Real-time computing without stable states: a new framework for neural computation based on perturbations.

Wolfgang Maass, Thomas Natschläger, Henry Markram

November 2002 Neural Comput

Synopsis of Social media discussions

The discussions include references to influential researchers like Wolfgang Maass and mentions of discovering new work, which shows moderate interest and recognition of the article’s relevance. The tone indicates curiosity and acknowledgment of its contribution without strong agreement or in-depth critique, reflecting a moderate overall impact and engagement.

A
Agreement
 Neither agree nor disagree

The discussions display mixed reactions with some curiosity but no clear consensus supporting or opposing the publication.

I
Interest
 Moderate level of interest

There is a moderate level of curiosity, especially with mentions of discovering related work and recognizing classics, indicating some engagement but not deep enthusiasm.

E
Engagement
 Neutral engagement

Posts touch upon the significance of the work, like noting classic papers and recent advances, yet lack detailed analysis or critique, reflecting a surface-level engagement.

I
Impact
 Moderate level of impact

Comments suggest the research is noteworthy and potentially influential in the field, but there is no direct indication of groundbreaking impact or widespread debate.

Social Mentions

YouTube

3 Videos

Twitter

4 Posts

Blogs

4 Articles

News

3 Articles

Metrics

Video Views

72

Total Likes

2

Extended Reach

17,775

Social Features

14

Timeline: Posts about article

Top Social Media Posts

Posts referencing the article

Reservoir CPG for Real-Time Neural Computation on Loihi

Reservoir CPG for Real-Time Neural Computation on Loihi

This video demonstrates a lamprey robot simulation controlled by reservoir-based CPG on an Intel Loihi board, showcasing novel neural computation principles based on transient dynamics rather than stable states, enabling real-time processing of changing inputs.

Master Thesis

November 15, 2020

44 views

Neural Computation Using Transient Dynamics in Reservoir CPGs

Neural Computation Using Transient Dynamics in Reservoir CPGs

This video demonstrates a lamprey robot controlled via a reservoir-based CPG on a SpiNNaker board, utilizing neural computation principles based on dynamic, transient neural states rather than stable ones, enabling real-time processing of changing inputs.

Master Thesis

November 15, 2020

18 views

Reservoir CPG for Real-Time Neural Control of Robots on Skylake CPU

Reservoir CPG for Real-Time Neural Control of Robots on Skylake CPU

This simulation showcases a lamprey robot controlled by a reservoir-based CPG, executed on an Intel Skylake CPU using Nengo. The model leverages transient neural dynamics for real-time processing, inspired by recent frameworks of neural computation based on perturbations.

Master Thesis

November 15, 2020

11 views

  • Kirito (e/acc)
    @bronzeagepapi (Twitter)

    RT @drmichaellevin: After Luis F. Seoane's excellent talk at SFI, am discovering the work of Wolfgang Maass (https://t.co/62GRdKdTIn) - res…
    view full post

    March 29, 2024

    2

  • (Sichu Lu(Sichu.Lu218@proton.me)
    @lu_sichu (Twitter)

    https://t.co/BbmQIyn04n wow really
    view full post

    October 12, 2023

  • ed fernandez
    @efernandez (Twitter)

    RT @drmichaellevin: After Luis F. Seoane's excellent talk at SFI, am discovering the work of Wolfgang Maass (https://t.co/62GRdKdTIn) - res…
    view full post

    April 8, 2020

    2

  • Dan Levenstein
    @dlevenstein (Twitter)

    @neuropoetic @neuromeditate @ldallap1 The classic: https://t.co/j7YVKQwYDY The recent classic: https://t.co/G5Ur21SHMR The classic classic: https://t.co/LVaZ8THDi6
    view full post

    July 10, 2019

    2

Abstract Synopsis

  • The paper introduces a novel model for real-time neural computation that relies on the transient dynamics of high-dimensional neural circuits rather than stable internal states, allowing for processing continuously changing input streams without task-specific circuit design.
  • This model, based on principles of dynamical systems and statistical learning, uses the concept of a 'liquid state machine' where readout neurons extract relevant information from the evolving neural activity to produce stable outputs, even without internal stability.
  • Unlike traditional Turing-based models requiring discrete, sequential states, this approach emphasizes the use of high-dimensional, transient neural states, offering new insights into neural coding, experimental design, and real-time processing in biological neural systems.