# One Fuzzer, Many Reasoners: Adaptive Neuro-Symbolic Fuzzing with Reliability-Calibrated and Context-Gated Guidance

DevFeed: [One Fuzzer, Many Reasoners: Adaptive Neuro-Symbolic Fuzzing with Reliability-Calibrated and Context-Gated Guidance](<https://devfeed.tech/articles/one-fuzzer-many-reasoners-adaptive-neuro-symbolic-fuzzing-with-reliability-calibrated-and-context-gated-guidance-132208.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.16211>)

Author: Jinghao Wen, Xianglong Yu, Shaohuang Wang, Xun Jiao

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Fuzzing/Fuzz testing](<https://devfeed.tech/topics/fuzzing.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [constraint](<https://devfeed.tech/tags/constraint.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [discovery](<https://devfeed.tech/tags/discovery.md>), [effective](<https://devfeed.tech/tags/effective.md>), [failure](<https://devfeed.tech/tags/failure.md>), [framework](<https://devfeed.tech/tags/framework.md>), [fuzzing](<https://devfeed.tech/tags/fuzzing.md>), [grounding](<https://devfeed.tech/tags/grounding.md>), [guidance](<https://devfeed.tech/tags/guidance.md>), [model](<https://devfeed.tech/tags/model.md>), [mutation](<https://devfeed.tech/tags/mutation.md>), [neural](<https://devfeed.tech/tags/neural.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [reliability](<https://devfeed.tech/tags/reliability.md>), [risk](<https://devfeed.tech/tags/risk.md>), [semantics](<https://devfeed.tech/tags/semantics.md>), [signal](<https://devfeed.tech/tags/signal.md>), [systems](<https://devfeed.tech/tags/systems.md>)

## AI overview

NSAI-Fuzz uses a shared, semantics-aware risk interface to guide fuzzing across heterogeneous neural-symbolic AI systems. It combines five availability-aware signals--boundary pressure, grounding uncertainty, representation drift, constraint violation, and archive novelty--with task-specific valid mutation. Reliability-Calibrated Weighting (RCW) assigns dataset-model-level signal weights using calibrated reliability, while Contextual Signal Gating (CSG) adapts weights to the current candidate pool. In the reported evaluation across four systems and four reasoning tasks, fixed five-signal guidance beat Random and Uniform in every compatible cell, and RCW and CSG improved over Fixed in all nine cells. With matched mutation and query budgets, they found 16.17% and 16.06% aggregate failures, versus 6.06% for the strongest neural-fuzzing baseline. TensorFuzz used 1.11 times RCW's runtime while delivering 0.45 times its failure throughput.

## Source excerpt

arXiv:2611.16211v1 Announce Type: new Abstract: Neural-symbolic AI (NSAI) combines learned perception with programs, constraints, and compositional reasoning, but remains difficult to test because small semantics-preserving changes can alter grounded symbols and trigger discontinuous reasoning failures. Existing neural fuzzers mainly rely on neural coverage or representation-space feedback and do not provide a common testing interface for heterogeneous NSAI systems. We introduce NSAI-Fuzz, a unified semantics-aware fuzzing framework that separates task-specific valid mutation from architecture-neutral guided search and represents each candidate using five availability-aware, rank-normalized risk signals: boundary pressure, grounding uncertainty, representation drift, constraint violation, and archive novelty. We further propose Reliability-Calibrated Weighting (RCW), which learns dataset-model-level signal allocations from calibrated reliability, and Contextual Signal Gating (CSG), which adapts signal weights to the current candidate pool. We evaluate four NSAI systems across four reasoning tasks, nine compatible dataset-model cells, 20 policies, and 1,800 on-policy runs. Fixed five-signal guidance outperforms Random and Uniform in every cell, while RCW and CSG further improve over Fixed in all nine cells. Under matched mutation and query budgets, RCW and CSG achieve 16.17% and 16.06% aggregate failure discovery, compared with 6.06% for the strongest neural-fuzzing baseline. Wall-clock evaluation further shows that TensorFuzz requires 1.11x the runtime of RCW while achieving only 0.45x its failure throughput. These results show that effective and efficient fuzzing of heterogeneous NSAI systems benefits from a common neural-symbolic risk interface with adaptive signal allocation.