Functional Concept Proxies and the Actually Smart Hans Problem: What’s Special About Deep Neural Networks in Science
Abstract
AbstractDeep Neural Networks (DNNs) are becoming increasingly important as scientific tools, as they excel in various scientific applications beyond what was considered possible. Yet from a certain vantage point, they are nothing but parametrized functions $$\varvec{f}_{\varvec{\theta }}(\varvec{x})$$
f
θ
(
x
)
of some data vector $$\varvec{x}$$
x
, and their ‘learning’ is nothing but an iterative, algorithmic fitting of the parameters to data. Hence, what could be special about them as a scientific tool or model? I will here suggest an integrated perspective that mediates between extremes, by arguing that what makes DNNs in science special is their ability to develop functional concept proxies (FCPs): Substructures that occasionally provide them with abilities that correspond to those facilitated by concepts in human reasoning. Furthermore, I will argue that this introduces a problem that has so far barely been recognized by practitioners and philosophers alike: That DNNs may succeed on some vast and unwieldy data sets because they develop FCPs for features that are not transparent to human researchers. The resulting breach between scientific success and human understanding I call the ‘Actually Smart Hans Problem’.
Funder
Deutsche Forschungsgemeinschaft
Publisher
Springer Science and Business Media LLC
Subject
General Social Sciences,Philosophy
Reference97 articles.
1. Aad, G., Abajyan, T., Abbott, B., Abdallah, J., Khalek, S. A., Abdelalim, A. A., Aben, R., Abi, B., Abolins, M., AbouZeid, O. S., & Abramowicz, H. (2012). Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC. Physics Letters B, 716(1), 1–29. 2. Albertsson, K., Altoe, P., Anderson, D., Andrews, M., Espinosa, J. P. A., Aurisano, A., Basara, L., Bevan, A., Bhimji, W., Bonacorsi, D., Calafiura P, Campanelli, M., Capps, L., Carminati, F., Carrazza, S., Childers, T., Coniavitis, E., Cranmer, K., David, C., ... Zapata, O. (2018). Machine learning in high energy physics community white paper. Journal of Physics: Conference Series, 1085(2), 022008. 3. Alcorn, M. A., Li, Q., Gong, Z., Wang, C., Mai, L., Ku, W.-S., & Nguyen, A. (2019). Strike (with) a pose: Neural networks are easily fooled by strange poses of familiar objects. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 4845–4854). 4. Bach, S., Binder, A., Montavon, G., Klauschen, F., Müller, K.-R., & Samek, W. (2015). On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation. PLoS ONE, 10(7), e0130140. 5. Baldi, P. (2021). Deep learning in science. Cambridge University Press.
|
|