Information fluctuation complexity explained

Information fluctuation complexity is an information-theoretic quantity defined as the fluctuation of information about entropy. It is derivable from fluctuations in the predominance of order and chaos in a dynamic system and has been used as a measure of complexity in many diverse fields. It was introduced in a 1993 paper by Bates and

Definition

The information fluctuation complexity of a discrete dynamic system is a function of the probability distribution of its states when it is subject to random external input data. The purpose of driving the system with a rich information source such as a random number generator or a white noise signal is to probe the internal dynamics of the system in much the same way as a frequency-rich impulse is used in signal processing.

If a system has $N$ possible states and the state probabilities $p\_i$are known, then its information entropy is

Η = \sum ^{N} _{i =1} p _{i} I _{i} = - \sum ^{N} _{i =1} p _{i} log p _{i,}

where $I\_i\; =\; -\backslash log\; p\_i$ is the information content of state $i$.

The information fluctuation complexity of the system is defined as the standard deviation or fluctuation of $I$ about its mean $\backslash Eta$:

\sigma _{I} = \sqrt{\sum ^{N} _{i =1} p _{i(I} _{i} - Η ) ^{2}} = \sqrt{\sum ^{N} _{i =1} p _{iI} ^{2} ^{i} - Η ^{2}}

or

\sigma _{I} = \sqrt{\sum ^{N} _{i =1} p _{i} log ^{2} p _{i} - l (\sum ^{N} _{i =1} p _{i} log p _{i} r ) ^{2}.}

The fluctuation of state information

\sigma _{I} is zero in a maximally disordered system with all p _{i} = 1 /N ; the system simply mimics its random inputs. \sigma _{I} is also zero when the system is perfectly ordered with just one fixed state (p _{1} = 1 ) , regardless of inputs. \sigma _{I} is non-zero between these two extremes with a mixture of both higher-probability states and lower-probability states populating state space.

Fluctuation of information allows for memory and computation

As a complex dynamic system evolves in time, how it transitions between states depends on external stimuli in an irregular way. At times it may be more sensitive to external stimuli (unstable) and at other times less sensitive (stable). If a particular state has several possible next-states, external information determines which one will be next and the system gains that information by following a particular trajectory in state space. But if several different states all lead to the same next-state, then upon entering the next-state the system loses information about which state preceded it. Thus, a complex system exhibits alternating information gain and loss as it evolves in time. The alternation or fluctuation of information is equivalent to remembering and forgetting — temporary information storage or memory — an essential feature of non-trivial computation.

The gain or loss of information associated with transitions between states can be related to state information. The net information gain of a transition from state

i to state j is the information gained when leaving state i less the information lost when entering state j :\Gamma _{ij} = -log p _{i} + log p _{i} .

Here $p\_$ is the forward conditional probability that if the present state is

i then the next state is j and p _{i} is the reverse conditional probability that if the present state is j then the previous state was i . The conditional probabilities are related to the transition probability p _{ij} , the probability that a transition from state i to state j occurs, by:p _{ij} = p _{i} p _{i} = p _{i} p _{j.}

Eliminating the conditional probabilities:

\Gamma _{ij} = -log (p _{ij} /p _{i)} + log (p _{ij} /p _{j)} = log p _{i} - log p _{j} = I _{j} - I _{i.}

Therefore, the net information gained by the system as a result of the transition depends only on the increase in state information from the initial to the final state. It can be shown that this is true even for multiple consecutive

\Gamma = \Delta I is reminiscent of the relation between force and potential energy. I is like potential \Phi and \Gamma is like force F in F = {\nabla \Phi} . External information “pushes” a system “uphill” to a state of higher information potential to accomplish information storage, much like pushing a mass uphill to a state of higher gravitational potential stores energy. The amount of energy storage depends only on the final height, not on the path up the hill. Likewise, the amount of information storage does not depend on the transition path between an initial common state and a final rare state. Once a system reaches a rare state with high information potential, it may then "fall" back to a common state, losing the previously stored information.

It may be useful to compute the standard deviation of

\Gamma about its mean (which is zero), namely the fluctuation of net information gain but \sigma _{I} takes into account multi-transition memory loops in state space and therefore should be more indicative of the computational power of a system. Moreover, \sigma _{I} is easier to calculate because there can be many more transitions than states.

Chaos and order

A dynamic system that is sensitive to external information (unstable) exhibits chaotic behavior whereas one that is insensitive to external information (stable) exhibits orderly behavior. A complex system exhibits both behaviors, fluctuating between them in dynamic balance when subject to a rich information source. The degree of fluctuation is quantified by

\sigma _{I} ; it captures the alternation in the predominance of chaos and order in a complex system as it evolves in time.

Example: rule 110 variant of the elementary cellular

Source:

The rule 110 variant of the elementary cellular automaton has been proven to be capable of universal computation. The proof is based on the existence and interactions of cohesive and self-perpetuating cell patterns known as gliders or spaceships (examples of emergent phenomena associated with complex systems), that imply the capability of groups of automaton cells to remember that a glider is passing through them. It is therefore to be expected that there will be memory loops in state space resulting from alternations of information gain and loss, instability and stability, chaos and order.

Consider a 3-cell group of adjacent automaton cells that obey rule 110: . The next state of the center cell depends on the present state of itself and the end cells as specified by the rule:

1-1-1

1-1-0	1-0-1	1-0-0	0-1-1	0-1-0	0-0-1	0-0-0
next center cell	0	1	1	0	1	1	1	0

To calculate the information fluctuation complexity of this system, attach a driver cell to each end of the 3-cell group to provide a random external stimulus like so, , such that the rule can be applied to the two end cells. Next, determine what the next state is for each possible present state and for each possible combination of driver cell contents, to determine the forward conditional probabilities.

The state diagram of this system is depicted below, with circles representing the states and arrows representing transitions between states. The eight states of this system, to are labeled with the octal equivalent of the 3-bit contents of the 3-cell group: 7 to 0. The transition arrows are labeled with forward conditional probabilities. Notice that there is variability in the divergence and convergence of arrows corresponding to variability in gain and loss of information from the driver cells. The forward conditional probabilities are determined by the proportion of possible driver cell contents that drive a particular transition. For example, for the four possible combinations of two driver cell contents, state 7 leads to states 5, 4, 1 and 0 and so

p _{7} , p _{7} , p _{7} , and p _{7} are each ¼ or 25%. Likewise, state 0 leads to states 0, 1, 0 and 1 and so p _{0} and p _{0} are each ½ or 50%. And so forth.

The state probabilities are related by

p _{j} = \sum ^{7} _{i =0} p _{i} p _{i} and \sum ^{7} _{i =0} p _{i} = 1 .

These linear algebraic equations can be solved manually or with the aid of a computer program for the state probabilities, with the following

Information entropy and complexity can then be calculated from the state probabilities:

Η = - \sum ^{7} _{i =0} p _{i} log _{2} p _{i} = 2 . 86 bits , \sigma _{I} = \sqrt{\sum ^{7} _{i =0} p _{i} log ^{2} _{2} p _{i} - Η ^{2}} = 0 . 56 bits .

Note that the maximum possible entropy for eight states is

log _{2} 8 = 3 bits , which would be the case if all eight states were equally likely with probabilities of ⅛ (randomness). Thus rule 110 has a relatively high entropy or state utilization at 2.86 bits. But this does not preclude a substantial fluctuation of state information about entropy and thus a substantial value of complexity. Whereas, maximum entropy would preclude complexity.

An alternative method can be used to obtain the state probabilities when the analytical method used above is unfeasible. Simply drive the system at its inputs (the driver cells) with a random source for many generations and observe the state probabilities empirically. When this is done via computer simulation for 10 million generations the results are as

Information variables for the rule 110 elementary cellular automaton!number of cells!3!4!5!6!7!8!9!10!11!12!13

Η (bits)	2.86	3.81	4.73	5.66	6.56	7.47	8.34	9.25	10.09	10.97	11.78
\sigma _{I} (bits)	0.56	0.65	0.72	0.73	0.79	0.81	0.89	0.90	1.00	1.01	1.15
\sigma _{I} / Η	0.20	0.17	0.15	0.13	0.12	0.11	0.11	0.10	0.10	0.09	0.10

Since both

Η and \sigma _{I} increase with system size, their dimensionless ratio \sigma _{I/Η} , the relative information fluctuation complexity, is included to better compare systems of different sizes. Notice that the empirical and analytical results agree for the 3-cell automaton and that the relative complexity levels off to about 0.10 by 10 cells.

In the paper by Bates and

\sigma _{I} is computed for all elementary cellular automaton rules and it was observed that the ones that exhibit slow-moving gliders and possibly stationary objects, as rule 110 does, are highly correlated with large values of \sigma _{I} . \sigma _{I} can therefore be used as a filter to select candidate rules for universal computation, which is difficult to prove.

Applications

Although the derivation of the information fluctuation complexity formula is based on information fluctuations in a dynamic system, the formula depends only on state probabilities and so is also applicable to any probability distribution, including those derived from static images or text.

Over the years the original paper has been referred to by researchers in many diverse fields: complexity theory,^[1] complex systems science, complex networks,^[2] chaotic dynamics,^[3] many-body localization entanglement,^[4] environmental engineering,^[5] ecological complexity,^[6] ecological time-series analysis, ecosystem sustainability,^[7] air^[8] and water^[9] pollution, hydrological wavelet analysis,^[10] soil water flow,^[11] soil moisture,^[12] headwater runoff,^[13] groundwater depth,^[14] air traffic control,^[15] flow patterns^[16] and flood events,^[17] topology, economics,^[18] market forecasting of metal^[19] and electricity^[20] prices, health informatics, human cognition,^[21] human gait kinematics,^[22] neurology,^[23] EEG analysis,^[24] education,^[25] investing,^[26] artificial life^[27] and aesthetics.^[28]

References

1. Atmanspacher. Harald. September 1997. Cartesian cut, Heisenberg cut, and the concept of complexity. World Futures. en. 49. 3–4. 333–355. 10.1080/02604027.1997.9972639. 0260-4027.
2. Huang . Min . Sun . Zhongkui . Donner . Reik V. . Zhang . Jie . Gua . Shuguang . Zou . Yong . 2021-03-09 . Characterizing dynamical transitions by statistical complexity measures based on ordinal pattern transition networks . . 31 . 3 . 033127 . 10.1063/5.0038876. 33810737 . 232771788 .
3. Wackerbauer. Renate. 1995-11-01. Noise-induced stabilization of the Lorenz system. Physical Review E. 52. 5. 4745–4749. 10.1103/PhysRevE.52.4745. 9963970. 1995PhRvE..52.4745W.
4. Hamilton . Gregory A. . Clark . Bryan K. . 2023-02-14 . Quantifying unitary flow efficiency and entanglement for many-body localization . Physical Review B . 107 . 6 . 064203 . 10.1103/PhysRevB.107.064203. 2110.10148 . 239024666 .
5. Book: Singh, Vijay P.. Entropy Theory and its Application in Environmental and Water Engineering. 2013-01-10. John Wiley & Sons. 978-1-118-42860-3. en.
6. Parrott. Lael. 2010-11-01. Measuring ecological complexity. Ecological Indicators. en. 10. 6. 1069–1076. 10.1016/j.ecolind.2010.03.014. 1470-160X.
7. Wang. Chaojun. Zhao. Hongrui. 2019-04-18. Analysis of remote sensing time-series data to foster ecosystem sustainability: use of temporal information entropy. International Journal of Remote Sensing. 40. 8. 2880–2894. 10.1080/01431161.2018.1533661. 2019IJRS...40.2880W. 135003743. 0143-1161.
8. Klemm. Otto. Lange. Holger. 1999-12-01. Trends of air pollution in the Fichtelgebirge Mountains, Bavaria. Environmental Science and Pollution Research. en. 6. 4. 193–199. 10.1007/BF02987325. 19005662. 35043. 1614-7499.
9. Wang. Kang. Lin. Zhongbing. 2018. Characterization of the nonpoint source pollution into river at different spatial scales. Water and Environment Journal. en. 32. 3. 453–465. 10.1111/wej.12345. 115667734 . 1747-6593.
10. Labat. David. 2005-11-25. Recent advances in wavelet analyses: Part 1. A review of concepts. Journal of Hydrology. en. 314. 1. 275–288. 10.1016/j.jhydrol.2005.04.003. 2005JHyd..314..275L. 0022-1694.
11. Pachepsky. Yakov. Guber. Andrey. Jacques. Diederik. Simunek. Jiri. Van Genuchten. Marthinus Th.. Nicholson. Thomas. Cady. Ralph. 2006-10-01. Information content and complexity of simulated soil water fluxes. Geoderma. Fractal Geometry Applied to Soil and Related Hierarchical Systems - Fractals, Complexity and Heterogeneity. en. 134. 3. 253–266. 10.1016/j.geoderma.2006.03.003. 2006Geode.134..253P. 0016-7061.
12. Kumar. Sujay V.. Dirmeyer. Paul A.. Peters-Lidard. Christa D.. Christa Peters-Lidard. Bindlish. Rajat. Bolten. John. 2018-01-01. Information theoretic evaluation of satellite soil moisture retrievals. Remote Sensing of Environment. en. 204. 392–400. 10.1016/j.rse.2017.10.016. 32636571. 7340154. 2018RSEnv.204..392K. 0034-4257. 2060/20180003069. free.
13. Hauhs. Michael. Lange. Holger. 2008. Classification of Runoff in Headwater Catchments: A Physical Problem?. Geography Compass. en. 2. 1. 235–254. 10.1111/j.1749-8198.2007.00075.x. 1749-8198.
14. Liu. Meng. Liu. Dong. Liu. Le. 2013-09-01. Complexity research of regional groundwater depth series based on multiscale entropy: a case study of Jiangsanjiang Branch Bureau in China. Environmental Earth Sciences. en. 70. 1. 353–361. 10.1007/s12665-012-2132-y. 2013EES....70..353L . 128958458. 1866-6299.
15. Xing. Jing. Manning. Carol A.. April 2005. Complexity and Automation Displays of Air Traffic Control: Literature Review and Analysis. https://web.archive.org/web/20220601070456/https://apps.dtic.mil/docs/citations/ADA460107. live. June 1, 2022. en.
16. Book: Wang. Kang. Li. Li. 2008 First International Conference on Intelligent Networks and Intelligent Systems . Characterizing Heterogeneous Flow Patterns Using Information Measurements . November 2008. 654–657. 10.1109/ICINIS.2008.110. 8867649.
17. Al Sawaf. Mohamad Basel. Kawanisi. Kiyosi. 2020-11-01. Assessment of mountain river streamflow patterns and flood events using information and complexity measures. Journal of Hydrology. en. 590. 125508. 10.1016/j.jhydrol.2020.125508. 2020JHyd..59025508A. 225261677. 0022-1694.
18. Jurado-González . Javier . Gómez-Barroso . José Luis . 2022-11-28 . Economic complexity and Information Society paradigms: a hybrid contribution to explain economic growth . Technological and Economic Development of Economy . en . 28 . 6 . 1871–1896 . 10.3846/tede.2022.17104 . 254339040 . 2029-4921. free .
19. He. Kaijian. Lu. Xingjing. Zou. Yingchao. Keung Lai. Kin. 2015-09-01. Forecasting metal prices with a curvelet based multiscale methodology. Resources Policy. en. 45. 144–150. 10.1016/j.resourpol.2015.03.011. 2015RePol..45..144H . 0301-4207.
20. He. Kaijian. Xu. Yang. Zou. Yingchao. Tang. Ling. 2015-05-01. Electricity price forecasts using a Curvelet denoising based approach. Physica A: Statistical Mechanics and Its Applications. en. 425. 1–9. 10.1016/j.physa.2015.01.012. 0378-4371.
21. Shi Xiujian. Sun Zhiqiang. Li Long. Xie Hongwei. Human Cognitive Complexity Analysis in Transportation Systems. Logistics. Proceedings. 2009. 4361–4368. 10.1061/40996(330)637. 9780784409961.
22. Book: Zhang. Shutao. Qian. Jinwu. Shen. Linyong. Wu. Xi. Hu. Xiaowu. 2015 International Symposium on Bioelectronics and Bioinformatics (ISBB) . Gait complexity and frequency content analyses of patients with Parkinson's disease . October 2015. 87–90. 10.1109/ISBB.2015.7344930. 978-1-4673-6609-0. 2891655.
23. Wang. Jisung. Noh. Gyu-Jeong. Choi. Byung-Moon. Ku. Seung-Woo. Joo. Pangyu. Jung. Woo-Sung. Kim. Seunghwan. Lee. Heonsoo. 2017-07-13. Suppressed neural complexity during ketamine- and propofol-induced unconsciousness. Neuroscience Letters. en. 653. 320–325. 10.1016/j.neulet.2017.05.045. 28572032. 13767209. 0304-3940.
24. Bola. Michał. Orłowski. Paweł. Płomecka. Martyna. Marchewka. Artur. 2019-01-30. EEG signal diversity during propofol sedation: an increase in sedated but responsive, a decrease in sedated and unresponsive subjects. bioRxiv. en. 444281. 10.1101/444281. 214726084.
25. Dilger. Alexander. 2012-01-01. Endogenous complexity, specialisation and general education. On the Horizon. 20. 1. 49–53. 10.1108/10748121211202062. 1074-8121.
26. Web site: Dynamic strategic investment portfolio management model. Ivanyuk. Vera Alekseevna. 2015. elibrary.ru.
27. Peña . Eric . Sayama . Hiroki . 2021-05-02 . Life Worth Mentioning: Complexity in Life-Like Cellular Automata . . 27 . 2 . 105–112 . 10.1162/artl_a_00348. 34727158 . 240424110 . free .
28. Aesthetic Automata: Synthesis and Simulation of Aesthetic Behaviour in Cellular Automata. Goldsmiths, University of London. 2019-11-30. doctoral. 10.25602/gold.00027681. Mohammad Ali. Javaheri Javid.