AMD+: A Design-Time Model-Selection Gate for Mixed-Intelligence Systems

Langiu (2026) introduced AMD, a specification language and methodology that types every service of a system by its intelligence mix and marks which services are served by a language model. We describe AMD+, the deterministic gate that turns such a specification into a model allocation before any code runs. For every model-served service the gate infers a capability-requirement vector from declared fields of the specification (interactions, contracts, fail modes, tools, counter-examples, declared payload size, provider tier), and matches it against a catalogue of candidate models whose cells carry a provenance mark. The cheapest model that meets the requirement on every axis is selected; when none does, the gate returns the closest model and the violated axes, which drives the variation loop of AMD. We give the inputs, the axes and their scale, every inference rule, the catalogue schema, the selection algorithm and a two-service example.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-02
DOI
https://doi.org/10.5281/zenodo.23885888
Primary Topic
Advanced Software Engineering Methodologies
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

AMD+: A Design-Time Model-Selection Gate for Mixed-Intelligence Systems

Alessio Langiu
2 links
Zenodo (CERN European Organization for Nuclear Research)
Advanced Software Engineering Methodologies
preprint

AMD+: A Design-Time Model-Selection Gate for Mixed-Intelligence Systems

Alessio Langiu
preprint en

Abstract

Langiu (2026) introduced AMD, a specification language and methodology that types every service of a system by its intelligence mix and marks which services are served by a language model. We describe AMD+, the deterministic gate that turns such a specification into a model allocation before any code runs. For every model-served service the gate infers a capability-requirement vector from declared fields of the specification (interactions, contracts, fail modes, tools, counter-examples, declared payload size, provider tier), and matches it against a catalogue of candidate models whose cells carry a provenance mark. The cheapest model that meets the requirement on every axis is selected; when none does, the gate returns the closest model and the violated axes, which drives the variation loop of AMD. We give the inputs, the axes and their scale, every inference rule, the catalogue schema, the selection algorithm and a two-service example.

Zenodo (CERN European Organization for Nuclear Research)
Istituto di Scienze Marine del Consiglio Nazionale delle Ricerche (IT), National Research Council (IT)
Advanced Software Engineering Methodologies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon