Next Article in Journal
Power-Optimal Control of a Stirling Engine’s Frictional Piston Motion
Next Article in Special Issue
Applying the Free Energy Principle to Complex Adaptive Systems
Previous Article in Journal
Deep Prediction Model Based on Dual Decomposition with Entropy and Frequency Statistics for Nonstationary Time Series
Previous Article in Special Issue
Permutation Entropy as a Universal Disorder Criterion: How Disorders at Different Scale Levels Are Manifestations of the Same Underlying Principle
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Entropy
Volume 24
Issue 3
10.3390/e24030361
Review Report
entropy-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

Printed Edition

A printed edition of this Special Issue is available at MDPI Books....
share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
Article
Peer-Review Record

How Active Inference Could Help Revolutionise Robotics

Entropy 2022, 24(3), 361; https://0-doi-org.brum.beds.ac.uk/10.3390/e24030361
by Lancelot Da Costa 1,2,*,†, Pablo Lanillos 3,†, Noor Sajid 2, Karl Friston 2 and Shujhat Khan 4
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Entropy 2022, 24(3), 361; https://0-doi-org.brum.beds.ac.uk/10.3390/e24030361
Submission received: 21 January 2022 / Revised: 24 February 2022 / Accepted: 28 February 2022 / Published: 2 March 2022
(This article belongs to the Special Issue Applying the Free-Energy Principle to Complex Adaptive Systems)

Round 1

Reviewer 1 Report

In this paper, the authors give a perspective on potential applications of active inference. The paper is well written and raises some interesting points. Therefore, I recommend it for publication in Entropy.

I only have a few optional remarks that the authors might consider prior to publication:
- Although I believe that the paper, for the most part, is accessible to a relatively broad audience, the introduction is quite technical. I believe that the impact of the paper might increase significantly if the authors were to include a paragraph with a less technical introduction to the basics of active inference.
- The way in which Landauer's principle is mentioned on page 4 seems wrong or misleading to me. Landauer's principle provides a minimal thermodynamic cost to computations and although I can see that, in theory, one can deduce from this that minimal complexity leads to minimal thermodynamic cost,  I do not believe that this is relevant for practical applications, as state-of-art experiments still produce several orders of magnitude more heat than that produced in Landauer's bound.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

  • I don't quite follow the novelty of this research design and proposal. So I think that this needs to be highlighted clearly early on. Referenced publications propose, in various ways and from different fields, the very same concept of applying Active Inference in robotics, so the unification is I believe, implicit. Hasn't this been discussed before?
  • There are references to equations, such as the forward Kolmogorov equation which are not presented or discussed adequately. Figure 1, visually is sufficient but it should accompany a mathematical formulation and examples.
  • Discussion is rather short and lacks depth. This is also evident in sections 2 and 3 which are rather short as well. I'm not an expert in this field, and I think that I was not given sufficient depth in this area.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Most of my suggestions were met by the authors.

Back to TopTop