task demands. Additionally, both tasks provide unique insight into plasticity mechanisms, which has important implications for future studies of aberrant plasticity in clinical populations.

scholarly journals Optimistic amnesia: how online and offline processing shape belief updating and memory biases in immediate and long-term optimism biases

2021 ◽  
Author(s):  
Ziqing Yao ◽  
Xuanyi Lin ◽  
Xiaoqing Hu
Keyword(s):  
Active Role ◽  
Prior Beliefs ◽  
Belief Updating ◽  
Estimation Task ◽  
Memory Biases ◽  
Brain Potentials ◽  
Adverse Life Event ◽  
Offline Processing

Abstract When people are confronted with feedback that counters their prior beliefs, they preferentially rely on desirable rather than undesirable feedback in belief updating, i.e. an optimism bias. In two pre-registered EEG studies employing an adverse life event probability estimation task, we investigated the neurocognitive processes that support the formation and the change of optimism biases in immediate and 24 h delayed tests. We found that optimistic belief updating biases not only emerged immediately but also became significantly larger after 24 h, suggesting an active role of valence-dependent offline consolidation processes in the change of optimism biases. Participants also showed optimistic memory biases: they were less accurate in remembering undesirable than desirable feedback probabilities, with inferior memories of undesirable feedback associated with lower belief updating in the delayed test. Examining event-related brain potentials (ERPs) revealed that desirability of feedback biased initial encoding: desirable feedback elicited larger P300s than undesirable feedback, with larger P300 amplitudes predicting both higher belief updating and memory accuracies. These results suggest that desirability of feedback could bias both online and offline memory-related processes such as encoding and consolidation, with both processes contributing to the formation and change of optimism biases.

scholarly journals A nonlinear updating algorithm captures suboptimal inference in the presence of signal-dependent noise

2018 ◽  
Author(s):  
Seth W. Egger ◽  
Mehrdad Jazayeri
Keyword(s):  
Probability Distributions ◽  
Nonlinear Function ◽  
Alternative View ◽  
Time Interval ◽  
Prior Beliefs ◽  
Optimal Behavior ◽  
Multiple Measurements ◽  
The Brain ◽  
Full Probability ◽  
Sequential Measurements

AbstractBayesian models of behavior have advanced the idea that humans combine prior beliefs and sensory observations to minimize uncertainty. How the brain implements Bayes-optimal inference, however, remains poorly understood. Simple behavioral tasks suggest that the brain can flexibly represent and manipulate probability distributions. An alternative view is that brain relies on simple algorithms that can implement Bayes-optimal behavior only when the computational demands are low. To distinguish between these alternatives, we devised a task in which Bayes-optimal performance could not be matched by simple algorithms. We asked subjects to estimate and reproduce a time interval by combining prior information with one or two sequential measurements. In the domain of time, measurement noise increases with duration. This property makes the integration of multiple measurements beyond the reach of simple algorithms. We found that subjects were able to update their estimates using the second measurement but their performance was suboptimal, suggesting that they were unable to update full probability distributions. Instead, subjects’ behavior was consistent with an algorithm that predicts upcoming sensory signals, and applies a nonlinear function to errors in prediction to update estimates. These results indicate that inference strategies humans deploy may deviate from Bayes-optimal integration when the computational demands are high.

scholarly journals Brain signatures of surprise in EEG and MEG data

2020 ◽  
Author(s):  
Zahra Mousavi ◽  
Mohammad Mahdi Kiani ◽  
Hamid Aghajan
Keyword(s):  
Sensory Modality ◽  
Ideal Observer ◽  
Neural Signals ◽  
Sensory Data ◽  
Temporal Features ◽  
Recorded Data ◽  
Recent Trends ◽  
Temporal Sample ◽  
Past Experiences ◽  
The Brain

AbstractThe brain is constantly anticipating the future of sensory inputs based on past experiences. When new sensory data is different from predictions shaped by recent trends, neural signals are generated to report this surprise. Existing models for quantifying surprise are based on an ideal observer assumption operating under one of the three definitions of surprise set forth as the Shannon, Bayesian, and Confidence-corrected surprise. In this paper, we analyze both visual and auditory EEG and auditory MEG signals recorded during oddball tasks to examine which temporal components in these signals are sufficient to decode the brain’s surprise based on each of these three definitions. We found that for both recording systems the Shannon surprise is always significantly better decoded than the Bayesian surprise regardless of the sensory modality and the selected temporal features used for decoding.Author summaryA regression model is proposed for decoding the level of the brain’s surprise in response to sensory sequences using selected temporal components of recorded EEG and MEG data. Three surprise quantification definitions (Shannon, Bayesian, and Confidence-corrected surprise) are compared in offering decoding power. Four different regimes for selecting temporal samples of EEG and MEG data are used to evaluate which part of the recorded data may contain signatures that represent the brain’s surprise in terms of offering a high decoding power. We found that both the middle and late components of the EEG response offer strong decoding power for surprise while the early components are significantly weaker in decoding surprise. In the MEG response, we found that the middle components have the highest decoding power while the late components offer moderate decoding powers. When using a single temporal sample for decoding surprise, samples of the middle segment possess the highest decoding power. Shannon surprise is always better decoded than the other definitions of surprise for all the four temporal feature selection regimes. Similar superiority for Shannon surprise is observed for the EEG and MEG data across the entire range of temporal sample regimes used in our analysis.

Exploring human resource work design, learning mechanism and TQM: through the lens of activity theory and contextual learning theory

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Suchitra Ajgaonkar ◽  
Netra Neelam ◽  
Abhishek Behl ◽  
Le Trung Dao ◽  
Le Dang Lang
Keyword(s):  
Activity Theory ◽  
Learning Theory ◽  
Contextual Learning ◽  
Work Design ◽  
Learning Behaviors ◽  
Content Type ◽  
Learning Mechanism ◽  
Design Learning ◽  
Well Differentiated ◽  
Practical Implications

PurposeThis research examines the effects of the context on the relationship between work design, learning mechanism and total quality management (TQM). The exploratory study examines the differential effects in context on how human resources and their activities are strategically managed for achieving TQM. Two theoretical frameworks – activity theory and contextual learning theory – are concurrently used for analysis. Specifically, the manufacturing companies, the authors examine are (1) technology-intensive company which has bought technology from a global foreign establishment (MU1), (2) technology-intensive companies having their own technology (MU2) and (3) labor-intensive units (MU3) of varying organizational sizes.Design/methodology/approachThis case study-based research consists of 27 in-depth interviews with managers and employees of different hierarchies in each manufacturing unit. The authors interviewed them using semi-structured questions that were pre-validated by five senior HR experts from the manufacturing industry. Document analysis, multiple site visits and website content helped triangulation. The data are coded and analyzed using Dedoose software for qualitative research.FindingsActivity diagrams for each manufacturing unit provides task and interaction analysis. Within and cross-case analysis address complexity and challenges of contextual reality, influences on work design and learning mechanism. HRD executives must recognize that there may be well-differentiated learning behaviors that align with organizational strategy. The learning behaviors may not be well-differentiated and become very dynamic. This dynamism may be characterized by double loop and single-loop learning feeding into each other.Practical implicationsThis study provides substantial practical implications for HRD and other managers in the manufacturing sector.Originality/valueThe new theoretical framework adds to organizational behavior studies through multi-level and cross-contextual approach. It informs strategic combinations and interactions between internal and external context, and learning needs implicating work design and TQM.

How the positivity of new information influences belief updating in depression - The more, the better?

2021 ◽  
Author(s):  
Tobias Kube ◽  
Lukas Kirchner ◽  
Gunnar Lemmer ◽  
Julia Glombiewski
Keyword(s):  
Depressive Symptoms ◽  
Traditional Learning ◽  
False Feedback ◽  
Prior Beliefs ◽  
Belief Updating ◽  
Positive Information ◽  
New Information ◽  
New Learning ◽  
Interaction Task ◽  
Disconfirming Information

Aberrant belief updating has been linked to psychopathology, e.g., depressive symptoms. While previous research used to treat belief-confirming vs. -disconfirming information as binary concepts, the present research varied the extent to which new information deviates from prior beliefs and examined its influence on belief updating. In a false feedback task (Study 1; N = 379) and a social interaction task (Study 2; N = 292), participants received slightly positive, moderately positive or extremely positive information in relation to their prior beliefs. In both studies, new information was deemed most reliable if it was moderately positive. Yet, differences in the positivity of new information had only small effects on belief updating. In Study 1, depressive symptoms were related to difficulties in generalizing positive new learning experiences. The findings suggest that, contrary to traditional learning models, the larger the differences between prior beliefs and new information, the more beliefs are not updated.

Sign in / Sign up

Export Citation Format

Share Document