Lark: Biologically Inspired Neuroevolution for Multi-Stakeholder LLM Agents

TL;DR

<3-5 sentence high-level summary> Lark presents a biologically inspired, compute-aware neuroevolution framework that couples LLM-driven reasoning with a discrete-generation evolutionary loop to address multi-stakeholder trade-offs in MAS. It integrates four mechanisms—plasticity, duplication/maturation, influence-weighted Borda voting, and compute-efficiency penalties—to enable diverse strategy generation, modular specialization, and concise reasoning under compute budget. In synthetic benchmarks, Lark Full demonstrates statistically reliable gains over ablations and competitive performance versus strong baselines, with a clear understanding of each mechanism’s contribution and cost-quality trade-offs. The work argues for practical, interpretable, population-based search over RL in complex, stakeholder-driven decision settings and lays groundwork for real-world validation and broader deployment considerations.

Abstract

We present Lark, a biologically inspired decision-making framework that couples LLM-driven reasoning with an evolutionary, stakeholder-aware Multi-Agent System (MAS). To address verbosity and stakeholder trade-offs, we integrate four mechanisms: (i) plasticity, which applies concise adjustments to candidate solutions; (ii) duplication and maturation, which copy high-performing candidates and specialize them into new modules; (iii) ranked-choice stakeholder aggregation using influence-weighted Borda scoring; and (iv) compute awareness via token-based penalties that reward brevity. The system iteratively proposes diverse strategies, applies plasticity tweaks, simulates stakeholder evaluations, aggregates preferences, selects top candidates, and performs duplication/maturation while factoring compute cost into final scores. In a controlled evaluation over 30 rounds comparing 14 systems, Lark Full achieves a mean rank of 2.55 (95% CI [2.17, 2.93]) and a mean composite score of 29.4/50 (95% CI [26.34, 32.46]), finishing Top-3 in 80% of rounds while remaining cost competitive with leading commercial models ($0.016 per task). Paired Wilcoxon tests confirm that all four mechanisms contribute significantly as ablating duplication/maturation yields the largest deficit (ΔScore = 3.5, Cohen's d_z = 2.53, p < 0.001), followed by plasticity (ΔScore = 3.4, d_z = 1.86), ranked-choice voting (ΔScore = 2.4, d_z = 1.20), and token penalties (ΔScore = 2.2, d_z = 1.63). Rather than a formal Markov Decision Process with constrained optimization, Lark is a practical, compute-aware neuroevolutionary loop that scales stakeholder-aligned strategy generation and makes trade-offs transparent through per-step metrics. Our work presents proof-of-concept findings and invites community feedback as we expand toward real-world validation studies.

Figures (1)

• Figure 1: The workflow of the Lark evolutionary framework with the four core mechanisms: plasticity tweaks, stakeholder evaluation, and duplication/maturation with compute-efficiency penalties across generations.