Large Reasoning Modelling for Scientific Computing

Dear Colleagues,

Advanced reasoning models—such as large language models (LLMs), neural–symbolic systems, and other hybrid AI approaches—are being integrated into scientific computing pipelines at an accelerating pace. These technologies can assist with hypothesis generation, automate simulation workflows, support data-driven discovery, and link symbolic reasoning with numerical solvers. As a result, they are reshaping research practices in physics, engineering, chemistry, biology, medicine, and related fields.

This Special Issue seeks original research articles, reviews, and well-documented case studies that examine how large-scale reasoning models contribute to scientific computing. We are especially interested in reproducible methods, theoretically grounded innovations, and practical demonstrations that show measurable benefits in real-world settings. Submissions from academic, industrial, and interdisciplinary teams are welcome.

Key themes include model architectures, scientific applications, automated workflows, knowledge representation, verification and validation, interpretability, human–AI collaboration, benchmarking, efficiency, and emerging cross-disciplinary directions. Work that provides open data, open source code, or detailed experimental protocols is strongly encouraged.

Topics of interest (non-exhaustive):

• Model Development and Architectures

1.1. Large reasoning models tailored to scientific tasks;

1.2. Domain-specific or foundation LLMs for computation and analysis;

1.3. Integration of LLMs with established scientific computing platforms;

1.4. Neural–symbolic and hybrid AI architectures for science.

• Applications in Scientific Domains

2.1. Physics-informed computing and simulation control;

2.2. Chemical and molecular design assistance;

2.3. Systems biology and bioinformatics pipelines;

2.4. Climate and environmental modeling;

2.5. Engineering simulations augmented by reasoning models.

• Scientific Workflows and Automation

3.1. AI-driven experiment design, planning, and optimization;

3.2. Automated theorem proving and symbolic mathematics;

3.3. Code generation, debugging, and refactoring for scientific software;

3.4. End-to-end workflow automation for data preparation, simulation, and analysis.

• Knowledge Representation and Reasoning

4.1. Ontology-guided reasoning for discipline-specific knowledge;

4.2. Structured data interpretation with LLMs;

4.3. Integration with scientific knowledge graphs and databases;

4.4. Causal inference techniques in computational science.

• Verification, Validation, and Assurance

5.1. Formal verification methods for LLM-driven simulations;

5.2. Statistical validation of AI-generated scientific outputs;

5.3. Certification frameworks for safety-critical scientific applications;

5.4. Debugging and testing strategies for hybrid reasoning systems.

• Interpretability, Trust, and Ethics

6.1. Explainable AI techniques tailored to scientific contexts;

6.2. Uncertainty quantification and risk assessment;

6.3. Responsible data use, compliance, and reproducibility standards.

• Human-AI Collaboration

7.1. Human-in-the-loop discovery processes;

7.2. Interactive assistants and cognitive interfaces for researchers.

• Benchmarking and Evaluation

8.1. Public datasets and benchmarks for reasoning in science;

8.2. Metrics for performance, robustness, and resource usage;

8.3. Comparative studies of symbolic, sub-symbolic, and hybrid approaches.

• Scalability and Computational Efficiency

9.1. Efficient training and inference strategies for large models;

9.2. Federated, distributed, and parallel reasoning systems;

9.3. Memory-efficient architectures and compression techniques.

• Cross-Disciplinary and Emerging Topics

10.1. Multi-modal reasoning across text, images, and numerical data;

10.2. Simulation-based inference with LLM support;

10.3. Zero-shot and few-shot learning in scientific tasks.

Dr. Nicholas Christakis
Prof. Dr. Dimitris Drikakis
Dr. William J. Rider
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Applied Sciences is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• large language models (LLMs)
• scientific computing
• artificial intelligence
• human-computer interaction
• computational efficiency