arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2610.04275

Symbolic Search Is Not Exhausted: Persistent Proof-Space Exploration in Lean4

Ruoran Xu
View original source ↗Read source document ↗Export metadata ↓

Abstract

Formal theorem proving increasingly combines learned semantic guidance with verified symbolic execution. The quality of the symbolic search substrate therefore determines how much useful mathematical structure can be accumulated, reused, and exposed under a finite inference budget. We introduce ViaLean, a Lean4 prover that organizes symbolic reasoning as bounded exploration of a persistent proof-state graph. Its search preserves coverage across complementary reasoning modes, merges semantically equivalent goals, retains verified intermediate structure, and observes short symbolic futures before committing to a transition. On the complete miniF2F test split, the model-free configuration solves 122/244 problems (50.0% pass@1). Historical and recent symbolic reference points range from Lean's earlier tidy search to modern grind and SMT-backed verification, showing that proof-space organization remains a substantial source of capability. The same persistent state also provides a natural interface for neural--symbolic agents: neural reasoning can operate over verified regions and intermediate objects while Lean continuously expands and validates the local proof space.

Explore related subjects

cs.LO

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Ruoran Xu. 2026-10-03. Symbolic Search Is Not Exhausted: Persistent Proof-Space Exploration in Lean4. https://arxiv.org/abs/2610.04275

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Coslice Colimits in Homotopy Type Theory

We contribute to the theory of (homotopy) colimits inside homotopy type theory. The heart of our work characterizes the connection between (graph-indexed) colimits in a type universe and colimits in coslices of the universe, called coslice colimits. To derive this characterization, we give a construction of coslice colimits that is tailored to reveal the connection. We use the construction to prove that the forgetful functor from a coslice creates colimits over trees. We also use it to study how coslice colimits interact with orthogonal factorization systems and with cohomology theories. As a result of their interaction with orthogonal factorization systems, all colimits of pointed types preserve $n$-connectedness, which implies that higher groups, in the sense of Buchholtz, van Doorn, and Rijke, are closed under colimits. We have formalized major portions of this work (see https://github.com/PHart3/colimits-agda for the Agda code), including our main construction of the coslice colimit functor.

cs.LO↗
arXiv2026

Non-Wellfounded and Cyclic Proofs for LTL: A Syntactic Correspondence with Linear Nested Sequents

We introduce the formalism of non-wellfounded and cyclic linear nested sequent calculi, developing concrete systems for linear temporal logic (LTL). The paper addresses two central problems, which we call "cycle recognition" and "unraveling." Cycle recognition concerns identifying cycles in non-wellfounded proofs in order to extract corresponding cyclic proofs, while unraveling studies the converse transformation, from cyclic proofs to non-wellfounded ones. Although these processes are well understood for Gentzen sequents, they have received little attention for more expressive sequent formalisms and become more challenging in the linear nested sequent setting. To address cycle recognition, we show the completeness of non-wellfounded proofs relative to a particular normal form exhibiting a property we call "saturation recurrence," which enables the systematic extraction of cyclic proofs. To address unraveling, we introduce a specialized procedure that shifts rule applications forward along linear nested sequents, allowing non-wellfounded proofs to be reconstructed from cyclic ones. Overall, our work provides new proof-theoretic techniques for cycle recognition and unraveling in expressive multisequent formalisms.

cs.LO↗
arXiv2026

Decreasing Diagrams are Complete for Confluence

Confluence is a fundamental property of nondeterministic computations, arising from parallelism, concurrency, or freedom in the evaluation order. It guarantees that such a computation always yields the same result, regardless of the order in which steps are taken. The decreasing diagrams technique of van Oostrom is one of the most versatile methods for establishing confluence of transition systems (abstract rewriting systems). It reduces global confluence to local confluence: a system is confluent whenever its transitions admit a locally decreasing labeling. Essentially all classical confluence criteria arise as corollaries. A central question, posed by van Oostrom in 1993, asks whether the decreasing diagrams technique is complete: Does every confluent transition system admit a locally decreasing labeling? This is Problem 56 of the RTA List of Open Problems. A positive answer was known only for countable systems, and recently up to the first uncountable cardinal $\aleph_1$. The general case remained open. We settle this thirty-three-year-old problem in full. We prove that every confluent transition system admits a locally decreasing labeling using only three labels. This bound is optimal, as two labels do not suffice even at the first uncountable cardinality. It follows that this single criterion can, in principle, certify the confluence of every confluent system, and hence of every confluent program. The entire development is machine-checked in the Isabelle/HOL and Lean proof assistants and relies only on classical logic and the axiom of choice.

cs.LO↗