{"slug": "inductive-deductive-synthesis-enabling-ai-to-generate-formally-verified-systems", "title": "Inductive Deductive Synthesis: Enabling AI to Generate Formally Verified Systems", "summary": "Researchers have developed Inductive Deductive Synthesis (IDS), an AI system that jointly generates formally verified code and mathematical proofs, solving a long-standing challenge in automated software verification. In tests on distributed key-value-store specifications, IDS succeeded on all seven tasks where state-of-the-art coding agents failed on five, completing each specification in about 6.8 hours at $106 per spec—roughly 200 times faster than human experts and 17% cheaper than existing AI agents. The system also optimizes for performance within the same loop, producing verified implementations up to three times faster than previously published systems.", "body_md": "arXiv:2605.23109v1 Announce Type: new\nAbstract: AI agents increasingly excel at generating, testing, and refining code. However, they fall short on tasks requiring formal guarantees of full coverage that testing alone cannot provide. Distributed systems are a prime example: properties such as consistency between reads and writes must hold under every possible interleaving of events. Mechanized formal verification can guarantee such correctness, but typically demands months to years of expert effort. As evidence, even SOTA coding agents (Codex with GPT-5.4 and Claude Code with Opus 4.6) succeed on only 2/7 distributed key-value-store specifications. In this paper, we present the first effective approach to addressing this gap, Inductive Deductive Synthesis (IDS), which jointly and incrementally synthesizes implementation and proof, and learns from failed attempts to systematically try promising strategies. Built as an agentic LLM system, IDS achieves 7/7 in about 6.8 hours and $106 per spec on average, roughly 200x faster than expert effort and 17% cheaper than SOTA agents. IDS further incorporates performance feedback into the same loop, yielding implementations up to 3x faster than published verified systems.", "url": "https://wpnews.pro/news/inductive-deductive-synthesis-enabling-ai-to-generate-formally-verified-systems", "canonical_source": "https://arxiv.org/abs/2605.23109", "published_at": "2026-05-25 04:00:00+00:00", "updated_at": "2026-05-25 15:16:32.995722+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-agents", "ai-research"], "entities": ["Codex", "GPT-5.4", "Claude Code", "Opus 4.6", "Inductive Deductive Synthesis", "IDS"], "alternates": {"html": "https://wpnews.pro/news/inductive-deductive-synthesis-enabling-ai-to-generate-formally-verified-systems", "markdown": "https://wpnews.pro/news/inductive-deductive-synthesis-enabling-ai-to-generate-formally-verified-systems.md", "text": "https://wpnews.pro/news/inductive-deductive-synthesis-enabling-ai-to-generate-formally-verified-systems.txt", "jsonld": "https://wpnews.pro/news/inductive-deductive-synthesis-enabling-ai-to-generate-formally-verified-systems.jsonld"}}