neuro-symbolic

Proves the long-standing conjecture W(2;3,21) = 416 by Kullmann (2010) using a Lean 4 kernel-checked proof over 209 cases and 38.6M leaf formulas, produced by the neuro-symbolic Large Proof Model (LPM) with a 13.6 GB verifiable evidence package.

This paper argues for layered verification of LLM-generated argument graphs, combining deterministic checks, solvers, and LLM reviewers, and shows that LLM reviewers detect structural defects well but are inconsistent, over-reject correct graphs, and slower than exact checks.

ACED-Bench introduces a neuro-symbolic benchmark combining natural-language scientific scenarios with hidden probabilistic graphical models to evaluate whether LLM agents can actively design experiments, gather causal evidence, and make correct causal decisions across 180 core tasks and a 150-question qualification set.

The paper introduces SAE-IDF, which uses sparse autoencoders and rarity-weighted features as a shared coordinate system to match symbolic units across heterogeneous systems, outperforming dense LLM baselines on 12 tasks while remaining human-interpretable.

A generatively informed neuro-symbolic framework combines syntactic theory with AraBERT to resolve structural ambiguity in Modern Standard Arabic noun phrases, achieving 96.88% accuracy on unseen data. The study shows formal linguistic structure can be operationalized as an interpretable interface in Transformer-based NLP.

Law&Order is a neuro-symbolic framework that combines LLM synthesis with cell-level symbolic verification and iterative error repair to automatically formalize tax forms into executable programs, achieving 100% accuracy on 51 held-out TaxCalcBench returns compared to 66% for the best LLM alone.

Introduces Hillock, an open-source engine that uses a neuro-symbolic approach for long-term memory in local agents, improving efficiency with techniques like fact extraction and hyperdimensional gating.

Introduces the Epistemic Probabilistic Language Agents (EPLA) model, a neuro-symbolic architecture for enhancing multi-agent LLM coordination by integrating epistemic logic, probabilistic belief engines, and symbolic guards to ensure protocol-governed actions under uncertainty.

This paper presents Signal2Symbol, a neuro-symbolic framework for explainable anomaly detection in physiological time-series data like ECG and EEG, leveraging symbolic tokenization and temporal reasoning to produce interpretable explanations of anomalous patterns.

LogicTrack is a neuro-symbolic framework that audits reasoning trajectories of large language models using formal logic solvers to ensure logical validity, improving both reasoning chain verifiability and final answer accuracy.

A fully differentiable neuro-soft-symbolic framework integrates visual perception and task planning in a single computational graph, achieving higher success rates on benchmarks like Blocksworld and PlanBench with reduced computation.

This paper introduces a neuro-symbolic Hierarchical Planning Decoder (HPD) for anticipating human intentions from multimodal episodes. It demonstrates improved performance in goal inference and logic constraint satisfaction on a benchmark dataset.

This paper presents a self-demonstration-driven approach for schema-ontology mapping using LLMs, combining neuro-symbolic task decomposition to achieve state-of-the-art performance on the RODI benchmark.

This paper introduces VERDICT, a neuro-symbolic system that uses LLMs and SMT solvers to ensure accountable decision-making in clinical trial matching, providing consistent policies and faithful rationales.

NSIDDx is a neuro-symbolic design framework for differential diagnosis in low-resource settings that prioritizes clinician collaboration and transparent reasoning, addressing gaps in LLM-based systems by ensuring verifiable outputs and auditability.

ProofEvolve is a neuro-symbolic framework that evolves formally verified proof structures using neural models to enhance automated theorem proving, achieving high solve rates on Lean benchmarks by preserving verified knowledge from incomplete attempts.

The paper proposes a neuro-symbolic framework that decouples reasoning into symbolic validity and semantic groundedness, using a verifier and a trained PRM to improve reliability in scientific reasoning tasks for LLMs.

Hillock v0.5.0 is a lightweight neuro-symbolic memory engine for AI agents that manages persistent memory efficiently using structured triples in SQLite and hypervectors, designed for local setups with limited VRAM.

The paper proposes a neuro-symbolic pipeline for pairwise logical selection of enthymeme completions, introducing the PWAL method that uses logical-resistance scores to improve accuracy and reduce ties across various reasoning tasks.

The paper introduces SDDL, a neuro-symbolic framework that improves combinatorial optimization accuracy in resource-constrained language models by translating natural-language problems into formal representations, resulting in higher feasibility rates compared to direct-generation and solver-code baselines.