Abstract
Concepts of scale are at the heart of diverse scientific endeavors that seek to understand processes and how observations and analyses influence our understanding. While disciplinary discretions exist, researchers commonly devise spatial, temporal, and organizational scales in scoping phenomena of interest and determining measurements and representational frameworks in research design. The rise of the Fourth Paradigm for science drives data-intensive computing without preconceived notions regarding at what scale the phenomena or processes of interest operate, or at which level of details meaningful patterns may emerge. While scale is the a priori consideration for theory-driven research to seek ontological and relational affirmations, big data analytics and machine learning embed scale in algorithms and model outputs. In this paper, we examine embedded scale in data-driven machine learning research, connect the embedding scale to scale operating in the general theory of geographic representation in GIS and scaffold our arguments with a study of using machine learning to detect archeological features in drone-collected high-density images.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Data availability
Data and codes used in the case study are available at https://doi.org/10.6084/m9.figshare.19166348.
Notes
Unsupervised machine learning clusters data into different groups based on distance measures and does not depend on labeled data. Recent developments in self-supervised machine learning attempts to reduce the reliance on massive, labeled data.
A “feature” is a special term used by archeology to indicate any physical structure or element that is made or altered by humans, is not portable, cannot be removed from a site, and must be studied in context (ref. Glossary at Archaeological Institute of America—https://archaeological.org). To avoid conflating with the term “feature” in image processing and CNN feature maps, we use “archaeological features” for the physical structures noted on Kvamme’s map.
References
Ahl V, Allen TFH (1996) Hierarchy theory: a vision, vocabulary, and epistemology. Columbia University Press
Anderson C (2008). The end of theory: data deluge makes the scientific method obsolete. Wired Mag. pp 10–12
Basaeed E, Bhaskar H, Hill P, Al-Mualla M, Bull D (2016) A supervised hierarchical segmentation of remote-sensing images using a committee of multi-scale convolutional neural networks. Int J Remote Sens 37(7):1671–1691. https://doi.org/10.1080/01431161.2016.1159745
Brown M, Lowe DG (2005) Unsupervised 3D object recognition and reconstruction in unordered datasets. In: Proceedings of the international conference on 3D digital imaging and modelling (pp. 56–63). https://doi.org/10.1109/3DIM.2005.81
De Reu J, Bourgeois J, Bats M, Zwertvaegher A, Gelorini V, De Smedt P, Chu W, Antrop M, De Maeyer P, Finke P, Van Miervenne M, Verniers J, Crombé P (2013) Application of the topographic position index to heterogeneous landscapes. Geomorphology 186:39–49
ESRI (2021) Curvature function for ArcPro 2.7. Retrieved April 20, 2021 from https://pro.arcgis.com/en/pro-app/latest/help/analysis/raster-functions/curvature-function.htm
Gong Y, Wang L, Guo R, Lazebnik S (2014) Multi-scale orderless pooling of deep convolutional activation features. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8695 LNCS(PART 7), 392–407. https://doi.org/10.1007/978-3-319-10584-0_26
Goodchild MF, Yuan M, Cova TJ (2007) Towards a general theory of geographic representation in GIS. Int J Geogr Inf Sci 21(3):239–260. https://doi.org/10.1080/13658810600965271
Harris TM (2006) Scale as artifact: GIS, ecological fallacy, and archaeological analysis. In: GR Lock, B Molyneaux (Eds.) Confronting scale in archaeology: issues of theory and practice. pp. 39–53
He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp 770–778
Hey T, Tansley S, Tolle K (2009) The fourth paradigm: data-intensive scientific discovery. Microsoft Research
Jadon, S. (2020). A survey of loss functions for semantic segmentation. 2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2020, 1–7. https://doi.org/10.1109/CIBCB48159.2020.9277638
Jungers WL, Falsetti AB, Wall CE (1995) Shape, relative size, and size-adjustments in morphometrics. Am J Phys Anthropol 38(S21):137–161. https://doi.org/10.1002/ajpa.1330380608
Krizhevsky, B. A., Sutskever, I., and Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Proceedings of the 25th International Conference on Neural Information Processing Systems., 1097–1105. https://doi.org/10.1145/3065386
Kvamme KL (2018) Experiments in the automatic detection of archaeological features in remotely sensed data from Great Plains villages. CAA, USA, p 2016
Kvamme KL, Ernenwein EG, Markussen CJ (2006) Robotic total station for microtopographic mapping: an example from the Northern Great Plains. Archaeol Prospect 13(2):91–102. https://doi.org/10.1002/arp.270
LeCun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86(11):2278–2323. https://doi.org/10.1109/5.726791
Lindsay JB, Cockburn JMH, Russell HAJ (2015) An integral image approach to performing multi-scale topographic position analysis. Geomorphology 245:51–61
Lowe DG (2004) Distinctive image features from scale-invariant keypoints. Int J Comput Vision 60(2):91–110. https://doi.org/10.1023/B:VISI.0000029664.99615.94
Mandelbrot BB, Frame M (1987) Fractals. Encycl Phys Sci Technol 5:579–593
Mitchell MD (2011) Continuity and change in the organization of Mandan craft production, 1400–1750 (Publication Number 3453760) [Ph.D., University of Colorado at Boulder]. ProQuest Dissertations and Theses Global. Ann Arbor
Msonda P, Uymaz SA, Karaaǧaç SS (2020) Spatial pyramid pooling in deep convolutional networks for automatic tuberculosis diagnosis. Traitement Du Signal 37(6):1075–1084
O’Neill RV, Johnson AR, King AW (1989) A hierarchical framework for the analysis of scale. Landscape Ecol 3(3–4):193–205. https://doi.org/10.1007/BF00131538
Peters-Lidard CD, Clark M, Samaniego L, Verhoest NEC, Van Emmerik T, Uijlenhoet R, Achieng K, Franz TE, Woods R (2017) Scaling, similarity, and the fourth paradigm for hydrology. Hydrol Earth Syst Sci 21(7):3701–3713. https://doi.org/10.5194/hess-21-3701-2017
Ronneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. Int Conf Med Image Comput Comput Assist Interv. https://doi.org/10.1007/978-3-319-24574-4
Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. In: 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings. pp 1–14
Verhoeven G (2007) Becoming a NIR-sensitive aerial archaeologist. In: Remote Sensing for Agriculture, Ecosystems, and Hydrology IX (Vol. 6742, p. 67420Y). International Society for Optics and Photonics
Verhoeven GJ, Smet PF, Poelman D, Vermeulen F (2009) Spectral characterization of a digital still camera’s NIR modification to enhance archaeological observation. IEEE Trans Geosci Remote Sens 47(10):3456–3468. https://doi.org/10.1109/TGRS.2009.2021431
Wang S, Zhang X, Ye P, Du M, Lu Y, Xue H (2019) Geographic knowledge graph (GeoKG): a formalized geographic knowledge representation. ISPRS Int J Geo Inf 8(4):184
Wojna Z, Ferrari V, Guadarrama S, Silberman N, Chen LC, Fathi A, Uijlings J (2017) The devil is in the decoder. British Machine Vision Conference 2017, BMVC 2017, 1–13. https://doi.org/10.5244/c.31.10
Wood RW (1967) An interpretation of mandan culture history. Smithsonian Institution
Wu J, Li H (2006) Concepts of scale and scaling. Scaling Uncertain Anal Ecol Methods Appl. https://doi.org/10.1007/1-4020-4663-4_1
Yan B, Janowicz K, Mai G, Zhu R (2019) A spatially explicit reinforcement learning model for geographic knowledge graph summarization. Trans GIS 23(3):620–640
Acknowledgements
We are indebted to the numerous individuals who have contributed to this paper. The PaleoCultural Research Group and the State Historical Society of North Dakota provided access to the Huff Indian Village State Historic Site and other sites surveyed for this project. Dr. Ken Kvamme graciously provided access and guidance for using his data from the site. We also thank the anonymous reviewers who’s comments through the review process made this paper stronger.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Yuan, M., McKee, A. Embedding scale: new thinking of scale in machine learning and geographic representation. J Geogr Syst 24, 501–524 (2022). https://doi.org/10.1007/s10109-022-00378-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10109-022-00378-6
Keywords
- Scale
- Machine learning
- Geographic representation
- High-density measurements
- Archeological feature extraction