{"slug": "pareto-optimal-rtl-code-generation-via-multi-objective-reinforcement-learning", "title": "Pareto-Optimal RTL Code Generation via Multi-Objective Reinforcement Learning with Large Language Models", "summary": "A new arXiv paper (2610.26904v1) proposes a multi-objective reinforcement learning method that trains large language models to generate sets of Pareto-optimal RTL circuits from natural-language specifications rather than a single design. The method combines a generation-based training pipeline with a Hypervolume Contribution (HVC) reward function that scores how much each generated circuit adds to the Pareto front across power, performance and area trade-offs. On the RTLLM v2 benchmark, the approach gains 13.3 percentage points in relative Hypervolume over a baseline without PPA optimization and 14.4 percentage points over a single-objective reinforcement learning method, while also showing better circuit diversity.", "body_md": "arXiv:2610.26904v1 Announce Type: new \nAbstract: Recent studies have proposed reinforcement learning with verifiable rewards (RLVR) methods that enable LLMs to generate high-quality RTL circuits from natural-language specifications. Existing approaches optimize circuit PPA (Power, Performance, Area) by aggregating multiple objectives into a single scalar reward, making it difficult to simultaneously generate diverse Pareto-optimal solutions across different trade-offs such as low power, low latency, and small area. However, in practical design, the relative importance of each objective varies depending on product requirements and design constraints; thus, obtaining a diverse and high-quality candidate set is beneficial in that designers can choose a preferred trade-off a posteriori. In this work, we are the first to propose a learning method for LLM-based circuit generation that optimizes, not the quality of a single circuit, but the quality and diversity of the entire circuit set obtained through multiple generations in the PPA objective space. The proposed method consists of (1) a generation-based training pipeline that reformulates coverage improvement of the entire Pareto solution set into a sequential circuit generation problem, and (2) a reward function based on Hypervolume Contribution (HVC) that measures how much a generated circuit contributes to the Pareto front. Evaluation on the RTLLM v2 benchmark shows that the proposed method achieves +13.3 percentage points in relative Hypervolume over a baseline without PPA optimization and +14.4 percentage points over a single-objective reinforcement learning method, while also demonstrating superiority in circuit diversity.", "url": "https://wpnews.pro/news/pareto-optimal-rtl-code-generation-via-multi-objective-reinforcement-learning", "canonical_source": "https://arxiv.org/abs/2610.26904", "published_at": "2026-10-21 04:00:00+00:00", "updated_at": "2026-10-21 04:19:48.723290+00:00", "lang": "en", "topics": ["large-language-models", "machine-learning", "ai-research", "ai-chips"], "entities": ["RTLLM v2", "arXiv", "Hypervolume Contribution"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/pareto-optimal-rtl-code-generation-via-multi-objective-reinforcement-learning", "markdown": "https://wpnews.pro/news/pareto-optimal-rtl-code-generation-via-multi-objective-reinforcement-learning.md", "text": "https://wpnews.pro/news/pareto-optimal-rtl-code-generation-via-multi-objective-reinforcement-learning.txt", "jsonld": "https://wpnews.pro/news/pareto-optimal-rtl-code-generation-via-multi-objective-reinforcement-learning.jsonld"}}