Mohammed I. Radaideh
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
Mohammed I. Radaideh has a total of 118 co-authors. Here are the top co-authors by number of shared publications:
- Tomasz Kozlowski
University of Illinois Urbana-Champaign, Department of Nuclear, Plasma, and Radiological Engineering, Urbana, Illinois 61801
7 shared publications - Dean Price
University of Michigan
6 shared publications Sarah Cousineau
Spallation Neutron Source, Oak Ridge National Laboratory, Oak Ridge, TN 37830, United States
6 shared publicationsTomasz Kozlowski
Department of Nuclear, Plasma, and Radiological Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801, United States
6 shared publicationsChris Pappas
Oak Ridge Natl Lab, Oak Ridge, TN 37830 USA
5 shared publicationsLinyu Lin
5 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Springer Nature in Artificial Intelligence Review
When computers call the shots: a systematic review exploring the impact of AI in high-stakes decision-making
While applications abound for the modern surge of artificial-intelligence (AI)-based technologies, we owe special care to those in which human lives are at stake . Perilously, AI advances faster than the regulations that protect us from it. As such, in this review, we explore the impact of AI on high-stakes decision-making. Though reviews have attempted adjacent analyses before, none have effectively targeted safety-critical industries or considered the relation between AI developers, users, and regulators. To address these gaps, we conduct a systematic literature review. We cast a wide net in our search of two databases, Web of Science and Scopus, then filter with a series of criteria that qualify a system’s sensitivity, decision-making framework, and risk. After distilling our initial pool of 1135 papers retrieved on December 15, 2025 into a final subset of 72 reports, we identify critical features of AI systems and industry implementation practices that demonstrate: (1) the safety risks of using AI in high-stakes decision-making, (2) how we can mitigate these risks, and (3) how we can assign responsibility for AI-based decisions. We perform a quality assessment, including risk of bias, on each study in our final corpus to ensure adequate rigor of retrieved material. We extract categorical and binary data from the articles reviewed to provide quantitative trends towards the types of AI used across the industries considered and the expertise of the end-user of these applications, as well as the decision-making time those users are allotted. We also perform a qualitative analysis of the trends described in these studies, on a per-industry basis.
- 9 November 2026
- Published by Springer Nature in Artificial Intelligence Review
Open AccessWhen computers call the shots: a systematic review exploring the impact of AI in high-stakes decision-making
While applications abound for the modern surge of artificial-intelligence (AI)-based technologies, we owe special care to those in which human lives are at stake . Perilously, AI advances faster than the regulations that protect us from it. As such, in this review, we explore the impact of AI on high-stakes decision-making. Though reviews have attempted adjacent analyses before, none have effectively targeted safety-critical industries or considered the relation between AI developers, users, and regulators. To address these gaps, we conduct a systematic literature review. We cast a wide net in our search of two databases, Web of Science and Scopus, then filter with a series of criteria that qualify a system’s sensitivity, decision-making framework, and risk. After distilling our initial pool of 1135 papers retrieved on December 15, 2025 into a final subset of 72 reports, we identify critical features of AI systems and industry implementation practices that demonstrate: (1) the safety risks of using AI in high-stakes decision-making, (2) how we can mitigate these risks, and (3) how we can assign responsibility for AI-based decisions. We perform a quality assessment, including risk of bias, on each study in our final corpus to ensure adequate rigor of retrieved material. We extract categorical and binary data from the articles reviewed to provide quantitative trends towards the types of AI used across the industries considered and the expertise of the end-user of these applications, as well as the decision-making time those users are allotted. We also perform a qualitative analysis of the trends described in these studies, on a per-industry basis.
- Open Access
- 6 November 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 279
Spatially aware generative models for full-field sensor reconstruction in safety-critical digital twins
- 6 November 2026
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 279
Open AccessSpatially aware generative models for full-field sensor reconstruction in safety-critical digital twins
- Open Access
- 25 September 2026
- Published by Elsevier in Proceedings of the Combustion Institute
- Vol. 42
Uncertainty-aware smoke-point screening of SAF-rich aviation fuels via optimization and Bayesian inference
- 25 September 2026
- Published by Elsevier in Proceedings of the Combustion Institute
- Vol. 42
Open AccessUncertainty-aware smoke-point screening of SAF-rich aviation fuels via optimization and Bayesian inference
- Open Access
- 17 August 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 458
MAEO: Scalable Multiobjective Animorphic Ensemble Optimization for large-scale optimization applications
- 17 August 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 458
Open AccessMAEO: Scalable Multiobjective Animorphic Ensemble Optimization for large-scale optimization applications
- Open Access
High-fidelity modeling of pressurized water reactor flow fields for data-driven applications
- Open Access
- 25 July 2026
- Published by Elsevier in Applied Thermal Engineering
- Vol. 303
Multifidelity surrogate modeling of onset of natural circulation following a depressurized loss of forced cooling in high-temperature gas reactors
- 25 July 2026
- Published by Elsevier in Applied Thermal Engineering
- Vol. 303
Open AccessMultifidelity surrogate modeling of onset of natural circulation following a depressurized loss of forced cooling in high-temperature gas reactors
- Open Access
- 15 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 81885-81900
On the Impact of Language Nuances on Sentiment Analysis With Large Language Models: Paraphrasing, Sarcasm, and Emojis
This research investigates how textual nuances like emojis and sarcasm impact the performance of Large Language Models (LLMs) in sentiment analysis, especially when using social media data. To improve data quality, the study explores text paraphrasing techniques and addresses the scarcity of labeled sarcasm data by creating a human-labeled dataset of 5,929 tweets, enabling a more robust evaluation of LLMs in handling sarcasm. The results show that when topic-specific datasets related to nuclear power are used to finetune LLMs, these models are not able to comprehend accurate sentiment in the presence of sarcasm due to less diverse text, requiring external interventions like sarcasm removal to boost model accuracy. Sarcasm removal led to up to 21% improvement in sentiment accuracy, as LLMs trained on nuclear power-related content struggled with sarcastic tweets, achieving only 30% accuracy. In contrast, LLMs trained on dataset of 690,000 general tweets, covering a broader range of topics, showed considerable improvements in predicting sentiment for sarcastic tweets (60% accuracy), indicating that incorporating general text data can enhance sarcasm detection. The study also utilized adversarial text paraphrasing, showing that creating synthetic text variants by making minor changes significantly increased model robustness and accuracy for sarcastic tweets (approximately 85%). Additionally, text paraphrasing of tweets with fragmented language transformed around 40% of approximately 363,000 tweets, improving LLMs sentiment analysis accuracy by 6%. Finally, emojis did not significantly affect sentiment analysis for nuclear power-related content, suggesting that emojis may only reinforce, rather than reveal, sentiment in certain contexts.
- 15 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 81885-81900
Open AccessOn the Impact of Language Nuances on Sentiment Analysis With Large Language Models: Paraphrasing, Sarcasm, and Emojis
This research investigates how textual nuances like emojis and sarcasm impact the performance of Large Language Models (LLMs) in sentiment analysis, especially when using social media data. To improve data quality, the study explores text paraphrasing techniques and addresses the scarcity of labeled sarcasm data by creating a human-labeled dataset of 5,929 tweets, enabling a more robust evaluation of LLMs in handling sarcasm. The results show that when topic-specific datasets related to nuclear power are used to finetune LLMs, these models are not able to comprehend accurate sentiment in the presence of sarcasm due to less diverse text, requiring external interventions like sarcasm removal to boost model accuracy. Sarcasm removal led to up to 21% improvement in sentiment accuracy, as LLMs trained on nuclear power-related content struggled with sarcastic tweets, achieving only 30% accuracy. In contrast, LLMs trained on dataset of 690,000 general tweets, covering a broader range of topics, showed considerable improvements in predicting sentiment for sarcastic tweets (60% accuracy), indicating that incorporating general text data can enhance sarcasm detection. The study also utilized adversarial text paraphrasing, showing that creating synthetic text variants by making minor changes significantly increased model robustness and accuracy for sarcastic tweets (approximately 85%). Additionally, text paraphrasing of tweets with fragmented language transformed around 40% of approximately 363,000 tweets, improving LLMs sentiment analysis accuracy by 6%. Finally, emojis did not significantly affect sentiment analysis for nuclear power-related content, suggesting that emojis may only reinforce, rather than reveal, sentiment in certain contexts.
- Open Access
- 5 May 2026
- Published by Elsevier in Applied Energy
- Vol. 416
Physics-informed digital twin development of thermal energy distribution systems with active learning
- 5 May 2026
- Published by Elsevier in Applied Energy
- Vol. 416
Open AccessPhysics-informed digital twin development of thermal energy distribution systems with active learning
- Open Access
- 28 November 2025
- Published by Elsevier in Energy and AI
- Vol. 22
Surrogate-driven design optimization with uncertainty constraints in Monte Carlo simulations
- 28 November 2025
- Published by Elsevier in Energy and AI
- Vol. 22
Open AccessSurrogate-driven design optimization with uncertainty constraints in Monte Carlo simulations