📋 AI Review from DeepReviewer will be automatically processed
📋 AI Review from ZGCA will be automatically processed
This paper reviews the emerging field of intelligent rock mechanics, tracing AI methods from early BP/SVM models to modern deep learning (CNNs, LSTMs, GANs, transformers), physics-informed/hybrid solvers (PINNs and variants, graph-based models), and multi-fidelity/ML-accelerated computational mechanics. It is organized by a pipeline: (i) data-driven estimation of properties (UCS, stiffness, wave velocities) from indirect indicators and logs (Section 2); (ii) image-based microstructure reconstruction and fracture detection, including generative models and 2D/3D segmentation/point cloud methods (Section 3); (iii) AI-assisted constitutive modeling and physics-informed simulation, including neural constitutive laws, FEM-PINN hybrids, and variants such as PIRBN, PI-TCN (Section 4); and (iv) applications in rock engineering (classification/characterization, rockburst/geohazard prediction, tunneling/boring, slope stability, and other use cases) (Section 5). Section 6 discusses cross-cutting challenges (data scarcity/standardization, interpretability/trust, computational efficiency, workflow integration, and lifecycle maintenance) and highlights forward-looking directions including LLMs for code generation/decision support and spatial intelligence for multi-scale integration. The paper positions itself as a systematic review and an agenda-setting perspective for the next decade.
Cross‑Modal Consistency: 26/50
Textual Logical Soundness: 22/30
Visual Aesthetics & Clarity: 8/20
Overall Score: 56/100
Detailed Evaluation (≤500 words):
Image‑first understanding (visual ground truth)
• Figure 1 / (a) Schematic of message passing in a graph; arrows between nodes; no readable labels.
• Figure 1 / (b) “Multi‑fidelity GNN” block diagram; two DNN branches; no variable names.
• Figure 1 / (c) Bar chart (computation time) and line plot (error vs NDOF); axes/units illegible.
• Figure 1 / (d) Simplified FEA cases with color maps of |u|; qualitative comparison FEA vs MFGNN.
• Rockburst visuals / (a) Geological cross‑section showing lithologies/faults and rockburst depths; symbols legend tiny.
• Rockburst visuals / (b) Time series: daily microseismic events and S‑value; dual‑axis; labels unreadable.
• Rockburst visuals / (c) Time series: daily cumulative energy and incidence; dual‑axis; text too small.
Synopsis: Fig. 1 appears to illustrate GNN/MFGNN methodology and performance for FEM surrogates. The rockburst set seems to show site geology and monitoring trends; none are tied to captions in the text here.
1. Cross‑Modal Consistency
• Major 1: “see Figure 1” is used to support PINN/Shield‑tunneling discussion, but Fig. 1 shows MFGNN for FEM, not PINN. Evidence: Sec 5.3 “…PINN formulations… for shield tunneling… (Zhang et al., 2023b)(see Figure 1).”
• Major 2: Three rockburst figures appear without caption numbers or in‑text citation in Sec 5.2. Evidence: Uncaptioned rockburst cross‑section and two time‑series plots after Sec 5.5.
• Major 3: Fig. 1 sub‑panes lack labels (a/b/…), while prose never specifies which pane supports which claim. Evidence: Fig. 1 caption “Machine learning methods for solving FEM problems (Black & Najafi, 2022)” without sub‑labels.
• Minor 1: Mixed figure sourcing—caption credits Black & Najafi (2022), but nearby prose ties to Zhang et al. (2023b). Evidence: Sec 5.3 around Fig. 1.
• Minor 2: No axis titles/units visible in performance plots, hindering claim verification. Evidence: Fig. 1(c) small, unreadable axes.
2. Text Logic
• Major 1: The Introduction mentions “A bibliometric overview of 17 journals,” but no figure/table/method is provided. Evidence: Sec 1 “A bibliometric overview of 17 journals…” (no supporting visual/data).
• Minor 1: Occasional forward‑looking claims (e.g., “co‑pilot” for operations) lack quantitative backing in Results sections.
• Minor 2: Some reference-year mix (2025) cited for methods without clarifying preprint/accepted status.
3. Figure Quality
• Major 1: Multiple panes are illegible at print size; tick labels and legends cannot be read. Evidence: Panel resolutions ≈136–175 px height; text unreadable.
• Major 2: Missing sub‑figure labels and legends impede independent interpretation. Evidence: Fig. 1 lacks (a–d) markers; rockburst figures lack captions.
• Minor 1: Color choices appear adequate, but no colorbars for displacement fields.
Key strengths:
• Broad, well‑structured survey bridging properties, imaging, constitutive ML, and applications.
• Good coverage of physics‑aware ML (PINNs, PIRBN, PI‑TCN, MFGNN) with relevant citations.
• Practical framing of challenges and workflow integration.
Key weaknesses:
• Core figure–text mismatches (PINN vs MFGNN) and uncited/uncaptioned rockburst visuals.
• Illegible plots block claim verification.
• Bibliometric claim lacks presented evidence.
Recommendations:
• Relabel Fig. 1 with sub‑panes, readable axes/units, and map each pane to specific sentences.
• Provide a dedicated figure/table for the “17‑journal” bibliometric overview.
• Add captions and in‑text references for rockburst figures or remove if not discussed.
📋 AI Review from SafeReviewer will be automatically processed
This paper provides a comprehensive review of the application of artificial intelligence (AI) and machine learning (ML) techniques in the field of rock mechanics. The authors trace the evolution of AI in this domain, starting from early methods like backpropagation and support vector machines to modern deep learning architectures such as convolutional neural networks (CNNs) and transformers. The paper highlights the growing integration of AI in various aspects of rock engineering, including data-driven estimation of rock properties, image-based modeling and fracture detection, AI-assisted constitutive modeling, and physics-informed simulation. The authors also discuss the practical applications of these techniques in rock engineering, such as rock mass classification, rockburst and geohazard prediction, tunneling and boring operations, and slope stability analysis. The paper concludes by identifying key challenges and future directions, emphasizing the need for standardized datasets, interdisciplinary collaboration, and the development of transparent and reproducible AI workflows. While the paper provides a valuable synthesis of recent advancements, it falls short in offering a critical evaluation of the limitations and challenges associated with the adoption of AI in rock mechanics, particularly in terms of practical implementation and the need for domain expertise.
One of the paper's core strengths lies in its comprehensive and well-structured review of the integration of AI and ML techniques in rock mechanics. The authors effectively trace the historical development of AI in this field, starting from early methods like backpropagation and support vector machines to modern deep learning architectures such as CNNs and transformers. This historical context is crucial for understanding the current state of the field and the trajectory of future research. The paper also provides a detailed overview of the various applications of AI in rock mechanics, including data-driven estimation of rock properties, image-based modeling and fracture detection, AI-assisted constitutive modeling, and physics-informed simulation. Each of these sections is well-supported with specific examples and references, which adds to the paper's credibility and depth. For instance, the discussion on data-driven estimation of rock properties includes examples of using ANNs to infer rock strength and stiffness from indirect indicators, while the section on image-based modeling highlights the use of CNNs for automating the segmentation and quantification of rock fractures. The paper's emphasis on the practical applications of AI in rock engineering, such as rock mass classification, rockburst and geohazard prediction, tunneling and boring operations, and slope stability analysis, is particularly valuable. These examples demonstrate the potential of AI to enhance the efficiency and accuracy of rock engineering tasks. Additionally, the authors acknowledge the importance of interdisciplinary collaboration and the need for standardized datasets, which are essential for the advancement of AI in rock mechanics. The paper's forward-looking perspective on the development of next-generation intelligent frameworks capable of coupling physical knowledge, spatial reasoning, and adaptive learning is also commendable, as it sets a clear direction for future research.
Despite its comprehensive review and well-structured content, the paper has several limitations that need to be addressed. One significant weakness is the lack of a critical evaluation of the limitations and challenges associated with the adoption of AI and ML techniques in rock mechanics. While the paper mentions challenges such as data quality, model generalization, and interpretability, it does not delve deeply into these issues or provide specific examples of where current methods fall short. For instance, the paper could benefit from a more detailed discussion on the potential for overfitting in complex models, especially given the often limited size of datasets in rock mechanics. The authors should explore regularization techniques, cross-validation strategies, and the importance of domain expertise in feature engineering to mitigate overfitting. This would provide a more balanced and critical perspective on the current state of AI in rock mechanics. Another limitation is the paper's focus on specific ML techniques without a broader discussion of alternative approaches. The paper extensively discusses methods like CNNs, LSTMs, and GANs, but it does not compare these with other relevant techniques such as reinforcement learning or causal inference. This omission limits the paper's scope and could be addressed by including a section that explores the potential of these alternative methods and the challenges associated with their implementation. The paper also lacks a detailed discussion on the practical implications of using AI in rock mechanics. While the authors mention the potential for improved efficiency and accuracy, they do not address the practical challenges of deploying these models in real-world scenarios. For example, the paper could discuss the computational resources required for training and deploying complex models, the need for specialized expertise, and the potential for model bias. A more thorough analysis of these practical considerations would enhance the paper's relevance and utility for practitioners in the field. Furthermore, the paper does not provide a clear articulation of its novel contributions. While it serves as a valuable synthesis of recent advancements, it lacks a critical analysis of the limitations of existing methods and the potential for future research. The authors should explicitly state the unique aspects of their review, such as a novel framework for understanding the challenges of AI in rock mechanics or a new perspective on the practical implications of these technologies. This would help to differentiate the paper from other review articles and highlight its significance. Lastly, the paper could benefit from a more detailed discussion on the importance of physics-informed constraints in ML models. While the authors mention the use of physics-informed neural networks (PINNs), they do not fully explore the implications of incorporating physical constraints into ML models. For example, the paper could discuss how physics-informed constraints can improve the generalization capabilities of ML models, reduce the need for large datasets, and enhance the interpretability of model predictions. The authors should also provide specific examples of how physics-informed constraints have been successfully applied in rock mechanics and discuss the challenges associated with implementing these constraints in practice. This would provide a more nuanced understanding of the role of physics-informed constraints in the development of reliable and trustworthy AI tools for rock mechanics.
To address the identified weaknesses, the paper should include a dedicated section that critically evaluates the limitations and challenges associated with the adoption of AI and ML techniques in rock mechanics. This section should explore issues such as the potential for overfitting in complex models, especially given the often limited size of datasets in rock mechanics. The authors should discuss regularization techniques, cross-validation strategies, and the importance of domain expertise in feature engineering to mitigate overfitting. Additionally, the paper should broaden its scope by including a discussion of alternative ML techniques such as reinforcement learning and causal inference. The authors should explore the potential of these methods in rock mechanics and the challenges associated with their implementation. This would provide a more comprehensive overview of the field and highlight areas for future research. The paper should also delve deeper into the practical implications of using AI in rock mechanics. The authors should discuss the computational resources required for training and deploying complex models, the need for specialized expertise, and the potential for model bias. A detailed analysis of these practical considerations would enhance the paper's relevance and utility for practitioners in the field. Furthermore, the paper should clearly articulate its novel contributions. The authors should explicitly state the unique aspects of their review, such as a novel framework for understanding the challenges of AI in rock mechanics or a new perspective on the practical implications of these technologies. This would help to differentiate the paper from other review articles and highlight its significance. Finally, the paper should provide a more detailed discussion on the importance of physics-informed constraints in ML models. The authors should explore how physics-informed constraints can improve the generalization capabilities of ML models, reduce the need for large datasets, and enhance the interpretability of model predictions. Specific examples of how physics-informed constraints have been successfully applied in rock mechanics should be included, along with a discussion of the challenges associated with implementing these constraints in practice. This would provide a more nuanced understanding of the role of physics-informed constraints in the development of reliable and trustworthy AI tools for rock mechanics.
1. How do the authors envision the integration of physics-informed constraints in ML models to address the challenges of data scarcity and model generalization in rock mechanics? Could the authors provide specific examples of how these constraints have been successfully applied in the field and discuss the potential limitations? 2. What are the computational and practical challenges of deploying complex AI models in real-world rock engineering scenarios, and how can these challenges be mitigated? 3. How does the paper's focus on specific ML techniques, such as CNNs and LSTMs, limit its scope, and what alternative approaches, such as reinforcement learning and causal inference, could be explored to address the unique challenges of rock mechanics? 4. In what ways can the paper more clearly articulate its novel contributions, and how can it differentiate itself from other review articles in the field of AI and rock mechanics? 5. How can the paper better address the practical implications of using AI in rock mechanics, particularly in terms of the need for specialized expertise and the potential for model bias, and what strategies can be employed to ensure the responsible and ethical use of AI in this domain?