dorsal/arxiv
View SchemaImproving Zero-shot ADL Recognition with Large Language Models through Event-based Context and Confidence
| Authors | Michele Fiori, Gabriele Civitarese, Marco Colussi, Claudio Bettini |
|---|---|
| Categories | |
| ArXiv ID | 2601.08241vv1 |
| URL | https://arxiv.org/abs/2601.08241 |
| License | http://creativecommons.org/licenses/by-nc-sa/4.0/ |
Abstract
Unobtrusive sensor-based recognition of Activities of Daily Living (ADLs) in smart homes by processing data collected from IoT sensing devices supports applications such as healthcare, safety, and energy management. Recent zero-shot methods based on Large Language Models (LLMs) have the advantage of removing the reliance on labeled ADL sensor data. However, existing approaches rely on time-based segmentation, which is poorly aligned with the contextual reasoning capabilities of LLMs. Moreover, existing approaches lack methods for estimating prediction confidence. This paper proposes to improve zero-shot ADL recognition with event-based segmentation and a novel method for estimating prediction confidence. Our experimental evaluation shows that event-based segmentation consistently outperforms time-based LLM approaches on complex, realistic datasets and surpasses supervised data-driven methods, even with relatively small LLMs (e.g., Gemma 3 27B). The proposed confidence measure effectively distinguishes correct from incorrect predictions.
{
"annotation_id": "e9db9fdb-89e6-4cbf-a216-9b50cbf416a4",
"date_created": "2026-02-17T05:53:15.433000Z",
"date_modified": "2026-02-17T05:53:15.433000Z",
"file_hash": "0ed853b3207fd61e91880d83a96f463a0c8d6e646446c5aff731316e265d84af",
"private": false,
"record": {
"abstract": "Unobtrusive sensor-based recognition of Activities of Daily Living (ADLs) in smart homes by processing data collected from IoT sensing devices supports applications such as healthcare, safety, and energy management. Recent zero-shot methods based on Large Language Models (LLMs) have the advantage of removing the reliance on labeled ADL sensor data. However, existing approaches rely on time-based segmentation, which is poorly aligned with the contextual reasoning capabilities of LLMs. Moreover, existing approaches lack methods for estimating prediction confidence. This paper proposes to improve zero-shot ADL recognition with event-based segmentation and a novel method for estimating prediction confidence. Our experimental evaluation shows that event-based segmentation consistently outperforms time-based LLM approaches on complex, realistic datasets and surpasses supervised data-driven methods, even with relatively small LLMs (e.g., Gemma 3 27B). The proposed confidence measure effectively distinguishes correct from incorrect predictions.",
"arxiv_id": "2601.08241",
"authors": [
"Michele Fiori",
"Gabriele Civitarese",
"Marco Colussi",
"Claudio Bettini"
],
"categories": [
"cs.CV",
"cs.DC"
],
"license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
"title": "Improving Zero-shot ADL Recognition with Large Language Models through Event-based Context and Confidence",
"url": "https://arxiv.org/abs/2601.08241",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "edf63334-b61b-4227-a3dc-17c8eb3dd5d5",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}