Xinbo Ai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Xinbo Ai has a total of 53 co-authors. Here are the top co-authors by number of shared publications:
Yanzhu Hu
Beijing University of Posts and Telecommunications, No. 10 Xi Tu Cheng Road, Beijing 100876, China
6 shared publications- Yanjun Guo3 shared publications
- Yu Hu
School of Automation, Beijing Information Science and Technology University, Beijing, China
3 shared publications - Jiao Wang
School of Automation
2 shared publications Shaoyang Cheng
2 shared publicationsWenbin Luo
Beijing University of Posts and Telecommunications
2 shared publications
Latest Publications
- 8 March 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 272
Physics-informed multi-task generation: Toward automated risk control optimization
- 8 March 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 272
Physics-informed multi-task generation: Toward automated risk control optimization
- Open Access
Beyond Cosine Similarity
Cosine similarity, the standard metric for measuring semantic similarity in vector spaces, is mathematically grounded in the Cauchy-Schwarz inequality, which inherently limits it to capturing linear relationships--a constraint that fails to model the complex, nonlinear structures of real-world semantic spaces. We advance this theoretical underpinning by deriving a tighter upper bound for the dot product than the classical Cauchy-Schwarz bound. This new bound leads directly to recos, a similarity metric that normalizes the dot product by the sorted vector components. recos relaxes the condition for perfect similarity from strict linear dependence to ordinal concordance, thereby capturing a broader class of relationships. Extensive experiments across 11 embedding models--spanning static, contextualized, and universal types--demonstrate that recos consistently outperforms traditional cosine similarity, achieving higher correlation with human judgments on standard Semantic Textual Similarity (STS) benchmarks. Our work establishes recos as a mathematically principled and empirically superior alternative, offering enhanced accuracy for semantic analysis in complex embedding spaces.
Open AccessBeyond Cosine Similarity
Cosine similarity, the standard metric for measuring semantic similarity in vector spaces, is mathematically grounded in the Cauchy-Schwarz inequality, which inherently limits it to capturing linear relationships--a constraint that fails to model the complex, nonlinear structures of real-world semantic spaces. We advance this theoretical underpinning by deriving a tighter upper bound for the dot product than the classical Cauchy-Schwarz bound. This new bound leads directly to recos, a similarity metric that normalizes the dot product by the sorted vector components. recos relaxes the condition for perfect similarity from strict linear dependence to ordinal concordance, thereby capturing a broader class of relationships. Extensive experiments across 11 embedding models--spanning static, contextualized, and universal types--demonstrate that recos consistently outperforms traditional cosine similarity, achieving higher correlation with human judgments on standard Semantic Textual Similarity (STS) benchmarks. Our work establishes recos as a mathematically principled and empirically superior alternative, offering enhanced accuracy for semantic analysis in complex embedding spaces.
- 20 January 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 337
Exploring prompt distributions and probability bias for long-tailed multi-label image recognition
- 20 January 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 337
Exploring prompt distributions and probability bias for long-tailed multi-label image recognition
- 19 August 2025
- Published by Springer Nature
- p. 504-513
Degree-Aware Graph Contrastive Learning for Long-Tail Recommendations: An Empirical Analysis for Hazard Inspection
- 19 August 2025
- Published by Springer Nature
- p. 504-513
Degree-Aware Graph Contrastive Learning for Long-Tail Recommendations: An Empirical Analysis for Hazard Inspection
- 4 December 2024
- Published by Springer Nature
- p. 45-54
Mitigating Hallucination in Large Language Model by Leveraging Decoder Layer Contrasting
- 4 December 2024
- Published by Springer Nature
- p. 45-54
Mitigating Hallucination in Large Language Model by Leveraging Decoder Layer Contrasting
- 15 April 2024
- Published by Elsevier in Process Safety and Environmental Protection
- Vol. 186, 980-994
A multi-task learning risk assessment method for the chemical process industry
- 15 April 2024
- Published by Elsevier in Process Safety and Environmental Protection
- Vol. 186, 980-994
A multi-task learning risk assessment method for the chemical process industry
- Open Access
Adjust Pearson's r to Measure Nonlinear Monotone Dependence
As the most commonly used correlation coefficient, Pearson's r is generally considered as a measure only for linear dependence. The use of Pearson's r has been strongly discouraged for nonlinear relationships. Recent research shows that this opinion is myth and Pearson's r should not be ruled out a priori for measuring nonlinear monotone relationships. We find that Pearson's r does have potential to measure nonlinear dependence as long as the relationships are monotone, it just underestimates them. We derive a new inequality which is tighter than the famous Cauchy-Schwarz Inequality, leverage it to adjust the underestimation, and propose a new correlation coefficient, i.e., rearrangement correlation. The rearrangement correlation is able to capture arbitrary monotone relationships, both linear and nonlinear ones. It reverts to Pearson's r in linear scenarios. Simulation experiments and real-life investigations show that the rearrangement correlation is more accurate in measuring nonlinear monotone dependence than the three classical correlation coefficients, and other recently proposed dependence measures.
Open AccessAdjust Pearson's r to Measure Nonlinear Monotone Dependence
As the most commonly used correlation coefficient, Pearson's r is generally considered as a measure only for linear dependence. The use of Pearson's r has been strongly discouraged for nonlinear relationships. Recent research shows that this opinion is myth and Pearson's r should not be ruled out a priori for measuring nonlinear monotone relationships. We find that Pearson's r does have potential to measure nonlinear dependence as long as the relationships are monotone, it just underestimates them. We derive a new inequality which is tighter than the famous Cauchy-Schwarz Inequality, leverage it to adjust the underestimation, and propose a new correlation coefficient, i.e., rearrangement correlation. The rearrangement correlation is able to capture arbitrary monotone relationships, both linear and nonlinear ones. It reverts to Pearson's r in linear scenarios. Simulation experiments and real-life investigations show that the rearrangement correlation is more accurate in measuring nonlinear monotone dependence than the three classical correlation coefficients, and other recently proposed dependence measures.
- Open Access
- 1 January 2024
- Published by Wiley in International Journal of Intelligent Systems
- Vol. 2024 (1)
Complex Question Answering Method on Risk Management Knowledge Graph: Multi‐Intent Information Retrieval Based on Knowledge Subgraphs
The critical aspects of risk management include hazard identification, risk assessment, and risk control. Timely risk management is critical to company decision‐making, but the process of acquiring risk management knowledge is often time‐consuming and labor‐intensive. Knowledge graph question answering (KGQA) provides an effective solution by delivering knowledge through accurate reasoning. However, existing KGQA methods do not cover the critical risk management aspects and are difficult to retrieve quickly and accurately from large knowledge graphs. This study describes a complex question answering method for intelligently generating risk management knowledge, specifically through multi‐intent information retrieval based on knowledge subgraphs. The proposed method comprises three main modules. First, in the question understanding module, we propose an intent recognition method that integrates topic entity extraction with convolutional neural networks (CNNs) to identify eleven different user intents. To enhance the retrieval efficiency, we propose a hierarchical knowledge‐embedding subgraph constructed based on company and hazard descriptions. Once user intent is identified, the information retrieval module based on a novel approximate nearest neighbor (ANN) algorithm achieves deep semantic feature matching of company and hazard expressions from the knowledge embedding subgraph. After obtaining these two deep semantic features, in the answer generation module, we propose a rule‐based knowledge subgraph reasoning method to answer complex questions including single‐hop, multihop, constraints, and numerical calculations. On the real risk management dataset, the precision of the intent recognition module reaches 91.3% and the information retrieval module spends only 0.36 ms, verifying that the model outperforms the existing state‐of‐the‐art models. Meanwhile, a question answering system based on the proposed method is developed to acquire risk management knowledge: Xiao An. Compared to the popular search engine and expert system for acquiring knowledge, Xiao An achieves the best results regarding ease of use, time spent, and overall performance.
- 1 January 2024
- Published by Wiley in International Journal of Intelligent Systems
- Vol. 2024 (1)
Open AccessComplex Question Answering Method on Risk Management Knowledge Graph: Multi‐Intent Information Retrieval Based on Knowledge Subgraphs
The critical aspects of risk management include hazard identification, risk assessment, and risk control. Timely risk management is critical to company decision‐making, but the process of acquiring risk management knowledge is often time‐consuming and labor‐intensive. Knowledge graph question answering (KGQA) provides an effective solution by delivering knowledge through accurate reasoning. However, existing KGQA methods do not cover the critical risk management aspects and are difficult to retrieve quickly and accurately from large knowledge graphs. This study describes a complex question answering method for intelligently generating risk management knowledge, specifically through multi‐intent information retrieval based on knowledge subgraphs. The proposed method comprises three main modules. First, in the question understanding module, we propose an intent recognition method that integrates topic entity extraction with convolutional neural networks (CNNs) to identify eleven different user intents. To enhance the retrieval efficiency, we propose a hierarchical knowledge‐embedding subgraph constructed based on company and hazard descriptions. Once user intent is identified, the information retrieval module based on a novel approximate nearest neighbor (ANN) algorithm achieves deep semantic feature matching of company and hazard expressions from the knowledge embedding subgraph. After obtaining these two deep semantic features, in the answer generation module, we propose a rule‐based knowledge subgraph reasoning method to answer complex questions including single‐hop, multihop, constraints, and numerical calculations. On the real risk management dataset, the precision of the intent recognition module reaches 91.3% and the information retrieval module spends only 0.36 ms, verifying that the model outperforms the existing state‐of‐the‐art models. Meanwhile, a question answering system based on the proposed method is developed to acquire risk management knowledge: Xiao An. Compared to the popular search engine and expert system for acquiring knowledge, Xiao An achieves the best results regarding ease of use, time spent, and overall performance.
- 3 August 2023
- Published by Elsevier in Journal of Colloid and Interface Science
- Vol. 651, 525-533
Stabilizing iron single atoms with electrospun hollow carbon nanofibers as self-standing air-electrodes for long-time Zn − air batteries
- 3 August 2023
- Published by Elsevier in Journal of Colloid and Interface Science
- Vol. 651, 525-533
Stabilizing iron single atoms with electrospun hollow carbon nanofibers as self-standing air-electrodes for long-time Zn − air batteries