Landau theory of social clustering

Plewczyński, Dariusz

doi:10.1016/s0378-4371(98)00349-5

Cited by 25 publications

(27 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The system defined in this way is similar to previously postulated cellular automata models of opinion change in social sciences [44,49,50]. The main differences of those approaches from the previously described cellular automata models is given by the infinite-range interactions and fully connected cellular automata, that are better fitted to the learning context of the problem.…”

Section: Presence Of Noisementioning

confidence: 89%

“…The cellular automata with intrinsic disorder was later solved analytically in the continuous limit by Plewczynski [49], and proved that in the model of Cartesian social space (therefore not fully connected) and containing no learning rules, one can also observe different phases (small clusters in the sparse phase with large role of strong individuals, and high density phase with almost uniform opinion). The later results of Hołyst et al, where numerical simulations and analytical models were tested in simplified geometries, proved the usefulness of mean-field formalism in describing the social impact theory, and the presence of the equilibrium states of the system with complex intermittent behavior [50][51][52][53].…”

Section: In Nt Tr Ro Od Du Uc Ct Ti Io On Nmentioning

confidence: 99%

“…The above solutions for the system are given by the minority clusters surrounded by the majority agents, and the dynamic is of "staircase" character in the presence of small noise [44,49]. In general three initial states of the system can observed: sparse (no correlation between the agent strength and its state), middle density (a state of a agent starts to be correlated with its strength, lot of interesting meta-stable global configurations), and large density state (most of agents initially have similar state, therefore the role of coupling is not so important).…”

Section:  Asmentioning

confidence: 99%

“…[44], Plewczynski [49] and Hołyst et al [50,51,53] to novel problem of meta-learning. The CA model is characterized by long-range interactions between individual agents, the fully connected learning space (each individual learners exchange their prediction results with all others) and by an intrinsic disorder that allow for complex learning geometries to appear in the system.…”

mentioning

confidence: 99%

“…The equilibration of the system in clustered state is given by intermittent, consecutive steps in the form of "stair-case" dynamics [44,49]. First the strongest 17 agents change their state, then the weaker rest of minority cluster collapses.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Mean-Field Theory of Meta-learning

SpringerReference

View full text Add to dashboard Cite

A Ab bs st tr ra ac ct tWe discuss here the mean-field theory for a cellular automata model of meta-learning. The metalearning is the process of combining outcomes of individual learning procedures in order to determine the final decision with higher accuracy than any single learning method. Our method is constructed from an ensemble of interacting, learning agents, that acquire and process incoming information using various types, or different versions of machine learning algorithms. The abstract learning space, where all agents are located, is constructed here using a fully connected model that couples all agents with random strength values. The cellular automata network simulates the higher level integration of information acquired from the independent learning trials. The final classification of incoming input data is therefore defined as the stationary state of the meta-learning system using simple majority rule, yet the minority clusters that share opposite classification outcome can be observed in the system. Therefore, the probability of selecting proper class for a given input data, can be estimated even without the prior knowledge of its affiliation. The fuzzy logic can be easily introduced into the system, even if learning agents are build from simple binary classification machine learning algorithms by calculating the percentage of agreeing agents.

show abstract

Section: Presence Of Noisementioning

confidence: 89%