Reviewer has chosen to be Anonymous

Overall Impression: Good

Content:
Technical Quality of the paper: Good
Originality of the paper: Yes, but limited
Adequacy of the bibliography: Yes

Presentation:
Adequacy of the abstract: Yes
Introduction: background and motivation: Good
Organization of the paper: Satisfactory
Level of English: Satisfactory
Overall presentation: Good

Detailed Comments:

This paper provides a comprehensive survey of recent research in neuro-symbolic systems where Answer Set Programming (ASP) serves as the symbolic reasoning substrate. The authors systematically categorize existing frameworks according to how their neural and symbolic components are pre-trained, hard-coded, or jointly learned. The survey covers the conceptual background of ASP and neural networks, then organizes and analyzes the literature across four integration paradigms.

Reasons to Accept:
- The survey fills an important gap by focusing specifically on ASP-based neurosymbolic frameworks, a rapidly emerging yet under-reviewed subfield of NeSy AI.
- The coverage is extensive and up-to-date, including almost every relevant and recent paper (NeurASP, SLASH, dPASP, NSIL, NeuralFastLAS, GPT-ASP, etc.). Each approach is clearly categorized and contextualized within a coherent taxonomy.
- The article is well-written and structured, and figures effectively summarize architectures and information flow across frameworks.
- The authors critically evaluate weaknesses in existing work, for example, reliance on simple datasets (e.g., MNIST), limited ASP expressivity, and lack of true joint learning.

Revisions Required:
- While individual works are clearly described, the comparisons remain largely qualitative. Incorporating quantitative analyses such as performance metrics and scalability benchmarks would significantly strengthen the survey. A consolidated table summarizing dataset size, learning regime, and reported accuracy could help readers quickly grasp the empirical landscape of neurosymbolic ASP frameworks.
- Relatedly, in the Scalability Issues section, the discussion is primarily qualitative. As a result, it is unclear to what extent these systems actually face scalability bottlenecks. Including quantitative evidence such as runtime statistics, solver timeouts, or memory usage would help ground the discussion in measurable performance rather than remaining at a high conceptual level.

Reviewer has chosen not to be Anonymous

Overall Impression: Average

Content:
Technical Quality of the paper: Average
Originality of the paper: Yes, but limited
Adequacy of the bibliography: Yes, but see detailed comments

Presentation:
Adequacy of the abstract: Yes
Introduction: background and motivation: Good
Organization of the paper: Satisfactory
Level of English: Satisfactory
Overall presentation: Good

Detailed Comments:

This paper presents a survey of neurosymbolic AI approaches that specifically adopt Answer Set Programming (ASP) as the symbolic formalism. The authors introduce ASP, review recent frameworks according to whether neural and symbolic components are pre-defined or learned, and illustrate these systems withgraphical diagrams. The survey highlights the strengths of ASP-based approaches while also acknowledging key limitations in scalability and dependence on hand-crafted knowledge.

The paper has several merits. It is well written and reads easily. Neurosymbolic AI remains a highly varied and fragmented research area, and surveys that clarify specific subfields are valuable. By concentrating on ASP-based methods, the authors provide a structured overview of work that might otherwise be overlooked in broader surveys. The graphical illustrations of the frameworks are especially clear and aid in comparing approaches.

That said, the survey is limited in scope and depth. Restricting attention solely to ASP excludes related and relevant neurosymbolic systems. Without a stronger justification for this focus, the contribution risks being too narrow. Much of the discussion is descriptive rather than analytical. The reader is left without a clear comparative synthesis of why certain methods succeed or fail, what the main trade-offs are, and where the field should head next. A stronger conclusion section could articulate open challenges and research directions to inspire and guide the community.

Some technical and conceptual points would also benefit from clarification. The paper often emphasizes the superior expressivity of ASP over other formalisms, but it is not clear how frequently this added expressivity is practically relevant in NeSy applications. For many tasks, simpler languages such as Datalog are sufficient, and acknowledging this would provide a more balanced perspective.
The section on performance analysis is relatively weak. Claims are not well supported with systematic comparisons, and the survey misses an opportunity to synthesize insights from experimental results across different frameworks. Likewise, the related work section does not situate this survey with respect to existing, broader NeSy surveys. Readers would benefit from a short discussion of how this work complements or differs from prior surveys in the area.

Small remark:
The formal definition of semantics (Section 2) uses sets of ground literals, whereas answer sets are typically defined as sets of atoms.

Conclusion:
The topic of the survey is important, and the paper is a solid starting point. However, in its current form it is too narrow in scope and too descriptive in tone to meet the standards of a leading journal. To be publishable, the paper should (i) broaden or better justify its exclusive focus on ASP, (ii) move beyond description to provide deeper comparative and critical analysis, (iii) expand the conclusion to identify challenges and future directions, and (iv) strengthen the performance and related work sections with references and positioning.