Genliang Zhu

Genliang Zhu

Auto-generated profile, based on publicly available information

Genliang Zhu


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
GZGenliang Zhu
Genliang Zhu
CWChu Wang
Chu Wang
Tool-using AI agents commonly operate under integration credentials whose static permissions exceed a user's current request. We present Intent-Governed Access Control (IGAC), a server-side authorization layer that converts a trusted user request into a short-lived intent certificate, monotonically filters the statically authorized tool manifest, and checks proposed tool and payload effects before execution. The model and certificate generator remain untrusted: IGAC cannot grant authority outside static policy, and request-level confinement is conditional on certificate and effect-bound fidelity. We formalize these guarantees, implement a reusable IGAC path over an OpenPort governance substrate, and evaluate the mechanism using endpoint tests, 176 runtime-backed synthetic tasks, real-model classifier and planner pilots, 306 end-to-end model-task runtime trials, and a 36-trial benchmark-shaped external subset. In the deterministic runtime comparison, author-reference IGAC reduces the archived composite exposure-or-path indicator from 1.0000 to 0, while both OpenPort-only and IGAC observe zero completed unsafe executions. In the end-to-end model runs, the combined IGAC-OpenPort path again observes zero completed unsafe executions, but unsafe accepted authority remains 0.0909-0.2727; strict benign completion is 0, while the governed-artifact rate is 0.25-0.50. Every residual unsafe acceptance is a non-executed draft. A trace-backed normalizer counterfactual removes this residual authority at substantial utility cost. These results support mechanism-level static-policy non-expansion and expose certificate precision, rather than model intent classification alone, as the principal remaining bottleneck. They do not establish production robustness or deployment-level safety improvement.

16 October 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
GZGenliang Zhu
Genliang Zhu
ZWZiyuan Wang
Ziyuan Wang
XCXiaobo Chen
Xiaobo Chen
CWChu Wang
Chu Wang
An agent awaiting approval can retain reusable computation while its business evidence or approval becomes obsolete. We formulate online context scheduling with separate computation, evidence, and approval lifetimes, and implement retention, lossless externalization, version-addressed retrieval, and rebuilding under a common admission guard. Seven capability-matched controls are evaluated on 45,360 finite assignments and 1,008 assignments from a different event process. Joint forecasting improves main-grid on-time completion by 1.67 percentage points at 4.35% more modeled work; the alternative event process favors a simpler control. A separate actual-clock study assigns 240 episodes to four controllers and two local CPU models. It records 32 approved current-version SQLite commits, 160 timeouts, and 48 cancellations. All four controllers complete 3/18 held-out Qwen tasks and 2/18 SmolLM2 tasks, while their preparation trajectories and token costs differ. Of 83,829 prompt tokens, native telemetry identifies 3,474 reused cache tokens. Independent replay retains every allocation and checks version, approval, deadline, and complete costs. The combined evidence establishes a measurable contract for scheduling reusable context and shows how controller complexity, event exposure, and end-to-end completion interact under declared clocks and resource conditions.

16 October 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
GZGenliang Zhu
Genliang Zhu
ZWZiyuan Wang
Ziyuan Wang
XCXiaobo Chen
Xiaobo Chen
CWChu Wang
Chu Wang
A change of working language can preserve words while changing the entity, numeric condition, or effect of an agent's tool action. We study source-bound contract checks that separate representation validity, authorization, and task completion. Independent replay closes 432 controlled executions; on designated English-path counterexamples, semantic checks achieve 35/36 strict completions with no forbidden effects, compared with 25/36 completions and nine forbidden histories under schema or literal checks. A separate two-model study evaluates source extraction, working representations, reconstruction, and planning across nine matched policies. It closes 1,152 workflows over 16 authored roots, including 864 held-out workflows with paired English and Chinese instructions under clean and injected-binding-fault channels. In each 24-workflow Llama condition, local restoration completes six normal tasks in the clean channel and five under stress, with no observed forbidden effects; schema-only stress admits forbidden effects in 18/24. Qwen produces no normal completions across the held-out policies, distinguishing guarded execution from planning capability. Local restoration and fixed English tie on clean Llama completion, while local restoration consumes more input tokens. These results identify when contract checks improve source compliance and why a routing-efficiency claim requires a separate, equally protected comparison.

15 October 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
GZGenliang Zhu
Genliang Zhu
XCXiaobo Chen
Xiaobo Chen
ZWZiyuan Wang
Ziyuan Wang
CWChu Wang
Chu Wang
Enterprise AI agents can reach the intended business state while relying on outdated memory, mishandling approval waits or reporting completion without sufficient observations. We develop an executable lifecycle measurement contract that separates goal state, historical compliance and declaration-time evidence. A 160-episode scripted census verifies these distinctions in persistent synthetic business worlds. A separate prospectively specified local-model study runs 1,568 tool-using episodes over 98 paired instances from 14 held-out domain templates, crossing memory freshness with approval delay and partial effects with receipt visibility. Its prespecified endpoint combines correct goals, compliant effects and final-report consistency. For gpt-oss-20b, stale memory and long approval waits reduce completion, with a positive interaction of +0.214 (Holm-adjusted template sign-flip 𝑝 = 0.002): premature abandonment overlaps with stale-account payment failures. gemma-4-12b completes 783 of 784 episodes while its refund recovery effort increases in turns and tokens. An additional exploratory observation audit distinguishes accurate reports from observed confirmation: 60 of 96 correct gpt-oss-20b completions under partial refunds and delayed receipts lack an observed confirmation covering all required items. These findings show why lifecycle evaluation needs separate state, process, report and evidence measures, and how interacting conditions can conceal failure routes in a single terminal score.

14 October 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
GZGenliang Zhu
Genliang Zhu
XCXiaobo Chen
Xiaobo Chen
ZWZiyuan Wang
Ziyuan Wang
CWChu Wang
Chu Wang
In text-to-SQL and AI-agent analytics, successful SQL execution does not establish semantic correctness under changing business definitions. We study revalidation of fixed queries after changes to refund attribution, tax treatment, eligible populations and exchange-rate dates. An executable verification profile binds the requested revision and knowledge cutoff to legal counterexample data and exact monetary comparison. A preregistered confirmation on 32 newly authored revision families and two local producer bundles retains 192 query–definition cells: 144 nonconforming, 36 finite-suite agreeing and 12 unresolved. Of the nonconforming cells, 63 agree on the current snapshot. At an eight-ledger budget, obligation-guided checking detects 137 discrepancies versus 127 for generic ordering, a paired difference of 6.9 percentage points (family bootstrap 95% interval [2.1, 13.6]; exact family sign-flip p=0.0156). Every ledger arm rejects zero of the 36 agreeing cells. Current-obligation and equal-weight scoring match the guided result, identifying coverage rather than revision-specific priority as the observed source of benefit. Controlled assays, a natural-query pilot and repeated samples provide separate mechanism and robustness evidence. A legal discrepancy refutes conformance; agreement remains restricted to the recorded suite. The study establishes an auditable business-semantic revalidation profile within its authored schema and local-model scope.

13 September 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
GZGenliang Zhu
Genliang Zhu
By acquiring compute, credentials, accounts, services, and other agents, autonomous AI agents can introduce new authority into a task. Payment, budget, OAuth, mandate, and fulfillment checks can validate transaction conditions without deciding whether a returned resource may become usable authority. This post-fulfillment activation gap spans tool-mediated creation, inter-agent delegation, and agentic commerce. We present AcquireBound, a provenance-bounded runtime authorization architecture. It quarantines acquired outputs, resolves their actual capabilities from authenticated provider evidence through a versioned resolver, and activates them only through a current activation transaction that checks the resolved manifest, provenance, epochs, and a downward-closed relational envelope over a typed resource-capability hypergraph. The envelope preserves correlated identity, effect, data, delegation, and graph-wide limits. Single-use effect permits are revalidated and consumed at effect linearization. Under explicit assumptions, we prove eight safety properties covering quarantine, backing, non-amplification, split non-evasion, crash/retry, refunds, epochs, and effect confinement. Across five resource classes, reference semantics accepted 20/20 benign traces and rejected 40/40 registered unsafe traces over 810 events; an independent checker agreed on 60 base and 40 refinement traces and rejected 89/89 tamper tests. Frozen Codex and Gemini Model Context Protocol (MCP) client components completed 54/54 deterministic local stdio calls. In a registered 18-case staged MCP-to-Docker composition, both benign paths completed, and none of the 16 unsafe paths added an unauthorized Docker start request. A five-source audit classified 1,248 field pairs across 32 units; no unit alone supplied a complete activation profile.