In sectors such as energy, aquaculture and chemical processing, the effectiveness of both mechanistic and data-driven models is frequently limited by poor-quality training data, which obscures underlying physical laws and degrades predictive accuracy. Such quality issues commonly manifest as anomalies in the dataset, ranging from sensor faults to process upsets and incorrect operator inputs. While traditional rule-based filtering can remove a subset of the anomalies, it is unable to leverage contextual information stored as unstructured text, such as engineer comments and operational logs that explain under what conditions the data was collected. We propose a text-informed data filtering method that leverages large language models (LLMs) to integrate unstructured operational data with tabular sensor data. The method consists of two steps: (1) contextual anomaly detection by reasoning over both data modalities, and (2) generation of a binary mask that is applied to filter the training dataset for model development. To investigate the potential benefit, a steady-state soft-sensor regression model, namely a virtual flow meter that estimates well flow rates, is used as the downstream evaluation target. Experiments on a synthetic oil and gas dataset show improved model performance compared to traditional rule-based filtering. Ablation studies show that textual operator comments can provide filtering-relevant signals beyond sensor data alone, with the multimodal approach reducing downstream MAPE from 7.1% to 5.4% relative to sensor-only filtering. This work demonstrates that integrating unstructured operational knowledge with sensor data through LLMs can improve data quality for industrial machine learning applications, offering a scalable alternative to manual data curation.
Figures & tables
Figure 1: Traditional flow of information in industrial data curation. The operator synthesises background domain knowledge, operational context, and sensor readings into unstructured comments. These comments, together with the raw sensor data, then inform a manual curation step that determines which data points enter model training.
Category
Perturbation type
Physical mechanism
Scaling
Well parameter ( fD , D )
Mineral deposits restricting flow
Hydrate formation
Well parameter ( fD , D )
Ice-like plugs in the wellbore
Liquid loading
Well parameter ( α )
Inability to lift liquids
Emulsion
Well parameter ( fD )
Oil-water emulsions increasing friction
Wax deposition
Well parameter ( fD )
Wax build-up increasing friction
Data quality
Simulated parameter
Sensor faults and communication failures
Table 1: Anomaly groups.
Method
RQ1
RQ2
RQ3
Uniform (no filter)
—
—
—
Tukey
B
Mahalanobis
B
LLM (multimodal)
∙
∙
B
LLM (text only)
abl.
LLM (sensor only)
B
Table 2: Role of each method per research question. B = baseline, ∙ = tested, abl. = ablation, — = reference bound.
Figure 3: Example of a dataset for a single well and the model fitted to the nominal data points.
Method
MSE ((Sm 3 /h) 2 )
MAPE (%)
W10 (%)
Uniform (no filter)
148 [123, 176]
12.5 [11.5, 13.6]
51 [46, 55]
Tukey
112 [86, 144]
10.0 [8.9, 11.2]
66 [62, 70]
Mahalanobis
123 [98, 151]
9.4 [8.3, 10.6]
66 [62, 71]
LLM (multimodal)
53 [44, 64]
5.4 [4.9, 5.9]
87 [85, 89]
LLM (text only)
86 [71, 103]
8.0 [7.4, 8.7]
78 [76, 81]
LLM (sensor only)
88 [74, 103]
7.1 [6.5, 7.8]
84 [82, 86]
Table 3: MSE, MAPE and W10, pooled across wells and classification runs (95 % confidence intervals).
Figure 4: Results presented as a cumulative performance plot.
Method
Match (%)
LLM (multimodal)
96.6
LLM (text only)
100.0
LLM (sensor only)
92.7
LLM + Tukey
96.4
LLM + Mahalanobis
97.6
Table 4: Run-to-run consistency: Match(a,a) across all training wells (percent matching decisions).
Figure 5: Data points of the well with index 745 processed by the LLM (multimodal) method. Data points of particular interest are numbered.
Data preparation pipelines improve data quality in machine learning by transforming raw tables into learning-ready data through sequential cleaning and feature transformation operators. However, automatically constructing such pipelines is computationally difficult because operator sequences are combinatorial and end-to-end evaluation is expensive. Existing state-of-the-art (SOTA) Multi-DQN methods still face three key limitations: decoupled value estimators weaken long-horizon credit assignment, dataset context is only weakly injected into the policy, and exploration is inefficient in a sparse search space with many invalid states. To address these issues, we propose FlowPipe, a unified framework that formulates pipeline synthesis as conditional probabilistic flow generation over a directed acyclic graph. FlowPipe uses Conditional Generative Flow Networks (C-GFlowNets) with a Trajectory Balance objective to connect terminal validation rewards with early pipeline decisions. It further introduces Deep Semantic Modulation through Feature-wise Linear Modulation (FiLM), allowing LLM-derived logical priors to condition the policy's internal activations according to dataset semantics. In addition, FlowPipe incorporates failure awareness into the flow objective to avoid invalid states and concentrate search on high-potential regions. Experiments on two benchmark suites with 74 real-world datasets show that FlowPipe outperforms SOTA baselines, improving accuracy by 11.96% on average and achieving 12.5x faster training convergence. Source code is available at https://github.com/KunyuNi/FlowPipe.
Kunyu Ni, Lei Cao, Jie He +4
Ocean University of China China · University of Arizona USA · University of Science and Technology Beijing China +1
We introduce FailureSensorIQ, a novel Multi-Choice Question-Answering (MCQA) benchmarking system designed to assess the ability of Large Language Models (LLMs) to reason and understand complex, domain-specific scenarios in Industry 4.0. Unlike traditional QA benchmarks, our system focuses on multiple aspects of reasoning through failure modes, sensor data, and the relationships between them across various industrial assets. Through this work, we envision a paradigm shift where modeling decisions are not only data-driven using statistical tools like correlation analysis and significance tests, but also domain-driven by specialized LLMs which can reason about the key contributors and useful patterns that can be captured with feature engineering. We evaluate the Industrial knowledge of over a dozen LLMs-including GPT-4, Llama, and Mistral-on FailureSensorIQ from different lens using Perturbation-Uncertainty-Complexity analysis, Expert Evaluation study, Asset-Specific Knowledge Gap analysis, ReAct agent using external knowledge-bases. Even though closed-source models with strong reasoning capabilities approach expert-level performance, the comprehensive benchmark reveals a significant drop in performance that is fragile to perturbations, distractions, and inherent knowledge gaps in the models. We also provide a real-world case study of how LLMs can drive the modeling decisions on 3 different failure prediction datasets related to various assets. We release: (a) expert-curated MCQA for various industrial assets, (b) FailureSensorIQ benchmark and Hugging Face leaderboard based on MCQA built from non-textual data found in ISO documents, and (c) LLMFeatureSelector, an LLM-based feature selection scikit-learn pipeline. The software is available at https://github.com/IBM/FailureSensorIQ.
Christodoulos Constantinides, Dhaval Patel, Shuxin Lin +3
Large language model (LLM) agents have shown promise in automating complex data-analysis workflows, but their reliable deployment remains challenging in high-stakes industrial scenarios. Industrial anomaly detection (IAD) is essential for manufacturing quality, safety, and efficiency, yet existing LLM-based IAD agents mainly focus on execution while under-exploiting strategy formulation. Consequently, they struggle to handle heterogeneous modalities in a unified and cost-effective manner. Inspired by the DMAIC quality-management framework, we propose DMAIC-IAD (DMAIC-inspired Agentic Industrial Anomaly Detection), a "Plan First, Judge Later" multi-agent system that aligns LLM agents with structured industrial problem-solving. DMAIC-IAD distills heterogeneous references into standardized operating procedures (SOPs) before strategy generation, and introduces a pre-trained execution-free judge model to rank candidate strategies without costly runtime trials. Extensive experiments across four modalities show that DMAIC-IAD improves average detection performance over applicable agentic baselines by 37.76%.
Yongzi Yu, Ao Li, Le Wang +4
1The Hong Kong University of Science and Technology (Guangzhou) · 2The Hong Kong University of Science and Technology · 3Shanghai University of Finance and Economics +2