Refined Kolmogorov complexity of analog, evolving and stochastic recurrent neural networks

Kolmogorov complexity measures the compressibility of real numbers. We provide a refined characterization of the hypercomputational power of analog, evolving, and stochastic neural networks based on the Kolmogorov complexity of their real weights, evolving weights, and real probabilities, respective...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Information sciences Ročník 711; s. 122104
Hlavní autoři: Cabessa, Jérémie, Strozecki, Yann
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Inc 01.09.2025
Témata:
ISSN:0020-0255
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Kolmogorov complexity measures the compressibility of real numbers. We provide a refined characterization of the hypercomputational power of analog, evolving, and stochastic neural networks based on the Kolmogorov complexity of their real weights, evolving weights, and real probabilities, respectively. First, we retrieve the infinite hierarchy of complexity classes of analog networks, defined in terms of the Kolmogorov complexity of their real weights. This hierarchy lies between the complexity classes P and P/poly. Next, using a natural identification between real numbers and infinite sequences of bits, we generalize this result to evolving networks, obtaining a similar hierarchy of complexity classes within the same bounds. Finally, we extend these results to stochastic networks that employ real probabilities as randomness, deriving a new infinite hierarchy of complexity classes situated between BPP and BPP/log⁎. Beyond providing examples of such hierarchies, we describe a generic method for constructing them based on classes of functions of increasing complexity. As a practical application, we show that the predictive capabilities of recurrent neural networks are strongly impacted by the quantization applied to their weights. Overall, these results highlight the relationship between the computational power of neural networks and the intrinsic information contained by their parameters.
AbstractList Kolmogorov complexity measures the compressibility of real numbers. We provide a refined characterization of the hypercomputational power of analog, evolving, and stochastic neural networks based on the Kolmogorov complexity of their real weights, evolving weights, and real probabilities, respectively. First, we retrieve the infinite hierarchy of complexity classes of analog networks, defined in terms of the Kolmogorov complexity of their real weights. This hierarchy lies between the complexity classes P and P/poly. Next, using a natural identification between real numbers and infinite sequences of bits, we generalize this result to evolving networks, obtaining a similar hierarchy of complexity classes within the same bounds. Finally, we extend these results to stochastic networks that employ real probabilities as randomness, deriving a new infinite hierarchy of complexity classes situated between BPP and BPP/log⁎. Beyond providing examples of such hierarchies, we describe a generic method for constructing them based on classes of functions of increasing complexity. As a practical application, we show that the predictive capabilities of recurrent neural networks are strongly impacted by the quantization applied to their weights. Overall, these results highlight the relationship between the computational power of neural networks and the intrinsic information contained by their parameters.
ArticleNumber 122104
Author Cabessa, Jérémie
Strozecki, Yann
Author_xml – sequence: 1
  givenname: Jérémie
  orcidid: 0000-0002-5394-5249
  surname: Cabessa
  fullname: Cabessa, Jérémie
  email: jeremie.cabessa@uvsq.fr
  organization: Laboratoire DAVID, UVSQ – University Paris-Saclay, 78035 Versailles, France
– sequence: 2
  givenname: Yann
  orcidid: 0000-0002-0891-3766
  surname: Strozecki
  fullname: Strozecki, Yann
  email: yann.strozecki@uvsq.fr
  organization: Laboratoire DAVID, UVSQ – University Paris-Saclay, 78035 Versailles, France
BookMark eNp9kM1KAzEUhbOoYKs-gLs8gDPepJmf4EqKf1gQRDduQiZzU1OnSUmmo317p9S1qwMHvsO934xMfPBIyCWDnAErr9e58ynnwIuccc5ATMgUgEM2NsUpmaW0BgBRleWUfLyidR5b-hy6TViFGAZqwmbb4Y_r9zRYqr3uwuqK4hC6wfnVWLQ09cF86tQ7QyOaXYzoe-pxF3U3Rv8d4lc6JydWdwkv_vKMvN_fvS0es-XLw9PidpkZXrA-w6ZqRM1kI7istW5t0cjaclGXZg6AopKsbbQodY2C2UKgBGuqxrZcSjlvYH5G2HHXxJBSRKu20W103CsG6iBErdUoRB2EqKOQkbk5MjgeNjiMKhmH3mDrxn961Qb3D_0LWLdutQ
Cites_doi 10.1007/BF02478259
10.1037/h0042519
10.1142/S0129065714500294
10.1016/j.neunet.2020.05.006
10.1109/18.605580
10.1073/pnas.2001893117
10.1006/jcss.1995.1013
10.1006/inco.1996.0062
10.1016/j.neunet.2014.09.003
10.1006/jcom.1999.0505
10.1016/j.jcss.2004.04.001
10.1162/089976698300017359
10.1016/0304-3975(94)90178-3
10.1145/235809.235811
10.1006/jcss.1999.1693
10.1142/S0129054107004772
10.1016/0893-6080(95)00095-X
10.1126/science.268.5210.545
10.1016/j.neunet.2020.03.019
10.1162/089976699300016656
10.1162/neco.1996.8.1.1
10.1016/j.tcs.2003.12.014
10.1016/j.neunet.2018.08.025
10.1016/j.jcss.2018.11.003
10.1109/TNNLS.2016.2582924
10.1016/j.neunet.2019.04.019
10.1126/science.1091277
10.1007/s11047-011-9291-8
10.1016/j.cosrev.2009.03.005
ContentType Journal Article
Copyright 2025 Elsevier Inc.
Copyright_xml – notice: 2025 Elsevier Inc.
DBID AAYXX
CITATION
DOI 10.1016/j.ins.2025.122104
DatabaseName CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
Library & Information Science
ExternalDocumentID 10_1016_j_ins_2025_122104
S0020025525002361
GroupedDBID --K
--M
--Z
-~X
.DC
.~1
0R~
1B1
1OL
1RT
1~.
1~5
29I
4.4
457
4G.
5GY
5VS
7-5
71M
8P~
9JN
9JO
AAAKF
AAAKG
AABNK
AAEDT
AAEDW
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAQXK
AARIN
AATTM
AAXKI
AAXUO
AAYFN
AAYWO
ABAOU
ABBOA
ABEFU
ABFNM
ABJNI
ABMAC
ABUCO
ABWVN
ABXDB
ACDAQ
ACGFS
ACNNM
ACRLP
ACRPL
ACZNC
ADBBV
ADEZE
ADGUI
ADJOM
ADMUD
ADNMO
ADTZH
ADVLN
AEBSH
AECPX
AEIPS
AEKER
AENEX
AFFNX
AFJKZ
AFTJW
AFXIZ
AGCQF
AGHFR
AGQPQ
AGRNS
AGUBO
AGYEJ
AHHHB
AHJVU
AHZHX
AIALX
AIEXJ
AIGVJ
AIIUN
AIKHN
AITUG
AKRWK
ALMA_UNASSIGNED_HOLDINGS
AMRAJ
ANKPU
AOUOD
APLSM
APXCP
ARUGR
ASPBG
AVWKF
AXJTR
AZFZN
BJAXD
BKOJK
BLXMC
BNPGV
CS3
DU5
EBS
EFJIC
EJD
EO8
EO9
EP2
EP3
F5P
FDB
FEDTE
FGOYB
FIRID
FNPLU
FYGXN
G-Q
GBLVA
GBOLZ
HAMUX
HLZ
HVGLF
HZ~
H~9
IHE
J1W
JJJVA
KOM
LG9
LY1
M41
MHUIS
MO0
MS~
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
Q38
R2-
RIG
ROL
RPZ
SBC
SDF
SDG
SDP
SDS
SES
SEW
SPC
SPCBC
SSB
SSD
SSH
SST
SSV
SSW
SSZ
T5K
TN5
TWZ
UHS
WH7
WUQ
XPP
YYP
ZMT
ZY4
~02
~G-
77I
9DU
AAYXX
ACLOT
ACVFH
ADCNI
AEUPX
AFPUW
AIGII
AKBMS
AKYEP
CITATION
EFKBS
EFLBG
~HD
ID FETCH-LOGICAL-c251t-eb7b4819b4298aadf5b98f2486c300e4791dba46a8e41f54e90fc7bfd29993b03
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001456228400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0020-0255
IngestDate Sat Nov 29 07:57:28 EST 2025
Sat May 24 17:06:04 EDT 2025
IsPeerReviewed true
IsScholarly true
Keywords Recurrent neural networks
Computability theory
Computational power
Echo state networks
Kolmogorov complexity
Quantization
Analog computation
Stochastic computation
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c251t-eb7b4819b4298aadf5b98f2486c300e4791dba46a8e41f54e90fc7bfd29993b03
ORCID 0000-0002-5394-5249
0000-0002-0891-3766
ParticipantIDs crossref_primary_10_1016_j_ins_2025_122104
elsevier_sciencedirect_doi_10_1016_j_ins_2025_122104
PublicationCentury 2000
PublicationDate September 2025
2025-09-00
PublicationDateYYYYMMDD 2025-09-01
PublicationDate_xml – month: 09
  year: 2025
  text: September 2025
PublicationDecade 2020
PublicationTitle Information sciences
PublicationYear 2025
Publisher Elsevier Inc
Publisher_xml – name: Elsevier Inc
References Maass, Orponen (br0220) 1998; 10
Jin, Du, Huang, Liu, Luan, Wang, Xiong (br0440) 2024
Hebb (br0100) 1949
Copeland (br0400) 2004; 317
Šíma (br0240) 2019; 116
Churchland, Sejnowski (br0010) 2016
Păun (br0320) 2000; 61
Grigoryeva, Ortega (br0390) 2018; 108
Turing (br0070) 1948
Merrill, Weiss, Goldberg, Schwartz, Smith, Yahav (br0360) 2020
Minsky, Papert (br0110) 1969
Maass (br0270) 1999
Lukoševičius, Jaeger (br0380) 2009; 3
Minsky (br0040) 1967
Pollack (br0080) 1987
Omlin, Giles (br0060) 1996; 43
Greff, Srivastava, Koutnìk, Steunebrink, Schmidhuber (br0350) 2017; 28
Siegelmann, Sontag (br0140) 1995; 50
Horne, Hush (br0050) 1996; 9
Cabessa, Tchaptchet (br0300) 2020; 126
Kilian, Siegelmann (br0150) 1996; 128
Siegelmann (br0210) 1999; 15
Siegelmann, Sontag (br0170) 1994; 131
Maass (br0280) 1996; 8
Gheorghe, Stannett (br0340) 2012; 11
McCulloch, Pitts (br0020) 1943; 5
Neumann (br0130) 1958
Balcázar, Gavaldà, Siegelmann (br0260) 1997; 43
Gholami, Kim, Dong, Yao, Mahoney, Keutzer (br0430) 2021
Maass, Markram (br0450) 2004; 69
Maass (br0290) November 27-30, 1995
Maass, Sontag (br0230) 1999; 11
Bournez, Pouly (br0420) 2021
Siegelmann (br0160) 1999
Kleene (br0030) 1956
Schmidhuber (br0120) 2015; 61
Siegelmann (br0180) 1995; 268
Cabessa, Siegelmann (br0200) 2014; 24
Jaeger, Haas (br0370) 2004; 304
Šíma (br0250) 2020; 128
Arora, Barak (br0410) 2006
Cabessa, Finkel (br0190) 2019; 101
Păun, Pérez-Jiménez, Salomaa (br0330) 2007; 18
Rosenblatt (br0090) 1958; 65
Papadimitriou, Vempala, Mitropolsky, Collins, Maass (br0310) 2020; 117
Turing (10.1016/j.ins.2025.122104_br0070) 1948
Pollack (10.1016/j.ins.2025.122104_br0080) 1987
Minsky (10.1016/j.ins.2025.122104_br0040) 1967
Lukoševičius (10.1016/j.ins.2025.122104_br0380) 2009; 3
Maass (10.1016/j.ins.2025.122104_br0450) 2004; 69
Hebb (10.1016/j.ins.2025.122104_br0100) 1949
Minsky (10.1016/j.ins.2025.122104_br0110) 1969
Merrill (10.1016/j.ins.2025.122104_br0360) 2020
Siegelmann (10.1016/j.ins.2025.122104_br0180) 1995; 268
Arora (10.1016/j.ins.2025.122104_br0410) 2006
Churchland (10.1016/j.ins.2025.122104_br0010) 2016
Neumann (10.1016/j.ins.2025.122104_br0130) 1958
Balcázar (10.1016/j.ins.2025.122104_br0260) 1997; 43
Copeland (10.1016/j.ins.2025.122104_br0400) 2004; 317
Grigoryeva (10.1016/j.ins.2025.122104_br0390) 2018; 108
Maass (10.1016/j.ins.2025.122104_br0220) 1998; 10
Schmidhuber (10.1016/j.ins.2025.122104_br0120) 2015; 61
Siegelmann (10.1016/j.ins.2025.122104_br0160) 1999
Gholami (10.1016/j.ins.2025.122104_br0430)
Gheorghe (10.1016/j.ins.2025.122104_br0340) 2012; 11
Kleene (10.1016/j.ins.2025.122104_br0030) 1956
Siegelmann (10.1016/j.ins.2025.122104_br0140) 1995; 50
Šíma (10.1016/j.ins.2025.122104_br0240) 2019; 116
Maass (10.1016/j.ins.2025.122104_br0280) 1996; 8
Cabessa (10.1016/j.ins.2025.122104_br0200) 2014; 24
Omlin (10.1016/j.ins.2025.122104_br0060) 1996; 43
Păun (10.1016/j.ins.2025.122104_br0330) 2007; 18
Bournez (10.1016/j.ins.2025.122104_br0420) 2021
Jaeger (10.1016/j.ins.2025.122104_br0370) 2004; 304
Maass (10.1016/j.ins.2025.122104_br0290) 1995
Rosenblatt (10.1016/j.ins.2025.122104_br0090) 1958; 65
Maass (10.1016/j.ins.2025.122104_br0270) 1999
Siegelmann (10.1016/j.ins.2025.122104_br0170) 1994; 131
Siegelmann (10.1016/j.ins.2025.122104_br0210) 1999; 15
Păun (10.1016/j.ins.2025.122104_br0320) 2000; 61
Jin (10.1016/j.ins.2025.122104_br0440) 2024
Kilian (10.1016/j.ins.2025.122104_br0150) 1996; 128
Maass (10.1016/j.ins.2025.122104_br0230) 1999; 11
Papadimitriou (10.1016/j.ins.2025.122104_br0310) 2020; 117
Cabessa (10.1016/j.ins.2025.122104_br0190) 2019; 101
McCulloch (10.1016/j.ins.2025.122104_br0020) 1943; 5
Horne (10.1016/j.ins.2025.122104_br0050) 1996; 9
Cabessa (10.1016/j.ins.2025.122104_br0300) 2020; 126
Šíma (10.1016/j.ins.2025.122104_br0250) 2020; 128
Greff (10.1016/j.ins.2025.122104_br0350) 2017; 28
References_xml – year: 1958
  ident: br0130
  article-title: The Computer and the Brain
– volume: 8
  start-page: 1
  year: 1996
  end-page: 40
  ident: br0280
  article-title: Lower bounds for the computational power of networks of spiking neurons
  publication-title: Neural Comput.
– volume: 131
  start-page: 331
  year: 1994
  end-page: 360
  ident: br0170
  article-title: Analog computation via neural networks
  publication-title: Theor. Comput. Sci.
– volume: 11
  start-page: 771
  year: 1999
  end-page: 782
  ident: br0230
  article-title: Analog neural nets with Gaussian or other common noise distributions cannot recognize arbitary regular languages
  publication-title: Neural Comput.
– volume: 24
  year: 2014
  ident: br0200
  article-title: The super-Turing computational power of plastic recurrent neural networks
  publication-title: Int. J. Neural Syst.
– year: 1969
  ident: br0110
  article-title: Perceptrons: An Introduction to Computational Geometry
– year: 1967
  ident: br0040
  article-title: Computation: Finite and Infinite Machines
– year: 1999
  ident: br0160
  article-title: Neural Networks and Analog Computation: Beyond the Turing Limit
– year: 1948
  ident: br0070
– volume: 108
  start-page: 495
  year: 2018
  end-page: 508
  ident: br0390
  article-title: Echo state networks are universal
  publication-title: Neural Netw.
– volume: 43
  start-page: 1175
  year: 1997
  end-page: 1183
  ident: br0260
  article-title: Computational power of neural networks: a characterization in terms of Kolmogorov complexity
  publication-title: IEEE Trans. Inf. Theory
– start-page: 443
  year: 2020
  end-page: 459
  ident: br0360
  article-title: A formal hierarchy of RNN architectures
  publication-title: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, ACL 2020, Online, July 5-10, 2020, Association for Computational Linguistics
– start-page: 3
  year: 1956
  end-page: 41
  ident: br0030
  article-title: Representation of events in nerve nets and finite automata
  publication-title: Automata Studies
– volume: 126
  start-page: 312
  year: 2020
  end-page: 334
  ident: br0300
  article-title: Automata complete computation with Hodgkin-Huxley neural networks composed of synfire rings
  publication-title: Neural Netw.
– volume: 116
  start-page: 208
  year: 2019
  end-page: 223
  ident: br0240
  article-title: Subrecursive neural networks
  publication-title: Neural Netw.
– volume: 10
  start-page: 1071
  year: 1998
  end-page: 1095
  ident: br0220
  article-title: On the effect of analog noise in discrete-time analog computations
  publication-title: Neural Comput.
– year: November 27-30, 1995
  ident: br0290
  article-title: On the computational power of noisy spiking neurons
  publication-title: Advances in Neural Information Processing Systems 8, NIPS Conference
– volume: 117
  start-page: 14464
  year: 2020
  end-page: 14472
  ident: br0310
  article-title: Brain computation by assemblies of neurons
  publication-title: Proc. Natl. Acad. Sci.
– volume: 128
  start-page: 48
  year: 1996
  end-page: 56
  ident: br0150
  article-title: The dynamic universality of sigmoidal neural networks
  publication-title: Inf. Comput.
– volume: 11
  start-page: 253
  year: 2012
  end-page: 259
  ident: br0340
  article-title: Membrane system models for super-Turing paradigms
  publication-title: Nat. Comput.
– volume: 15
  start-page: 451
  year: 1999
  end-page: 475
  ident: br0210
  article-title: Stochastic analog networks and computational complexity
  publication-title: J. Complex.
– year: 2016
  ident: br0010
  article-title: The Computational Brain
– start-page: 55
  year: 1999
  end-page: 85
  ident: br0270
  article-title: Computing with spiking neurons
  publication-title: Pulsed Neural Networks
– volume: 43
  start-page: 937
  year: 1996
  end-page: 972
  ident: br0060
  article-title: Constructing deterministic finite-state automata in recurrent neural networks
  publication-title: J. ACM
– year: 1949
  ident: br0100
  article-title: The Organization of Behavior: A Neuropsychological Theory
– volume: 61
  start-page: 108
  year: 2000
  end-page: 143
  ident: br0320
  article-title: Computing with membranes
  publication-title: J. Comput. Syst. Sci.
– volume: 28
  start-page: 2222
  year: 2017
  end-page: 2232
  ident: br0350
  article-title: LSTM: a search space odyssey
  publication-title: IEEE Trans. Neural Netw. Learn. Syst.
– volume: 65
  start-page: 386
  year: 1958
  end-page: 408
  ident: br0090
  article-title: The perceptron: a probabilistic model for information storage and organization in the brain
  publication-title: Psychol. Rev.
– volume: 5
  start-page: 115
  year: 1943
  end-page: 133
  ident: br0020
  article-title: A logical calculus of the ideas immanent in nervous activity
  publication-title: Bull. Math. Biophys.
– volume: 317
  start-page: 251
  year: 2004
  end-page: 267
  ident: br0400
  article-title: Hypercomputation: philosophical issues
  publication-title: Theor. Comput. Sci.
– volume: 268
  start-page: 545
  year: 1995
  end-page: 548
  ident: br0180
  article-title: Computation beyond the Turing limit
  publication-title: Science
– volume: 9
  start-page: 243
  year: 1996
  end-page: 252
  ident: br0050
  article-title: Bounds on the complexity of recurrent neural network implementations of finite state machines
  publication-title: Neural Netw.
– volume: 50
  start-page: 132
  year: 1995
  end-page: 150
  ident: br0140
  article-title: On the computational power of neural nets
  publication-title: J. Comput. Syst. Sci.
– volume: 304
  start-page: 78
  year: 2004
  end-page: 80
  ident: br0370
  article-title: Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication
  publication-title: Science
– volume: 101
  start-page: 86
  year: 2019
  end-page: 99
  ident: br0190
  article-title: Computational capabilities of analog and evolving neural networks over infinite input streams
  publication-title: J. Comput. Syst. Sci.
– year: 1987
  ident: br0080
  article-title: On connectionist models of natural language processing
– volume: 3
  start-page: 127
  year: 2009
  end-page: 149
  ident: br0380
  article-title: Reservoir computing approaches to recurrent neural network training
  publication-title: Comput. Sci. Rev.
– start-page: 173
  year: 2021
  end-page: 226
  ident: br0420
  article-title: A Survey on Analog Models of Computation
– volume: 128
  start-page: 199
  year: 2020
  end-page: 215
  ident: br0250
  article-title: Analog neuron hierarchy
  publication-title: Neural Netw.
– year: 2006
  ident: br0410
  article-title: Computational Complexity: A Modern Approach
– volume: 69
  start-page: 593
  year: 2004
  end-page: 616
  ident: br0450
  article-title: On the computational power of circuits of spiking neurons
  publication-title: J. Comput. Syst. Sci.
– volume: 61
  start-page: 85
  year: 2015
  end-page: 117
  ident: br0120
  article-title: Deep learning in neural networks: an overview
  publication-title: Neural Netw.
– volume: 18
  start-page: 435
  year: 2007
  end-page: 455
  ident: br0330
  article-title: Spiking neural P systems: an early survey
  publication-title: Int. J. Found. Comput. Sci.
– year: 2021
  ident: br0430
  article-title: A survey of quantization methods for efficient neural network inference
– start-page: 12186
  year: 2024
  end-page: 12215
  ident: br0440
  article-title: A comprehensive evaluation of quantization strategies for large language models
  publication-title: Findings of the Association for Computational Linguistics
– volume: 5
  start-page: 115
  year: 1943
  ident: 10.1016/j.ins.2025.122104_br0020
  article-title: A logical calculus of the ideas immanent in nervous activity
  publication-title: Bull. Math. Biophys.
  doi: 10.1007/BF02478259
– start-page: 3
  year: 1956
  ident: 10.1016/j.ins.2025.122104_br0030
  article-title: Representation of events in nerve nets and finite automata
– volume: 65
  start-page: 386
  issue: 6
  year: 1958
  ident: 10.1016/j.ins.2025.122104_br0090
  article-title: The perceptron: a probabilistic model for information storage and organization in the brain
  publication-title: Psychol. Rev.
  doi: 10.1037/h0042519
– volume: 24
  issue: 8
  year: 2014
  ident: 10.1016/j.ins.2025.122104_br0200
  article-title: The super-Turing computational power of plastic recurrent neural networks
  publication-title: Int. J. Neural Syst.
  doi: 10.1142/S0129065714500294
– year: 1948
  ident: 10.1016/j.ins.2025.122104_br0070
– year: 2006
  ident: 10.1016/j.ins.2025.122104_br0410
– start-page: 12186
  year: 2024
  ident: 10.1016/j.ins.2025.122104_br0440
  article-title: A comprehensive evaluation of quantization strategies for large language models
– volume: 128
  start-page: 199
  year: 2020
  ident: 10.1016/j.ins.2025.122104_br0250
  article-title: Analog neuron hierarchy
  publication-title: Neural Netw.
  doi: 10.1016/j.neunet.2020.05.006
– start-page: 443
  year: 2020
  ident: 10.1016/j.ins.2025.122104_br0360
  article-title: A formal hierarchy of RNN architectures
– year: 1999
  ident: 10.1016/j.ins.2025.122104_br0160
– year: 1987
  ident: 10.1016/j.ins.2025.122104_br0080
– volume: 43
  start-page: 1175
  issue: 4
  year: 1997
  ident: 10.1016/j.ins.2025.122104_br0260
  article-title: Computational power of neural networks: a characterization in terms of Kolmogorov complexity
  publication-title: IEEE Trans. Inf. Theory
  doi: 10.1109/18.605580
– volume: 117
  start-page: 14464
  issue: 25
  year: 2020
  ident: 10.1016/j.ins.2025.122104_br0310
  article-title: Brain computation by assemblies of neurons
  publication-title: Proc. Natl. Acad. Sci.
  doi: 10.1073/pnas.2001893117
– volume: 50
  start-page: 132
  issue: 1
  year: 1995
  ident: 10.1016/j.ins.2025.122104_br0140
  article-title: On the computational power of neural nets
  publication-title: J. Comput. Syst. Sci.
  doi: 10.1006/jcss.1995.1013
– volume: 128
  start-page: 48
  issue: 1
  year: 1996
  ident: 10.1016/j.ins.2025.122104_br0150
  article-title: The dynamic universality of sigmoidal neural networks
  publication-title: Inf. Comput.
  doi: 10.1006/inco.1996.0062
– volume: 61
  start-page: 85
  year: 2015
  ident: 10.1016/j.ins.2025.122104_br0120
  article-title: Deep learning in neural networks: an overview
  publication-title: Neural Netw.
  doi: 10.1016/j.neunet.2014.09.003
– volume: 15
  start-page: 451
  issue: 4
  year: 1999
  ident: 10.1016/j.ins.2025.122104_br0210
  article-title: Stochastic analog networks and computational complexity
  publication-title: J. Complex.
  doi: 10.1006/jcom.1999.0505
– volume: 69
  start-page: 593
  issue: 4
  year: 2004
  ident: 10.1016/j.ins.2025.122104_br0450
  article-title: On the computational power of circuits of spiking neurons
  publication-title: J. Comput. Syst. Sci.
  doi: 10.1016/j.jcss.2004.04.001
– volume: 10
  start-page: 1071
  issue: 5
  year: 1998
  ident: 10.1016/j.ins.2025.122104_br0220
  article-title: On the effect of analog noise in discrete-time analog computations
  publication-title: Neural Comput.
  doi: 10.1162/089976698300017359
– year: 1958
  ident: 10.1016/j.ins.2025.122104_br0130
– volume: 131
  start-page: 331
  issue: 2
  year: 1994
  ident: 10.1016/j.ins.2025.122104_br0170
  article-title: Analog computation via neural networks
  publication-title: Theor. Comput. Sci.
  doi: 10.1016/0304-3975(94)90178-3
– start-page: 55
  year: 1999
  ident: 10.1016/j.ins.2025.122104_br0270
  article-title: Computing with spiking neurons
– year: 1995
  ident: 10.1016/j.ins.2025.122104_br0290
  article-title: On the computational power of noisy spiking neurons
– year: 1949
  ident: 10.1016/j.ins.2025.122104_br0100
– volume: 43
  start-page: 937
  issue: 6
  year: 1996
  ident: 10.1016/j.ins.2025.122104_br0060
  article-title: Constructing deterministic finite-state automata in recurrent neural networks
  publication-title: J. ACM
  doi: 10.1145/235809.235811
– volume: 61
  start-page: 108
  issue: 1
  year: 2000
  ident: 10.1016/j.ins.2025.122104_br0320
  article-title: Computing with membranes
  publication-title: J. Comput. Syst. Sci.
  doi: 10.1006/jcss.1999.1693
– start-page: 173
  year: 2021
  ident: 10.1016/j.ins.2025.122104_br0420
– volume: 18
  start-page: 435
  issue: 3
  year: 2007
  ident: 10.1016/j.ins.2025.122104_br0330
  article-title: Spiking neural P systems: an early survey
  publication-title: Int. J. Found. Comput. Sci.
  doi: 10.1142/S0129054107004772
– volume: 9
  start-page: 243
  issue: 2
  year: 1996
  ident: 10.1016/j.ins.2025.122104_br0050
  article-title: Bounds on the complexity of recurrent neural network implementations of finite state machines
  publication-title: Neural Netw.
  doi: 10.1016/0893-6080(95)00095-X
– volume: 268
  start-page: 545
  issue: 5210
  year: 1995
  ident: 10.1016/j.ins.2025.122104_br0180
  article-title: Computation beyond the Turing limit
  publication-title: Science
  doi: 10.1126/science.268.5210.545
– volume: 126
  start-page: 312
  year: 2020
  ident: 10.1016/j.ins.2025.122104_br0300
  article-title: Automata complete computation with Hodgkin-Huxley neural networks composed of synfire rings
  publication-title: Neural Netw.
  doi: 10.1016/j.neunet.2020.03.019
– volume: 11
  start-page: 771
  issue: 3
  year: 1999
  ident: 10.1016/j.ins.2025.122104_br0230
  article-title: Analog neural nets with Gaussian or other common noise distributions cannot recognize arbitary regular languages
  publication-title: Neural Comput.
  doi: 10.1162/089976699300016656
– volume: 8
  start-page: 1
  issue: 1
  year: 1996
  ident: 10.1016/j.ins.2025.122104_br0280
  article-title: Lower bounds for the computational power of networks of spiking neurons
  publication-title: Neural Comput.
  doi: 10.1162/neco.1996.8.1.1
– year: 2016
  ident: 10.1016/j.ins.2025.122104_br0010
– year: 1967
  ident: 10.1016/j.ins.2025.122104_br0040
– ident: 10.1016/j.ins.2025.122104_br0430
– volume: 317
  start-page: 251
  issue: 1–3
  year: 2004
  ident: 10.1016/j.ins.2025.122104_br0400
  article-title: Hypercomputation: philosophical issues
  publication-title: Theor. Comput. Sci.
  doi: 10.1016/j.tcs.2003.12.014
– volume: 108
  start-page: 495
  year: 2018
  ident: 10.1016/j.ins.2025.122104_br0390
  article-title: Echo state networks are universal
  publication-title: Neural Netw.
  doi: 10.1016/j.neunet.2018.08.025
– volume: 101
  start-page: 86
  year: 2019
  ident: 10.1016/j.ins.2025.122104_br0190
  article-title: Computational capabilities of analog and evolving neural networks over infinite input streams
  publication-title: J. Comput. Syst. Sci.
  doi: 10.1016/j.jcss.2018.11.003
– volume: 28
  start-page: 2222
  issue: 10
  year: 2017
  ident: 10.1016/j.ins.2025.122104_br0350
  article-title: LSTM: a search space odyssey
  publication-title: IEEE Trans. Neural Netw. Learn. Syst.
  doi: 10.1109/TNNLS.2016.2582924
– year: 1969
  ident: 10.1016/j.ins.2025.122104_br0110
– volume: 116
  start-page: 208
  year: 2019
  ident: 10.1016/j.ins.2025.122104_br0240
  article-title: Subrecursive neural networks
  publication-title: Neural Netw.
  doi: 10.1016/j.neunet.2019.04.019
– volume: 304
  start-page: 78
  issue: 5667
  year: 2004
  ident: 10.1016/j.ins.2025.122104_br0370
  article-title: Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication
  publication-title: Science
  doi: 10.1126/science.1091277
– volume: 11
  start-page: 253
  issue: 2
  year: 2012
  ident: 10.1016/j.ins.2025.122104_br0340
  article-title: Membrane system models for super-Turing paradigms
  publication-title: Nat. Comput.
  doi: 10.1007/s11047-011-9291-8
– volume: 3
  start-page: 127
  issue: 3
  year: 2009
  ident: 10.1016/j.ins.2025.122104_br0380
  article-title: Reservoir computing approaches to recurrent neural network training
  publication-title: Comput. Sci. Rev.
  doi: 10.1016/j.cosrev.2009.03.005
SSID ssj0004766
Score 2.47147
Snippet Kolmogorov complexity measures the compressibility of real numbers. We provide a refined characterization of the hypercomputational power of analog, evolving,...
SourceID crossref
elsevier
SourceType Index Database
Publisher
StartPage 122104
SubjectTerms Analog computation
Computability theory
Computational power
Echo state networks
Kolmogorov complexity
Quantization
Recurrent neural networks
Stochastic computation
Title Refined Kolmogorov complexity of analog, evolving and stochastic recurrent neural networks
URI https://dx.doi.org/10.1016/j.ins.2025.122104
Volume 711
WOSCitedRecordID wos001456228400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  issn: 0020-0255
  databaseCode: AIEXJ
  dateStart: 19950101
  customDbUrl:
  isFulltext: true
  dateEnd: 99991231
  titleUrlDefault: https://www.sciencedirect.com
  omitProxy: false
  ssIdentifier: ssj0004766
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1Nb9MwGLZKxwEOaBugjTHkA-IABKWOHTvHadoEmzQhNqTCJYq_9qHNmdquqvj1vKntJIxNYgcuUWS1byI_j948tt8PhN6KTGcmz3ViUpMl1FKeSKKLhGUWxKwkkiuxbDbBj47EeFx8HQyOYy7M_JI7JxaL4vq_Qg1jAHaTOvsAuFujMAD3ADpcAXa4_hPw34wF5ag_HNaXV_VpPannPm7cLEL0ReWaDZtmbg34pnnMUgQVqM6qqa_prELZpqbcJYDofLD4tC9lQyLTkj_hO9rq891KNq1VlhTxR_HhRP7qvCXS8WxS_zLKt83-UTnX34AgrI2w6hIC0qRZmvSdKg8u1LvFESGhy_BfHttvHlzAMqMpnk7Yp-63f1bHvvXVamMJY5jaRQkmysZE6U08QiuEs0IM0crOl73xQZcuy_0RdnzveNi9DPu79R53y5WeBDlZRc_C2gHveMzX0MC4dfS0V1FyHW2HPBT8DvfwwcGDP0c_Aztwxw7csQPXFnt2fMSRGzCgcccN3HIDe27gyI0X6Pv-3snu5yS010gUiNpZYiSXFAShBEkiqkpbJgthCRW5ytLUUF6MtKxoXglDR5ZRU6RWcWk1KJgik2n2Eg1d7cwGwsyAbCWWMMoVzbWRKTMiZzYHMSjAxCZ6H2exvPZVVMp7cdtENM5zGejr5V0JnLn_b68e8owt9KSj8ms0nE1uzDZ6rOaz8-nkTSDMb9cpfkc
linkProvider Elsevier
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Refined+Kolmogorov+complexity+of+analog%2C+evolving+and+stochastic+recurrent+neural+networks&rft.jtitle=Information+sciences&rft.au=Cabessa%2C+J%C3%A9r%C3%A9mie&rft.au=Strozecki%2C+Yann&rft.date=2025-09-01&rft.issn=0020-0255&rft.volume=711&rft.spage=122104&rft_id=info:doi/10.1016%2Fj.ins.2025.122104&rft.externalDBID=n%2Fa&rft.externalDocID=10_1016_j_ins_2025_122104
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0020-0255&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0020-0255&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0020-0255&client=summon