Evolutionary Intelligence for Scientific Discovery: From Evolutionary Computation to Cumulative Discovery Systems
TLDR
Introduces evolutionary intelligence (EI) as a framework for cumulative scientific discovery, extending evolutionary computation with experience retention.
Reasoning
The paper provides a conceptual framework and review, but lacks real-world experiments or empirical evaluations. Its strength lies in bridging evolutionary computation and cumulative discovery, but it does not present concrete implementations or benchmarks.
Read-first score
Read-first score 50.6, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 57.
Field roles
Rank sensitivity
Stability: volatile; rank range: 58.
Keyword Scores
Deep Analysis
Innovations
- Introduction of evolutionary intelligence (EI) as a concept bridging evolutionary computation and cumulative scientific discovery systems
- A five-dimensional analytical framework (what evolves, how, why, where, when) to characterize EI systems
- Characterization of EI systems that sustain exploration by linking candidate refinement with experience retention
- Demonstration of the EI paradigm across diverse discovery modes from evolving concrete entities to orchestrating automated workflows
- Identification of critical bottlenecks (evaluation, process traceability, shared infrastructure) and a roadmap for transition
Methodology
This is a review paper that proposes the concept of evolutionary intelligence and introduces a five-dimensional analytical framework. It analyzes scientific AI systems through the lens of feedback-driven candidate refinement combined with experience retention, and discusses diverse discovery modes and bottlenecks.
Key Results
The framework clarifies how evolutionary intelligence transforms isolated search trajectories into cumulative scientific insight. The paper identifies critical bottlenecks in evaluation, process traceability, and shared infrastructure, and provides a roadmap for advancing from evolutionary computation to cumulative discovery systems.