Geospatial XAI: A Review

Round 1

Reviewer 1 Report

As stated in the title, the proposed paper is a review. As a review, the paper maintains what is promised in the title and in the abstract. The review is well organised from a methodological point of view and, to my knowledge, quite complete. The topic is surely of interest. For these reasons, If the policy of this Journal is to accept review paper, I surely advice for the publication.

I suggest to the authors, just few very small (cosmetic) corrections.

Introduction. Line 3. Few examples, the classical ones, are given of application of geospatial data. Actually, the applications are much more. Please, stress better that the mentioned are just examples.
Figure 1. Use the whole width of the page to enlarge the description of steps.
Page 3. The last two lists of dotted points: compact in a comma separated list.
Table 3. The 'Formula' column is hardly readable: improve the formatting of formulas.
Figure 6. Absolutely not readable, starting with numbering of sub-plots: please, better format it.

Somewhere, some typo is present (for example, 6 lines below Fig. 6): please, clean them.

Best regards

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper reviews state-of-the-art research on geospatial XAI. The topic is interesting. Here are some of my comments:

1. In lines 30-31, page 1, you provide the definition of geospatial XAI, which is not accurate for me. It should aim to help understand the output or predictions made by the AI models, rather than analyzing geospatial data themselves. Please make it more clear. I also put the definition of XAI from Wikipedia here: “Explainable AI (XAI), also known as Interpretable AI, or Explainable Machine Learning (XML), is artificial intelligence (AI) in which humans can understand the reasoning behind decisions or predictions made by the AI.”

2. The research motivation should be described clearly in the introduction. There are already review papers about XAI, which cover the research on the application of XAI to both spatial and non-spatial data. What’s the novelty and contribution of this paper compared with previous studies?

3. Since explainable AI is also known as interpretable AI, interpretable AI should also be included in the key search words. If you search interpretable AI in google scholar, you can see a number of related studies. So I think it should be taken into account in the methodology. And the statistics on the number of studies should be updated accordingly (e.g., Figure 3, Figure 4).

4. In Figure 3a, you mentioned that the orange color represents the upscaled number of studies for the year 2023. How did you get the scale? I didn’t see a constant increase over the years.

5. In lines 189-190, ‘Tosun et al. [30] developed a prototype for breast core biopsy. Together with XAI, the prototype can assist pathologists.’ I didn’t see how this study is linked with geospatial data.

6. For section 3.2 RQ2, it seems pure machine learning theories are introduced in this section. I didn’t see any geospatial components here. Please consider whether this section is really required. Also for section 3.2 RQ3. I think it would be interesting to see how these XAI techniques are applied to geospatial fields to deal with various problems, rather than introducing these XAI techniques since they have been frequently introduced in XAI review papers.

7. A number of typical studies on explainable artificial intelligence are not included in this paper. Please do more literature review to include them. And if you search explainable/interpretable artificial intelligence and GIS, you can find many studies that are not included. Here are some examples:

Samek, W. and Müller, K.R., 2019. Towards explainable artificial intelligence. Explainable AI: interpreting, explaining and visualizing deep learning, pp.5-22.

Arrieta, A.B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., García, S., Gil-López, S., Molina, D., Benjamins, R. and Chatila, R., 2020. Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information fusion, 58, pp.82-115.

Xing, J. and Sieber, R., 2023. The challenges of integrating explainable artificial intelligence into GeoAI. Transactions in GIS.

Antzoulatos, G., Kouloglou, I.O., Bakratsas, M., Moumtzidou, A., Gialampoukidis, I., Karakostas, A., Lombardo, F., Fiorin, R., Norbiato, D., Ferri, M. and Symeonidis, A., 2022. Flood hazard and risk mapping by applying an explainable machine learning framework using satellite imagery and GIS data. Sustainability, 14(6), p.3251.

Dahal, A. and Lombardo, L., 2023. Explainable artificial intelligence in geoscience: A glimpse into the future of landslide susceptibility modeling. Computers & Geosciences, 176, p.105364.

Song, Z., Cao, S. and Yang, H., 2023. Assessment of solar radiation resource and photovoltaic power potential across China based on optimized interpretable machine learning model and GIS-based approaches. Applied Energy, 339, p.121005.

…

I think the search methodology in Figure 2 should be revised to include more key words.

8.Overall, the authors need to do more literature reviews to include the related studies. Besides, it is also necessary to consider the contributions of this study compared with the previous XAI review papers.

The quality of the English language is OK. Some minor typo errors need to be fixed.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The paper (in terms of title and abstract) claims that it reviews the new topic of "Geospatial XAI". With this, I expected a scientific review that explains the topic, describes the previous research and achievements, and concludes and discusses future trends. However, this paper does not go through these steps; instead, it starts by describing a method that has been suggested for review research, and then compares the previous efforts based on some defined criteria (questions).

Although the authors tried to introduce the literature, the output does not seem a scientific review to me, and so I cannot recommend it for publication.

    

The English language of the paper is fine. There are only a few minor issues.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

Very good and necessary paper. Congratulations to the authors.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Authors have applied the suggested comments.

Quality of English language is OK. Minor typo errors need to be fixed.