arXiv subjects

Thomson D. Nguy

Publications and source records attributed to Thomson D. Nguy.

3 recordsLinked to original sources

arXiv2026

Which Rules Matter Now? Policy-Centroid Routing Before an Intelligent System Acts

Before an intelligent system can decide whether an action is allowed, it must first know which rules the action has approached. A single proposed action can implicate several policy regimes at once. Their requirements may stack, overlap, or qualify one another, yet many remain written in natural language while the action itself arrives as an incomplete description of intent. The first problem is not judgment. It is attention. Policy-centroid routing creates a layer before adjudication. It compresses expressions within each policy regime into one or more representative centroids, places the proposed action in the same semantic space, applies a declared measure, and routes every regime crossing a declared threshold to authoritative review. Several regimes may trigger at once. The output is a review agenda, not permission, prohibition, legality, breach, compliance, certification, or enforcement. The paper develops six falsifiable propositions and seven follow-on studies comparing the hypothesis with structured workflows, lexical and semantic retrieval, hierarchical and direct classification, and selective prediction under matched review burden. The studies are designed to identify where policy geometry recovers applicable regimes, where compression loses rare or overlapping obligations, and where the mechanism should abstain. The paper includes a synthetic worked example and reports no empirical efficacy result.

cs.AI↗

arXiv2026

More Context, Same Budget: Dual-Bounded Relational Recall Beyond Top-K Retrieval

More context does not require a larger retrieval budget. Under the same ceiling, a retrieval system can recover more of the evidence a question requires by following relationships between evidence that flat top-k ranking leaves behind. We test that proposition with Dual-Bounded Relational Recall (DBRR), which allocates a fixed retrieval budget between relevance-selected seeds and bounded graph-adjacent context, against matched flat top-k retrieval using the same relevance-ranking stage and the same maximum number of retrieval units and tokens. The outcome is complete recovery of the official HotpotQA supporting-evidence set for each question. Across 7,405 FullWiki questions, the Primary DBRR allocation increased complete supporting-evidence recovery by 23.8 percentage points over its matched flat baseline (paired risk difference 0.2377; question-level bootstrap 95% interval 0.2269 to 0.2489). It improved 1,952 questions, tied on 5,261, and harmed 192. Bridge questions drove the effect, with a 28.7-point increase; comparison questions showed a smaller 4.2-point difference. In a prespecified, evaluation-only diagnostic population, real relationships also outperformed random-neighbor and degree-preserving shuffled-graph controls. The result is straightforward: under the same context budget, complete-evidence retrieval depends not only on which items rank highest, but on how context is allocated around them. Relational allocation recovered complete evidence sets that flat top-k retrieval left incomplete.

cs.IR↗

arXiv2026

Allostatic Control Systems: Goal Governance in Changing Environments

Allostatic control systems govern not only how a system pursues a goal, but whether the goal itself remains appropriate as the environment changes. This matters whenever a controller can continue to regulate successfully against a reference that no longer serves the system. We develop this as a two-timescale problem: a fast loop regulates under the current reference, while a slow loop governs whether that reference should move. We then test one candidate mechanism in which slow-loop movement waits for mature outcome evidence. In a preregistered synthetic experiment, that mechanism increased decision cost by 1.51% relative to otherwise identical bounded adaptation. The central failure was timing: the evidence-to-effect pathway often could not make correction effective before the environment changed again. The result motivates a broader design principle: an allostatic controller must be able to revise an inappropriate goal faster than serviceability is lost by continuing to defend it. We therefore treat allostatic control systems as a general engineering problem of goal governance under changing environments, and the negative result as one step toward making that problem operational and falsifiable.

eess.SY↗