Learning action-oriented models through active inference
- Submitting institution
-
University of Sussex
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 108674_90817
- Type
- D - Journal article
- DOI
-
10.1371/journal.pcbi.1007805
- Title of journal
- PLoS Computational Biology
- Article number
- -
- First page
- 1
- Volume
- 16
- Issue
- 4
- ISSN
- 1553-734X
- Open access status
- Compliant
- Month of publication
- April
- Year of publication
- 2020
- URL
-
https://doi.org/10.1371/journal.pcbi.1007805
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- Yes
- Number of additional authors
-
2
- Research group(s)
-
-
- Citation count
- 9
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- In this work we outline a new theory for online model-based learning in embodied cognitive agents. This work has implications for cognitive science: fundamentally challenging the over reliance on veridical and faithful models in the Bayesian Brain Hypothesis, and artificial intelligence: suggesting new approaches for sample efficient and frugal learning algorithms. The reviews of the paper were exceptional, one reviewer stating “It is one of the best computational papers on the free-energy principle I have seen.” It has garnered significant citations in a short time and was cited by Karl Friston (h-index 240). Field-weighted citation impact 6.82 (Scopus).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -