Next-Generation Simulation Illuminates Scientific Problems of Organised Complexity

TL;DR

Problems of organised complexity cannot be fully resolved by any single scientific paradigm. The authors propose next-generation simulation (NGS) as a paradigms-integration platform realized through Sophisticated Behavioural Simulation (SBS), which combines foundation-model commonsense reasoning, professional-domain modules, and personalised knowledge to create authentic agent behaviours. A hierarchical model tower and reinforcement-learning-based learning enable dynamic adaptation and cross-paradigm integration, while SBS supports analysis through past reproduction and behavioural rehearsing to study causality and risk under high-order chaos. While acknowledging inherent uncertainties and practical constraints, the framework aims to enable cost-effective, risk-aware exploration and policy testing in complex social systems where real-world experiments are impractical.

Abstract

As artificial intelligence becomes increasingly prevalent in scientific research, data-driven methodologies appear to overshadow traditional approaches in resolving scientific problems. In this Perspective, we revisit a classic classification of scientific problems and acknowledge that a series of unresolved problems remain. Throughout the history of researching scientific problems, scientists have continuously formed new paradigms facilitated by advances in data, algorithms, and computational power. To better tackle unresolved problems, especially those of organised complexity, a novel paradigm is necessitated. While recognising that the strengths of new paradigms have expanded the scope of resolvable scientific problems, we aware that the continued advancement of data, algorithms, and computational power alone is hardly to bring a new paradigm. We posit that the integration of paradigms, which capitalises on the strengths of each, represents a promising approach. Specifically, we focus on next-generation simulation (NGS), which can serve as a platform to integrate methods from different paradigms. We propose a methodology, sophisticated behavioural simulation (SBS), to realise it. SBS represents a higher level of paradigms integration based on foundational models to simulate complex systems, such as social systems involving sophisticated human strategies and behaviours. NGS extends beyond the capabilities of traditional mathematical modelling simulations and agent-based modelling simulations, and therefore, positions itself as a potential solution to problems of organised complexity in complex systems.

Figures (5)

• Figure 1: Three classes of systems and their examples.
• Figure 2: The evolution of the four paradigms aligned with the three elements.
• Figure 3: The framework of SBS.
• Figure 4: The dice rolling experiments was conducted on LLaMA-2-7B-chat and LLaMA-2-13B-chat.
• Figure 5: (a) The illustration of causality analysis. In this example, $e_i, e_n$ have similar special states to $e_0$, and the behaviour $b_{i,t_j}^{a_k}, b_{n,t_i}^{a_j}$ of the simulated data $B_i, B_n$ is suspected by manual analysis. (b) The illustration of rule-based analysis. In this example, $e_i, e_n$ have the special states unexpectedly disobeying some rules, and through manual analysis of simulated data $B_i, B_n$ behaviour $b_{i,t_j}^{a_k}, b_{n,t_i}^{a_j}$ is suspected. (c) The illustration of deep learning analysis. In this example, $e_i, \cdots, e_n$ are simulated futures. By training a deep learning model based on simulated data, it will be able to detect anomalies like $b_{i,t_j}^{a_k}, b_{n,t_i}^{a_j}$ and further predict $e_i, e_n$ before $e_i, e_n$ happen.