Convergent cross sorting for estimating dynamic coupling

Breston, Leo; Leonardis, Eric J.; Quinn, Laleh K.; Tolston, Michael T.; Wiles, Janet; Chiba, Andrea A.

doi:10.1038/s41598-021-98864-2

Cited by 17 publications

(24 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is recommended to use techniques such as burst detection to capture the more transient agent-based brain responses within behavior. Future approaches should also examine inter-regional communication, such as coherence and dynamic coupling ( Fries, 2015 ; Breston et al, 2021 ). Additionally, future studies should also incorporate a richer repertoire of stimuli and robots.…”

Section: Discussionmentioning

confidence: 99%

Interactive neurorobotics: Behavioral and neural dynamics of agent interactions

et al. 2022

Self Cite

View full text Add to dashboard Cite

Interactive neurorobotics is a subfield which characterizes brain responses evoked during interaction with a robot, and their relationship with the behavioral responses. Gathering rich neural and behavioral data from humans or animals responding to agents can act as a scaffold for the design process of future social robots. This research seeks to study how organisms respond to artificial agents in contrast to biological or inanimate ones. This experiment uses the novel affordances of the robotic platforms to investigate complex dynamics during minimally structured interactions that would be difficult to capture with classical experimental setups. We then propose a general framework for such experiments that emphasizes naturalistic interactions combined with multimodal observations and complementary analysis pipelines that are necessary to render a holistic picture of the data for the purpose of informing robotic design principles. Finally, we demonstrate this approach with an exemplar rat–robot social interaction task which included simultaneous multi-agent tracking and neural recordings.

show abstract

Section: Discussionmentioning

confidence: 99%

Interactive neurorobotics: Behavioral and neural dynamics of agent interactions

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Where w i ¼ e ÀðdðXðtÞ,Xðt i ÞÞ=dðXt,Xt 1 ÞÞ and d( • , • ) denotes the Euclidean distance. This latter operation causes several problems with the CCM method mentioned in [29,[37][38][39], the most relevant being:…”

Section: Convergent Cross Sorting Methodsmentioning

confidence: 99%

“…Where wi=normale−false(normald(X(t),X(ti))/normald(Xt,Xt1)false) and d ( · , · ) denotes the Euclidean distance. This latter operation causes several problems with the CCM method mentioned in [29,37–39], the most relevant being: Requires relatively long time series to obtain reliable results, i.e. length n∼Ofalse(103false).Failure in oscillatory variables that show a highly dominant or shared single frequency.Poor performance for noisy and also for strongly coupled time series.Shows some problems when applied to time series that are not ‘fully deterministic’.Short, oscillatory and noisy time series that are somewhere in between the extremes of being ‘fully deterministic’ (i.e.…”

Section: Causality Detection Methodsmentioning

confidence: 99%

“…Accordingly, the main objectives of this study are: (i) to determine the conditions under which CS [32] and CCS [29] methods can be safely applied to detect causality in real-world time series, (ii) to highlight the advantages and limitations of both methods for scientists within the Ecology, Biology, Earth and Marine Science communities, among others, who can use these methods to detect causal networks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Usefulness and limitations of convergent cross sorting and continuity scaling methods for their application in simulated and real-world time series

2023

View full text Add to dashboard Cite

Causality detection methods are valuable tools for detecting causal links in complex systems. The efficiency of continuity scaling (CS) and the convergent cross sorting (CSS) methods to detect causality was analysed. Usefulness and limitations of both methods in their application to simulated and real-world time series was explored under different scenarios. We find that CS is more robust and efficient than the CSS method for all simulated systems, even when increasing noise levels were considered. Both methods were not able to infer causality when time series with a marked difference in their main frequencies were analysed. Minimum time-series length required for the detection of a causal link depends on intrinsic system dynamics and on the method selected to detect it. Using simulated time series, only the CS method was capable to detect bidirectional causality. Causality detection, using the CS method, should at least include: (i) causality strength convergence analysis, (ii) statistical tests of significance, (iii) time-series standardization, and (iv) causality strength ratios as a strength indicator of relative causality between systems. Causality cannot be detected by either method in simulated time series that exhibit generalized synchronization.

show abstract

“…In economics and ecosystems, statistical tests including Granger Causality [27] and Convergence Cross Mapping (CCM) [12] have been developed to detect the causality between time series of two variables. While Granger Causality assumes that variables are separable, CCM variants [28], [29], [30], [31] seek synergistic effects between variables in stochastic and dynamical systems. Causality has been rarely studied in multi-agent task allocation to quantitatively measure the level of coordination in strategies.…”

Section: Related Workmentioning

confidence: 99%

Evaluating Emergent Coordination in Multi-Agent Task Allocation Through Causal Inference and Sub-Team Identification

Ghadami

Bayrak

et al. 2023

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Coordination in multi-agent systems is a vital component in teaming effectiveness. In dynamically changing situations, agent decisions depict emergent coordination strategies from following pre-defined rules to exploiting incentive-driven policies. While multi-agent reinforcement learning shapes team behaviors from experience, interpreting learned coordination strategies offers benefits in understanding complex agent dynamics and further improvement in developing adaptive strategies for evolving and unexpected situations. In this work, we develop an approach to quantitatively measure team coordination by collecting decision time series data, detecting causality between agents, and identifying statistically high coordinated sub-teams in decentralized multiagent task allocation operations. We focus on multi-agent systems with homogeneous agents and homogeneous tasks as the strategy formation is more ambiguous and challenging than heterogeneous teams with specialized capabilities. Emergent team coordination is then analyzed using rule-based and reinforcement learning-based strategies for task allocation in operations at different demand stages (stress) levels. We also investigate correlation vs. causation and agent over-or under-estimating demand levels.

show abstract

Convergent cross sorting for estimating dynamic coupling

Cited by 17 publications

References 31 publications

Interactive neurorobotics: Behavioral and neural dynamics of agent interactions

Interactive neurorobotics: Behavioral and neural dynamics of agent interactions

Usefulness and limitations of convergent cross sorting and continuity scaling methods for their application in simulated and real-world time series

Evaluating Emergent Coordination in Multi-Agent Task Allocation Through Causal Inference and Sub-Team Identification

Contact Info

Product

Resources

About