Designing safety-critical controllers for the blast furnace is crucial yet challenging due to its complex inherent dynamics. Recent advances in reinforcement learning (RL) have shown promise in designing effective controllers for complex industrial systems. However, direct application of RL control to an on-site blast furnace is hindered by stringent safety requirements and model discrepancies. To address these challenges, this paper presents a duo-stage RL-based safety-critical control framework (DRLSC Framework) that integrates safety-critical offline pretraining and an online agent transfer mechanism for long-term safe control of the blast furnace. An offline pretraining algorithm is designed to provide an adequate initial policy for online control, which consists of a data-driven blast furnace environment and a Soft Actor-Critic (SAC) agent. The offline stage incorporates a neural control barrier function (NCBF) to guarantee system state safety, which is trained jointly with the RL agent. During the online stage, a transfer mechanism facilitates rapid agent adaptation to the actual system dynamics, mitigating performance degradation arising from model discrepancies. Finally, silicon content ([Si]) tracking control experiments are carried out using a real dataset obtained from the #2 blast furnace of Liuzhou Steel in China. Compared with model-predictive-control-based methods, our DRLSC Framework can decrease the long-term tracking error by 50.21% when facing simulation discrepancy and maintain state safety continuously.

E-commerce is increasinglymultimedia-enriched, with products exhibited in a broad-domain manner as images, short videos, or live stream promotions. A unified and vectorized cross-domain production representation is essential. Due to large intra-product variance and high inter-product similarity in the broad-domain scenario, a visual-only representation is inadequate. While Automatic Speech Recognition (ASR) text derived from the short or live-stream videos is readily accessible, how to de-noise the excessively noisy text for multimodal representation learning is mostly untouched. We proposeASR-enhancedMultimodalProduct Representation Learning (AMPere). In order to extract product-specific information from the raw ASR text,AMPereuses an easy-to-implement LLM-based ASR text summarizer. The LLM-summarized text, together with visual data, is then fed into a multi-branch network to generate compact multimodal embeddings. Extensive experiments on a large-scale tri-domain dataset verify the effectiveness ofAMPerein obtaining a unified multimodal product representation that clearly improves cross-domain product retrieval.

Contemporary recommendation systems predominantly rely on ID embedding to capture latent associations among users and items. However, this approach overlooks the wealth of semantic information embedded within textual descriptions of items, leading to suboptimal performance and poor generalizations. Leveraging the capability of large language models to comprehend and reason about textual content presents a promising avenue for advancing recommendation systems. To achieve this, we propose an Llm-driven knowlEdge Adaptive RecommeNdation (LEARN) framework that synergizes open-world knowledge with collaborative knowledge. We address computational complexity concerns by utilizing pretrained LLMs as item encoders and freezing LLM parameters to avoid catastrophic forgetting and preserve open-world knowledge. To bridge the gap between the open-world and collaborative domains, we design a twin-tower structure supervised by the recommendation task and tailored for practical industrial application. Through experiments on the real large-scale industrial dataset and online A/B tests, we demonstrate the efficacy of our approach in industry application. We also achieve state-of-the-art performance on six Amazon Review datasets to verify the superiority of our method.

With the advancement of Industry 4.0, digital twins have become increasingly widespread in the industrial sector, providing real-time data acquisition, simulation, and optimization capabilities. However, ensuring standardized operations remains a significant challenge, particularly in complex environments where the absence of experienced engineers can lead to non-compliant designs or unsafe practices. To address this, digital twins can act as self-checking tools or enable manual querying of standards, ensuring compliance throughout the industrial process. This paper proposes a framework for the standardized operation of digital twins based on industrial large language models (LLMs), integrating industrial standards and dynamically detecting compliance during operations. Experiments on power system standardization tested different LLM approaches, demonstrating the framework’s ability to guide task-specific LLM selection. This research offers a robust solution for enhancing compliance, safety, and efficiency in industrial AI-powered digital twins.

The prediction of sintering quality indicators is essential for the automation of the ironmaking process. However, various challenges, including variable time delays and relatively weak correlations, present difficulties in achieving precise modeling of the sintering process. To address these challenges, we introduce a novel method based on the nonuniform time delay embedding technique for predicting sintering quality indicators. First, we construct input variables with different compensated time delays. Second, we develop a low-dimensional approximation of the joint mutual information criterion to search the optimal variables and time delays. This approach effectively leverages valuable information from process variables while minimizing irrelevant information. Next, we implement a hybrid modeling technique that combines extreme gradient boosting and support vector regression to predict sintering quality indicators, allowing for adaptation to various operating conditions. Finally, we validate the effectiveness of our proposed method using real production data, specifically focusing on FeO content.

The proliferation of short video and live-streaming platforms has revolutionized how consumers engage in online shopping. Instead of browsing product pages, consumers are now turning to rich-content e-commerce, where they can purchase products through dynamic and interactive media like short videos and live streams. This emerging form of online shopping has introduced technical challenges, as products may be presented differently across various media domains. Therefore, a unified product representation is essential for achieving cross-domain product recognition to ensure an optimal user search experience and effective product recommendations. Despite the urgent industrial need for a unified cross-domain product representation, previous studies have predominantly focused only on product pages without taking into account short videos and live streams. To fill the gap in the rich-content e-commerce area, in this paper, we introduce a large-scale cRoss-dOmain Product rEcognition dataset, called ROPE. ROPE covers a wide range of product categories and contains over 180,000 products, corresponding to millions of short videos and live streams. It is the first dataset to cover product pages, short videos, and live streams simultaneously, providing the basis for establishing a unified product representation across different media domains. Furthermore, we propose a Cross-dOmain Product rEpresentation framework, namely COPE, which unifies product representations in different domains through multimodal learning including text and vision. Extensive experiments on downstream tasks demonstrate the effectiveness of COPE in learning a joint feature space for all product domains.