Jove
Visualize
Contact Us

Related Concept Videos

Protein Networks02:26

Protein Networks

4.6K
An organism can have thousands of different proteins, and these proteins must cooperate to ensure the health of an organism. Proteins bind to other proteins and form complexes to carry out their functions. Many proteins interact with multiple other proteins creating a complex network of protein interactions.
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
4.6K
Predicting Molecular Geometry02:27

Predicting Molecular Geometry

46.2K
VSEPR Theory for Determination of Electron Pair Geometries
46.2K
Network Covalent Solids02:18

Network Covalent Solids

16.3K
Network covalent solids contain a three-dimensional network of covalently bonded atoms as found in the crystal structures of nonmetals like diamond, graphite, silicon, and some covalent compounds, such as silicon dioxide (sand) and silicon carbide (carborundum, the abrasive on sandpaper). Many minerals have networks of covalent bonds.
To break or to melt a covalent network solid, covalent bonds must be broken. Because covalent bonds are relatively strong, covalent network solids are typically...
16.3K
Linear Equations01:27

Linear Equations

519
Linear equations form the foundation of many algebraic and real-world applications, characterized by their simplicity and utility. A linear equation is an algebraic statement in which each term is either a constant or a product of a constant and a single variable. These equations represent straight lines when plotted on a Cartesian coordinate plane, reflecting a constant rate of change between two quantities.A typical linear equation in one variable has the form: ax + b = c, where a, b, and c...
519
Linear Circuits01:17

Linear Circuits

904
A linear circuit is characterized by its output having a direct proportionality to its input, adhering to the linearity property, which encompasses the principles of homogeneity (scaling) and additivity. Homogeneity dictates that when the input, also referred to as the excitation, is multiplied by a constant factor, the output, known as the response, is correspondingly scaled by the same constant factor. For instance, if the current is multiplied by a constant 'k,' the voltage likewise...
904
System of Memory01:23

System of Memory

7.5K
Memory is categorized into three major systems: sensory memory, short-term memory (STM), and long-term memory (LTM). These systems differ in their capacity and the duration for which they can hold information. Sensory memory captures raw sensory input from the environment, holding it for just a few seconds or less. For example, on hearing a brief, loud sound, like a car horn honking, the sound seems to linger in the mind for a moment even after it stops. This is an instance of sensory memory...
7.5K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

New analytic formulas for memory and prediction functions in reservoir computers with time delays.

Chaos (Woodbury, N.Y.)·2025
Same author

The role of NMDA receptors in memory and prediction in cultured neural networks.

Journal of neural engineering·2025
Same author

From reinforcement learning to agency: Frameworks for understanding basal cognition.

Bio Systems·2023
Same author

Prediction in cultured cortical neural networks.

PNAS nexus·2023
Same author

How Well Can We Infer Selection Benefits and Mutation Rates from Allele Frequencies?

Entropy (Basel, Switzerland)·2023
Same author

On the Role of Theory and Modeling in Neuroscience.

The Journal of neuroscience : the official journal of the Society for Neuroscience·2023
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Feb 15, 2026

A Method for Investigating Age-related Differences in the Functional Connectivity of Cognitive Control Networks Associated with Dimensional Change Card Sort Performance
09:01

A Method for Investigating Age-related Differences in the Functional Connectivity of Cognitive Control Networks Associated with Dimensional Change Card Sort Performance

Published on: May 7, 2014

10.6K

Difference between memory and prediction in linear recurrent networks.

Sarah Marzen1

  • 1Department of Physics, Physics of Living Systems, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA.

Physical Review. E
|January 20, 2018
PubMed
Summary

Maximizing memory in recurrent networks does not guarantee better prediction. Networks optimized for prediction, even simple ones, can outperform memory-focused designs, suggesting distinct optimization goals.

More Related Videos

Developing Neuroimaging Phenotypes of the Default Mode Network in PTSD: Integrating the Resting State, Working Memory, and Structural Connectivity
10:43

Developing Neuroimaging Phenotypes of the Default Mode Network in PTSD: Integrating the Resting State, Working Memory, and Structural Connectivity

Published on: July 1, 2014

15.8K
A Prediction Error-driven Retrieval Procedure for Destabilizing and Rewriting Maladaptive Reward Memories in Hazardous Drinkers
08:05

A Prediction Error-driven Retrieval Procedure for Destabilizing and Rewriting Maladaptive Reward Memories in Hazardous Drinkers

Published on: January 5, 2018

10.3K

Related Experiment Videos

Last Updated: Feb 15, 2026

A Method for Investigating Age-related Differences in the Functional Connectivity of Cognitive Control Networks Associated with Dimensional Change Card Sort Performance
09:01

A Method for Investigating Age-related Differences in the Functional Connectivity of Cognitive Control Networks Associated with Dimensional Change Card Sort Performance

Published on: May 7, 2014

10.6K
Developing Neuroimaging Phenotypes of the Default Mode Network in PTSD: Integrating the Resting State, Working Memory, and Structural Connectivity
10:43

Developing Neuroimaging Phenotypes of the Default Mode Network in PTSD: Integrating the Resting State, Working Memory, and Structural Connectivity

Published on: July 1, 2014

15.8K
A Prediction Error-driven Retrieval Procedure for Destabilizing and Rewriting Maladaptive Reward Memories in Hazardous Drinkers
08:05

A Prediction Error-driven Retrieval Procedure for Destabilizing and Rewriting Maladaptive Reward Memories in Hazardous Drinkers

Published on: January 5, 2018

10.3K

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computational Neuroscience

Background:

  • Recurrent neural networks (RNNs) are often trained for input memorization, with the expectation of improved predictive capabilities.
  • The relationship between memorization and prediction in RNNs is not fully understood, potentially leading to suboptimal network designs.

Purpose of the Study:

  • To investigate the efficacy of training recurrent networks for memorization versus prediction.
  • To determine if maximizing memory capacity in RNNs correlates with enhanced predictive performance.
  • To compare the predictive performance of optimized small RNNs with theoretical upper bounds and larger, randomly initialized networks.

Main Methods:

  • Analyzing the performance of RNNs trained with different objectives (memorization vs. prediction).
  • Evaluating the predictive capacity of single-node RNNs against Wiener filter bounds.
  • Comparing the performance of optimized one-node networks with randomly generated five-node networks.

Main Results:

  • Networks trained for memorization can exhibit poor predictive abilities.
  • Single-node networks optimized for prediction approach theoretical performance limits (Wiener filters).
  • Optimized one-node networks perform comparably to randomly generated five-node networks.

Conclusions:

  • Maximizing memory capacity and predictive capacity in recurrent networks require distinct architectural and training strategies.
  • Optimizing for prediction can lead to significantly more efficient network designs, reducing required size.
  • The findings challenge the assumption that enhanced memorization directly translates to improved prediction in recurrent networks.