Towards Building General Purpose Embedding Models for Industry 4.0 Agents

2025-06-14

AI Summary

To address weak natural language query understanding, poor cross-asset generalization, and slow downtime response in Industry 4.0 asset operations, this paper proposes an LLM-augmented task embedding framework. First, it constructs domain-specific task representations by fusing expert knowledge with LLM-enhanced semantic encoding. Second, it introduces a contrastive learning strategy tailored for multi-item relational queries to improve semantic matching accuracy. Third, it enables end-to-end collaborative reasoning between the embedding model and a ReAct-based intelligent agent, supporting multi-step planning and tool invocation. Evaluated on real-world industrial datasets, the framework achieves significant improvements: HIT@1 increases by 54.2%, MAP@100 by 50.1%, and NDCG@10 by 54.7%. These gains substantially enhance engineers’ decision-making efficiency and system robustness in fault diagnosis and response.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Search and Optimization: Learning to Search
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

In this work we focus on improving language models' understanding for asset maintenance to guide the engineer's decisions and minimize asset downtime. Given a set of tasks expressed in natural language for Industry 4.0 domain, each associated with queries related to a specific asset, we want to recommend relevant items and generalize to queries of similar assets. A task may involve identifying relevant sensors given a query about an asset's failure mode. Our approach begins with gathering a qualitative, expert-vetted knowledge base to construct nine asset-specific task datasets. To create more contextually informed embeddings, we augment the input tasks using Large Language Models (LLMs), providing concise descriptions of the entities involved in the queries. This embedding model is then integrated with a Reasoning and Acting agent (ReAct), which serves as a powerful tool for answering complex user queries that require multi-step reasoning, planning, and knowledge inference. Through ablation studies, we demonstrate that: (a) LLM query augmentation improves the quality of embeddings, (b) Contrastive loss and other methods that avoid in-batch negatives are superior for datasets with queries related to many items, and (c) It is crucial to balance positive and negative in-batch samples. After training and testing on our dataset, we observe a substantial improvement: HIT@1 increases by +54.2%, MAP@100 by +50.1%, and NDCG@10 by +54.7%, averaged across all tasks and models. Additionally, we empirically demonstrate the model's planning and tool invocation capabilities when answering complex questions related to industrial asset maintenance, showcasing its effectiveness in supporting Subject Matter Experts (SMEs) in their day-to-day operations.

Problem

• Improving language models for asset maintenance decisions
• Recommending relevant items for Industry 4.0 queries
• Enhancing embeddings with LLM-augmented contextual information

Innovation

• LLM-augmented embeddings for contextual understanding
• ReAct agent for multi-step reasoning
• Contrastive loss for improved query-item matching