Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Leveraging LLMs for Dynamic IoT Systems Generation through Mixed-Initiative Interaction

Bassam Adnan, Sathvika Miryala, Aneesh Sambu, Karthik Vaidhyanathan, Martina De Sanctis, Romina Spalazzese

TL;DR

This work advances the IoT-Together paradigm by integrating Large Language Models into a mixed-initiative framework that enables goal-driven, runtime generation of IoT services and user interfaces. It introduces a three-pass Goal Management process, an LLM-powered Backend Generation pipeline, and a template-based Intelligent User Interface Generator grounded in Knowledge and Context Management. The Hyderabad smart-city case study and extensive experiments show accurate service identification, high-quality dynamically generated services, and efficient application generation, demonstrating practical adaptability in real-world IoT deployments. The findings highlight the potential of LLMs to enhance architectural adaptability and user-centric usability in dynamic IoT environments, while acknowledging limitations in generalizability and computational cost.

Abstract

IoT systems face significant challenges in adapting to user needs, which are often under-specified and evolve with changing environmental contexts. To address these complexities, users should be able to explore possibilities, while IoT systems must learn and support users in the process of providing proper services, e.g., to serve novel experiences. The IoT-Together paradigm aims to meet this demand through the Mixed-Initiative Interaction (MII) paradigm that facilitates a collaborative synergy between users and IoT systems, enabling the co-creation of intelligent and adaptive solutions that are precisely aligned with user-defined goals. This work advances IoT-Together by integrating Large Language Models (LLMs) into its architecture. Our approach enables intelligent goal interpretation through a multi-pass dialogue framework and dynamic service generation at runtime according to user needs. To demonstrate the efficacy of our methodology, we design and implement the system in the context of a smart city tourism case study. We evaluate the system's performance using agent-based simulation and user studies. Results indicate efficient and accurate service identification and high adaptation quality. The empirical evidence indicates that the integration of Large Language Models (LLMs) into IoT architectures can significantly enhance the architectural adaptability of the system while ensuring real-world usability.

Leveraging LLMs for Dynamic IoT Systems Generation through Mixed-Initiative Interaction

TL;DR

This work advances the IoT-Together paradigm by integrating Large Language Models into a mixed-initiative framework that enables goal-driven, runtime generation of IoT services and user interfaces. It introduces a three-pass Goal Management process, an LLM-powered Backend Generation pipeline, and a template-based Intelligent User Interface Generator grounded in Knowledge and Context Management. The Hyderabad smart-city case study and extensive experiments show accurate service identification, high-quality dynamically generated services, and efficient application generation, demonstrating practical adaptability in real-world IoT deployments. The findings highlight the potential of LLMs to enhance architectural adaptability and user-centric usability in dynamic IoT environments, while acknowledging limitations in generalizability and computational cost.

Abstract

IoT systems face significant challenges in adapting to user needs, which are often under-specified and evolve with changing environmental contexts. To address these complexities, users should be able to explore possibilities, while IoT systems must learn and support users in the process of providing proper services, e.g., to serve novel experiences. The IoT-Together paradigm aims to meet this demand through the Mixed-Initiative Interaction (MII) paradigm that facilitates a collaborative synergy between users and IoT systems, enabling the co-creation of intelligent and adaptive solutions that are precisely aligned with user-defined goals. This work advances IoT-Together by integrating Large Language Models (LLMs) into its architecture. Our approach enables intelligent goal interpretation through a multi-pass dialogue framework and dynamic service generation at runtime according to user needs. To demonstrate the efficacy of our methodology, we design and implement the system in the context of a smart city tourism case study. We evaluate the system's performance using agent-based simulation and user studies. Results indicate efficient and accurate service identification and high adaptation quality. The empirical evidence indicates that the integration of Large Language Models (LLMs) into IoT architectures can significantly enhance the architectural adaptability of the system while ensuring real-world usability.

Paper Structure

This paper contains 18 sections, 19 equations, 5 figures, 11 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Motivating Case Study
  3. Approach
  4. Goal Management
  5. Backend Generation
  6. Intelligent User Interface Generator
  7. Context Management
  8. Knowledge Management
  9. Experiments and Results
  10. Evaluation Setup
  11. Tourist-Guide Simulations
  12. User Evaluation
  13. Results & Discussions
  14. Performance Analysis
  15. Service Generation Analysis
  16. ...and 3 more sections

Figures (5)

  • Figure 1: Three-Pass Dialogue Flow: Progressive Identification of User Goals and Service Parameters enabling Goal-Driven Architecture
  • Figure 2: High-Level Architecture: System Components and Their Interactions
  • Figure 3: Sequence diagram for creating the web application
  • Figure 4: Token consumption scaling with increasing number of services for GPT-4o-mini and DeepSeek-V2.5. The x-axis represents the number of services and the y-axis shows the corresponding input token count.
  • Figure 5: Token distribution during application generation. The high proportion of processing tokens (89.51%) indicates potential for optimization through context management improvements.