Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Hierarchical Framework with Spatio-Temporal Consistency Learning for Emergence Detection in Complex Adaptive Systems

Siyuan Chen, Xin Du, Jiahai Wang

TL;DR

A hierarchical framework with spatio-temporal consistency learning (HSTCL) to solve emergence-related spatial patterns and the nonlinear relationships among agents by learning the system representation and agent representations, respectively.

Abstract

Emergence, a global property of complex adaptive systems (CASs) constituted by interactive agents, is prevalent in real-world dynamic systems, e.g., network-level traffic congestions. Detecting its formation and evaporation helps to monitor the state of a system, allowing to issue a warning signal for harmful emergent phenomena. Since there is no centralized controller of CAS, detecting emergence based on each agent's local observation is desirable but challenging. Existing works are unable to capture emergence-related spatial patterns, and fail to model the nonlinear relationships among agents. This paper proposes a hierarchical framework with spatio-temporal consistency learning to solve these two problems by learning the system representation and agent representations, respectively. Spatio-temporal encoders composed of spatial and temporal transformers are designed to capture agents' nonlinear relationships and the system's complex evolution. Agents' and the system's representations are learned to preserve the spatio-temporal consistency by minimizing the spatial and temporal dissimilarities in a self-supervised manner in the latent space. Our method achieves more accurate detection than traditional methods and deep learning methods on three datasets with well-known yet hard-to-detect emergent behaviors. Notably, our hierarchical framework is generic in incorporating other deep learning methods for agent-level and system-level detection.

A Hierarchical Framework with Spatio-Temporal Consistency Learning for Emergence Detection in Complex Adaptive Systems

TL;DR

A hierarchical framework with spatio-temporal consistency learning (HSTCL) to solve emergence-related spatial patterns and the nonlinear relationships among agents by learning the system representation and agent representations, respectively.

Abstract

Emergence, a global property of complex adaptive systems (CASs) constituted by interactive agents, is prevalent in real-world dynamic systems, e.g., network-level traffic congestions. Detecting its formation and evaporation helps to monitor the state of a system, allowing to issue a warning signal for harmful emergent phenomena. Since there is no centralized controller of CAS, detecting emergence based on each agent's local observation is desirable but challenging. Existing works are unable to capture emergence-related spatial patterns, and fail to model the nonlinear relationships among agents. This paper proposes a hierarchical framework with spatio-temporal consistency learning to solve these two problems by learning the system representation and agent representations, respectively. Spatio-temporal encoders composed of spatial and temporal transformers are designed to capture agents' nonlinear relationships and the system's complex evolution. Agents' and the system's representations are learned to preserve the spatio-temporal consistency by minimizing the spatial and temporal dissimilarities in a self-supervised manner in the latent space. Our method achieves more accurate detection than traditional methods and deep learning methods on three datasets with well-known yet hard-to-detect emergent behaviors. Notably, our hierarchical framework is generic in incorporating other deep learning methods for agent-level and system-level detection.
Paper Structure (40 sections, 25 equations, 13 figures, 3 tables)

This paper contains 40 sections, 25 equations, 13 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Emergence Detection
  4. Related Graph Mining Tasks
  5. Self-Supervised Learning for Spatio-Temporal Data
  6. Proposed Method
  7. Problem Formulation
  8. Motivation and Overview of HSTCL
  9. Agent-Level Detection
  10. Spatio-Temporal Encoder
  11. Spatio-Temporal Consistency Learning
  12. Temporal Consistency Loss
  13. Spatial Consistency Loss
  14. Optimization
  15. Communication
  16. ...and 25 more sections

Figures (13)

  • Figure 1: (a) illustrates the emergence through the traffic flow. (b) shows that emergence detection is framed as a change-point detection problem.
  • Figure 2: Overview of HSTCL. HSTCL contains three hierarchies, agents, region monitors, and a global monitor, which can be conceptually implemented by the end-edge-cloud collaborative framework. Agents sense the states of neighbors, measure the change in relationships, and communicate with them to make agent-level detection. The detecting results are coarse-grained to region states, whose spatial patterns are used for system-level detection. Representations of agents and the system are learned via spatio-temporal consistency learning (STCL) technique to support agent-level and system-level detection, respectively.
  • Figure 3: Procedure of agent-level STCL.
  • Figure 4: Procedure of system-level STCL.
  • Figure 5: Case study of spatial patterns on the Traffic dataset. In (a), each line segment represents a road segment. Segments in red are congested, while those in gray are normal. In (b), each grid represents a region state. The darker the color, the more agent-level detections are collected. Best viewed in color.
  • ...and 8 more figures

Theorems & Definitions (4)

  • Definition 1: Dyanmic graph
  • Definition 2: CPD in dynamic graphs
  • Definition 3: Distributed Setting for Emergence Detection
  • Definition 4: Criterion for CPD