scholarly journals Unsupervised manifold learning of collective behavior

2020 ◽  
Author(s):  
Mathew Titus ◽  
George Hagstrom ◽  
James R. Watson
Keyword(s):  
Complex Systems ◽  
Prior Knowledge ◽  
Collective Behavior ◽  
Generative Models ◽  
Emergent Behavior ◽  
New Method ◽  
Collective Behaviors ◽  
State Classification ◽  
Macro Scale ◽  
Fish Schooling

AbstractCollective behavior is an emergent property of numerous complex systems, from financial markets to cancer cells to predator-prey ecological systems. Characterizing modes of collective behavior is often done through human observation, training generative models, or other supervised learning techniques. Each of these cases requires knowledge of and a method for characterizing the macro-state(s) of the system. This presents a challenge for studying novel systems where there may be little prior knowledge. Here, we present a new unsupervised method of detecting emergent behavior in complex systems, and discerning between distinct collective behaviors. We require only metrics, d(1), d(2), defined on the set of agents, X, which measure agents’ nearness in variables of interest. We apply the method of diffusion maps to the systems (X, d(i)) to recover efficient embeddings of their interaction networks. Comparing these geometries, we formulate a measure of similarity between two networks, called the map alignment statistic (MAS). A large MAS is evidence that the two networks are codetermined in some fashion, indicating an emergent relationship between the metrics d(1) and d(2). Additionally, the form of the macro-scale organization is encoded in the covariances among the two sets of diffusion map components. Using these covariances we discern between different modes of collective behavior in a data-driven, unsupervised manner. This method is demonstrated on empirical fish schooling data. We show that our state classification subdivides the known behaviors of the school in a meaningful manner, leading to a finer description of the system’s behavior.Author summaryMany complex systems in society and nature exhibit collective behavior where individuals’ local interactions lead to system-wide organization. One challenge we face today is to identify and characterize these emergent behaviors, and here we have developed a new method for analyzing data from individuals, to detect when a given complex system is exhibiting system-wide organization. Importantly, our approach requires no prior knowledge of the fashion in which the collective behavior arises, or the macro-scale variables in which it manifests. We apply the new method to an agent-based model and empirical observations of fish schooling. While we have demonstrated the utility of our approach to biological systems, it can be applied widely to financial, medical, and technological systems for example.

scholarly journals Unsupervised manifold learning of collective behavior

2021 ◽  
Vol 17 (2) ◽  
pp. e1007811
Author(s):  
Mathew Titus ◽  
George Hagstrom ◽  
James R. Watson
Keyword(s):  
Complex Systems ◽  
Collective Behavior ◽  
Generative Models ◽  
Emergent Behavior ◽  
Collective Behaviors ◽  
State Classification ◽  
Learning Techniques ◽  
Macro Scale ◽  
Map Alignment ◽  
Observation Training

Collective behavior is an emergent property of numerous complex systems, from financial markets to cancer cells to predator-prey ecological systems. Characterizing modes of collective behavior is often done through human observation, training generative models, or other supervised learning techniques. Each of these cases requires knowledge of and a method for characterizing the macro-state(s) of the system. This presents a challenge for studying novel systems where there may be little prior knowledge. Here, we present a new unsupervised method of detecting emergent behavior in complex systems, and discerning between distinct collective behaviors. We require only metrics, d(1), d(2), defined on the set of agents, X, which measure agents’ nearness in variables of interest. We apply the method of diffusion maps to the systems (X, d(i)) to recover efficient embeddings of their interaction networks. Comparing these geometries, we formulate a measure of similarity between two networks, called the map alignment statistic (MAS). A large MAS is evidence that the two networks are codetermined in some fashion, indicating an emergent relationship between the metrics d(1) and d(2). Additionally, the form of the macro-scale organization is encoded in the covariances among the two sets of diffusion map components. Using these covariances we discern between different modes of collective behavior in a data-driven, unsupervised manner. This method is demonstrated on a synthetic flocking model as well as empirical fish schooling data. We show that our state classification subdivides the known behaviors of the school in a meaningful manner, leading to a finer description of the system’s behavior.

Particle anisotropy tunes emergent behavior in active colloidal systems

Soft Matter ◽  
2022 ◽  
Author(s):  
Shannon E. Moran ◽  
Isaac R. Bruss ◽  
Philip Shoenhofer ◽  
Sharon C Glotzer
Keyword(s):  
Collective Behavior ◽  
Particle Systems ◽  
Emergent Behavior ◽  
Active Particle ◽  
Colloidal Systems ◽  
Particle Anisotropy

Studies of active particle systems have demonstrated that particle anisotropy can impact the collective behavior of a system. However, systems studied to date have served as one-off demonstrations of concept,...

Emergent Behavior in Complex Organizations

2015 ◽  
Vol 6 (4) ◽  
pp. 14-27
Author(s):  
Ernesto R. Alvarez-Molina ◽  
Luis G. Martínez ◽  
Manuel Castañón-Puga ◽  
Antonio Rodriguez-Diaz
Keyword(s):  
Software Development ◽  
Complex Systems ◽  
Computational Modeling ◽  
Development Process ◽  
Work Teams ◽  
Focus Of Attention ◽  
Emergent Behavior ◽  
Software Development Process ◽  
Viable System ◽  
Emergent Behaviors

Organizations are complex systems, which are formed by other subsystems such as work teams, and are the focus of attention in this research. This article makes an approach to the teams involved software development process in IT companies using a viable system based model and computational modeling. An analysis of teamwork is made from a socio-technical perspective, where individuals and technology produce emergent behaviors that may be crucial to achieving goals, since fellowship, collaboration, and culture are relevant processes within these organizations and technology also playing an important role.

scholarly journals Confronting Functional Complexity Failures: The Case of a Runway Excursion in Munich

2019 ◽  
Vol 63 (1) ◽  
pp. 116-120
Author(s):  
Robert Mauro ◽  
Lance Sherry
Keyword(s):  
Complex Systems ◽  
Emergent Behavior ◽  
Operating Region ◽  
Geographically Distributed ◽  
Functional Complexity ◽  
The Individual ◽  
Emergent Behaviors ◽  
Normal Accidents

Complex systems often produce unanticipated emergent behavior as a result of the interactions between behaviorally complex sub-systems or agents. The sub-systems may be human or artificial. They may be co-located or geographically distributed and operate autonomously. Although the individual sub-systems may be tested and certified for high levels of reliability (e.g. 10-7), interactions between the sub-systems may occur so that emergent behaviors allow the system to migrate into an unsafe operating region. This may occur even when all of the sub-systems are behaving nominally and no equipment has failed. This phenomenon is called a “functional complexity failure.” In this paper, we present an analysis of a functional complexity failure that resulted in a runway excursion and discuss the options for detecting and mitigating the conditions for these “normal accidents” before the accident occurs.

Nanotectonica: Architectural Design Studio and Table Top SEM

2008 ◽  
Vol 16 (5) ◽  
pp. 40-43
Author(s):  
Jonas Coersmeier ◽  
Donovan N. Leonard
Keyword(s):  
Architectural Design ◽  
Generative Models ◽  
Design Studio ◽  
Research Seminar ◽  
Large Span ◽  
Mother Nature ◽  
Macro Scale ◽  
Buckminster Fuller ◽  
Design Students ◽  
Architectural Design Studio

Inspired by architect Frei Otto and design scientist Buckminster Fuller, third year Pratt Institute design students from Jonas Coersmeier’s design studio and research seminar (of Spring 2008) utilized a Table Top SEM to observe micro and nano-scale features produced solely by Mother Nature. After analyzing and documenting the intricacy, beauty and functionality of natural structures, students selected structural entities typically not observed on the macro scale, and utilized the micrograph data to generate analytical drawings followed by generative models for design of a large span structure that would become an aquatic center in the Williamsburg neighborhood of Brooklyn, N.Y.

scholarly journals Genetic architecture of collective behaviors in zebrafish

2018 ◽  
Author(s):  
Wenlong Tang ◽  
Guoqiang Zhang ◽  
Fabrizio Serluca ◽  
Jingyao Li ◽  
Xiaorui Xiong ◽  
...  
Keyword(s):  
Collective Behavior ◽  
Adrenergic Receptor ◽  
Genetic Architecture ◽  
Single Gene ◽  
Gene Mutations ◽  
Collective Actions ◽  
Adult Zebrafish ◽  
Inner Mitochondrial Membrane ◽  
Collective Behaviors ◽  
Species Survival

AbstractCollective behaviors of groups of animals, such as schooling and shoaling of fish, are central to species survival, but genes that regulate these activities are not known. Here we parsed collective behavior of groups of adult zebrafish using computer vision and unsupervised machine learning into a set of highly reproducible, unitary, several hundred millisecond states and transitions, which together can account for the entirety of relative positions and postures of groups of fish. Using CRISPR-Cas9 we then targeted for knockout 35 genes associated with autism and schizophrenia. We found mutations in three genes had distinctive effects on the amount of time spent in the specific states or transitions between states. Mutation in immp2l (inner mitochondrial membrane peptidase 2-like gene) enhances states of cohesion, so increases shoaling; mutation in in the Nav1.1 sodium channel, scn1lab+/− causes the fish to remain scattered without evident social interaction; and mutation in the adrenergic receptor, adra1aa−/−, keeps fish close together and retards transitions between states, leaving fish motionless for long periods. Motor and visual functions seemed relatively well-preserved. This work shows that the behaviors of fish engaged in collective activities are built from a set of stereotypical states. Single gene mutations can alter propensities to collective actions by changing the proportion of time spent in these states or the tendency to transition between states. This provides an approach to begin dissection of the molecular pathways used to generate and guide collective actions of groups of animals.

scholarly journals Weakly Supervised Disentanglement by Pairwise Similarities

2020 ◽  
Vol 34 (04) ◽  
pp. 3495-3502 ◽  
Author(s):  
Junxiang Chen ◽  
Kayhan Batmanghelich
Keyword(s):  
Real World ◽  
Latent Variables ◽  
Generative Models ◽  
Experimental Results ◽  
New Method ◽  
Weak Supervision ◽  
Real World Problem ◽  
Variational Autoencoder ◽  
Weakly Supervised

Recently, researches related to unsupervised disentanglement learning with deep generative models have gained substantial popularity. However, without introducing supervision, there is no guarantee that the factors of interest can be successfully recovered (Locatello et al. 2018). Motivated by a real-world problem, we propose a setting where the user introduces weak supervision by providing similarities between instances based on a factor to be disentangled. The similarity is provided as either a binary (yes/no) or real-valued label describing whether a pair of instances are similar or not. We propose a new method for weakly supervised disentanglement of latent variables within the framework of Variational Autoencoder. Experimental results demonstrate that utilizing weak supervision improves the performance of the disentanglement method substantially.

Sign in / Sign up

Export Citation Format

Share Document