AppleSeed: Intent-Based Multi-Domain Infrastructure Management via Few-Shot Learning
Bibliographic record
Abstract
Managing complex infrastructures in multi-domain settings is time-consuming and error-prone. Intent-based infrastructure management is a means to simplify management by allowing users to specify intents, i.e., high-level statements in natural language, that are automatically realized by the system. However, providing intent-based multi-domain infrastructure management poses a number of challenges: 1) intent translation; 2) plan execution and parallelization; 3) incompatible cross-domain abstractions. To tackle these challenges, we propose AppleSeed, an intent-based infrastructure management system that enables an end-to-end intent-to-deployment pipeline. AppleSeed uses few-shot learning for training a Large Language Model (LLM) to translate intents into intermediate programs, which are processed by a just-in-time compiler and a materialization module to automatically generate parallelizable, domain-specific executable programs. We evaluate the system in two use cases: Deep Packet Inspection (DPI); and machine learning training and inferencing. Our system achieves efficient intent translation into an execution plan with an average 22.3x lines of code to intent word ratio. It also speeds up the execution of the management plan by 1.7-2.6 times with our JIT compilation for parallelized execution compared to sequential execution.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".