---
title: "Trail: Accurate Local Smart Home Control Using Decision Models"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.00610v1"
permalink: "https://feed.astralor.com/items/itm_0b9575fcba2587a2be29a0fb"
published: "2026-10-31T04:13:21.000Z"
score: 56
channel: "tech"
tags: ["LLM","推理效率","Trail","Accurate Local Smart","Home Control Using","Decision Models Natural-language","Large Language Models","LLMs"]
---

# Trail: Accurate Local Smart Home Control Using Decision Models

## At a glance

Natural-language driven smart home control remains challenging: cloud-hosted Large Language Models (LLMs) raise privacy concerns, while local Small Language Models (SLMs) can generate invalid actions or misinterpret req…

## Key points

- Natural-language driven smart home control remains challenging: cloud-hosted Large Language Models (LLMs) raise privacy concerns, while local Small Language Models (SLMs) can generate invalid actions or misinterpret req…
- We introduce Trail (Typed Registry-Anchored Inference for Language commands in Smart Homes), a fully local framework that replaces open-ended command generation with hierarchical typed decisions.
- Trail exploits the natural structure of a home's action space, selecting a room, device, and operation from candidates derived from the runtime device registry.

## Full text

Natural-language driven smart home control remains challenging: cloud-hosted Large Language Models (LLMs) raise privacy concerns, while local Small Language Models (SLMs) can generate invalid actions or misinterpret requests. We introduce Trail (Typed Registry-Anchored Inference for Language commands in Smart Homes), a fully local framework that replaces open-ended command generation with hierarchical typed decisions. Trail exploits the natural structure of a home's action space, selecting a room, device, and operation from candidates derived from the runtime device registry. This ensures valid actions while enabling confidence-based abstention when requests cannot be reliably resolved. We also design a structured memory module that uses typed decisions to support user preferences, device references, and multi-turn interactions. To evaluate Trail, we extend HomeBench and generate three datasets: Trek-Base, which tests natural-language variation; Trek-Robust, which evaluates robustness to changes in the home and its description; and Trek-Memory, which assesses conversational memory across multi-turn interactions. Trail achieves 85.3% accuracy on Trek-Robust versus 69.8% for a finetuned Qwen3.5-4B baseline, while remaining competitive on Trek-Base requests. With memory, Trail resolves 75.6% of the requests that depend on earlier turns, compared with 42.5% for the finetuned baseline given conversation history.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_0b9575fcba2587a2be29a0fb