scholarly journals Hippocampal Pattern Separation Supports Reinforcement Learning

2018 ◽  
Author(s):  
Ian Ballard ◽  
Anthony D. Wagner ◽  
Samuel M. McClure
Keyword(s):  
Reinforcement Learning ◽  
Functional Mri ◽  
The Other ◽  
Pattern Separation ◽  
Poison Ivy ◽  
Feature Based ◽  
Individual Features ◽  
Object Features

1ABSTRACTAnimals rely on learned associations to make decisions. Associations can be based on relationships between object features (e.g., the three-leaflets of poison ivy leaves) and outcomes (e.g., rash). More often, outcomes are linked to multidimensional states (e.g., poison ivy is green in summer but red in spring). Feature-based reinforcement learning fails when the values of individual features depend on the other features present. One solution is to assign value to multifeatural conjunctive representations. We tested if the hippocampus formed separable conjunctive representations that enabled learning of response contingencies for stimuli of the form: AB+, B-, AC-, C+. Pattern analyses on functional MRI data showed the hippocampus formed conjunctive representations that were dissociable from feature components and that these representations influenced striatal PEs. Our results establish a novel role for hippocampal pattern separation and conjunctive representation in reinforcement learning.

Independent features form integrated objects: Using a novel shape-color “conjunction task” to reconstruct memory resolution for multiple object features simultaneously

2021 ◽  
Author(s):  
Aedan Yue Li ◽  
Keisuke Fukuda ◽  
Morgan Barense
Keyword(s):  
High Resolution ◽  
Object Representation ◽  
Empirical Support ◽  
Work Memory ◽  
Spatial Location ◽  
Feature Integration ◽  
Multiple Object ◽  
Individual Features ◽  
Independent Features ◽  
Object Features

Though much progress has been made to understand feature integration, debate remains regarding how objects are represented in mind based on their constituent features. Here, we advance this debate by introducing a novel shape-color “conjunction task” to reconstruct memory resolution for multiple object features simultaneously. In a first experiment, we replicated and extended a classic change detection paradigm using our task. Replicating previous work, memory resolution for individual features was reduced when the number of objects increased, regardless of the number of to-be-remembered features. Extending previous work, we found that high resolution memory near perfect in resemblance to the target was selectively impacted by the number of to-be-remembered features. Applying a statistical model of stochastic dependence, we found evidence primarily for integration of low-resolution feature memories, but less evidence for integration of high-resolution feature memories. These results suggest a resolution trade-off, such that memory resolution for individual features can be higher when those features are represented independently compared to when those features are integrated. In a second experiment which manipulated the nature of distracting information, we examined whether object features were directly bound to each other or by virtue of shared spatial location. Feature integration was disrupted by distractors sharing visual features of target objects but not disrupted when distractors shared spatial location – suggesting that feature integration was driven by direct binding between shape and color features. Our results constrain theoretical models of object representation, providing empirical support for hierarchical representations of both integrated and independent features.

Dynamic Binding of Features and Solid Modelers

1993 ◽  
Author(s):  
Jami J. Shah ◽  
Viren Pherwani
Keyword(s):  
Geometric Modeling ◽  
Object Oriented ◽  
The Other ◽  
Dynamic Binding ◽  
Communication Architecture ◽  
X Window ◽  
Feature Based

Abstract The work described in this paper investigates the feasibility of standardizing communications between geometric modeling core systems and generic feature-based applications. Since geometric modelers differ in the functionality they provide and feature applications vary in the level of geometric operations they can support internally, a multi-layered communication architecture is proposed. The methodology is analogous to the X-Window standard for graphics. At the lowest level is a library of functions named Geo-lib, which are translated into geometric modeler specific commands. If there was to be a future dynamic interfacing standard, such as STEP-SDAI, these specific calls could be replaced by standard calls, analogous to Geo-Protocol. At the next layer is a library, called Geo-widgets, which are written entirely using Geo-lib functions. At the highest level Geo-Tools, functions used commonly by generic applications. Feature applications can choose to use the library at any level, as necessary. This multi-layered geometric toolkit creates a seamless object oriented bond between the feature application and the geometric modeling core, in such a way that either one could be replaced without requiring any changes to the other.

scholarly journals Hippocampal pattern separation supports reinforcement learning

2019 ◽  
Vol 10 (1) ◽  
Author(s):  
Ian C. Ballard ◽  
Anthony D. Wagner ◽  
Samuel M. McClure
Keyword(s):  
Reinforcement Learning ◽  
Pattern Separation

Ensemble Classification System for Scientific Chart Recognition from PDF Files

2012 ◽  
Vol 2 (4) ◽  
pp. 1-10 ◽  
Author(s):  
S. Nagarajan ◽  
V. Karthikeyani
Keyword(s):  
Feature Extraction ◽  
Classification System ◽  
Texture Features ◽  
Portable Document Format ◽  
Ensemble Classification ◽  
3 Dimensional ◽  
Feature Based ◽  
Separation Feature ◽  
Object Features ◽  
Document Format

Portable Document Format (PDF) is the most frequently used universal document format on the Internet and E-Publishing. Wide usage of PDF files has increased the need of conversion tools that convert PDF file content to text or HTML formats. A PDF converter can be categorized into two domains, namely, text recognition and graphics recognition. This paper focus on graphic recognition, especially chart type identification, which is concerned with developing algorithms that has the ability to determine the type of a given chart image from a PDF file. In the proposed system, initially an enhanced connected component and statistical feature based method is used to separate the chart region from other regions. The chart region is then analyzed and grouped as either 2-dimensional or 3-dimensional chart. After separating the graphic component from the text components, feature extraction is performed. The features can be grouped as object features, texture features and shape features. The combined feature vector is then classified using ensemble classification system. Experimental results show that the chart separation, feature extraction and ensemble classification models significantly improve the quality of chart identification.

scholarly journals Monocular mechanisms determine plaid motion coherence

1996 ◽  
Vol 13 (4) ◽  
pp. 615-626 ◽  
Author(s):  
David Alais ◽  
Maarten J. van der Smagt ◽  
Frans A. J. Verstraten ◽  
W. A. van de Grind
Keyword(s):  
Cortical Area ◽  
The Other ◽  
Coherent Motion ◽  
Spatial Period ◽  
Two Dimensional ◽  
Middle Temporal ◽  
Spatial Frequencies ◽  
Motion Coherence ◽  
Plaid Motion ◽  
Feature Based

AbstractAlthough the neural location of the plaid motion coherence process is not precisely known, the middle temporal (MT) cortical area has been proposed as a likely candidate. This claim rests largely on the neurophysiological findings showing that in response to plaid stimuli, a subgroup of cells in area MT responds to the pattern direction, whereas cells in area V1 respond only to the directions of the component gratings. In Experiment 1, we report that the coherent motion of a plaid pattern can be completely abolished following adaptation to a grating which moves in the plaid direction and has the same spatial period as the plaid features (the so-called “blobs”). Interestingly, we find this phenomenon is monocular: monocular adaptation destroys plaid coherence in the exposed eye but leaves it unaffected in the other eye. Experiment 2 demonstrates that adaptation to a purely binocular (dichoptic) grating does not affect perceived plaid coherence. These data suggest several conclusions: (1) that the mechanism determining plaid coherence responds to the motion of plaid features, (2) that the coherence mechanism is monocular, and thus (3), that it is probably located at a relatively low level in the visual system and peripherally to the binocular mechanisms commonly presumed to underlie two-dimensional (2-D) motion perception. Experiment 3 examines the spatial tuning of the monocular coherence mechanism and our results suggest it is broadly tuned with a preference for lower spatial frequencies. In Experiment 4, we examine whether perceived plaid direction is determined by the motion of the grating components or the features. Our data strongly support a feature-based model.

scholarly journals Don't Think, Just Feel the Music: Individuals with Strong Pavlovian-to-Instrumental Transfer Effects Rely Less on Model-based Reinforcement Learning

2016 ◽  
Vol 28 (7) ◽  
pp. 985-995 ◽  
Author(s):  
Miriam Sebold ◽  
Daniel J. Schad ◽  
Stephan Nebe ◽  
Maria Garbusow ◽  
Elisabeth Jünger ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Individual Variation ◽  
The Other ◽  
Transfer Effects ◽  
Behavioral Choice ◽  
Model Based ◽  
Model Free ◽  
Underlying Substrate

Behavioral choice can be characterized along two axes. One axis distinguishes reflexive, model-free systems that slowly accumulate values through experience and a model-based system that uses knowledge to reason prospectively. The second axis distinguishes Pavlovian valuation of stimuli from instrumental valuation of actions or stimulus–action pairs. This results in four values and many possible interactions between them, with important consequences for accounts of individual variation. We here explored whether individual variation along one axis was related to individual variation along the other. Specifically, we asked whether individuals' balance between model-based and model-free learning was related to their tendency to show Pavlovian interferences with instrumental decisions. In two independent samples with a total of 243 participants, Pavlovian–instrumental transfer effects were negatively correlated with the strength of model-based reasoning in a two-step task. This suggests a potential common underlying substrate predisposing individuals to both have strong Pavlovian interference and be less model-based and provides a framework within which to interpret the observation of both effects in addiction.

A multiagent reinforcement learning method based on the model inference of the other agents

2002 ◽  
Vol 33 (12) ◽  
pp. 67-76
Author(s):  
Yoichiro Matsuno ◽  
Tatsuya Yamazaki ◽  
Jun Matsuda ◽  
Shin Ishii
Keyword(s):  
Reinforcement Learning ◽  
The Other ◽  
Learning Method ◽  
Model Inference ◽  
Multiagent Reinforcement Learning
Sign in / Sign up

Export Citation Format

Share Document