Empiricism without magic: transformational abstraction in deep convolutional neural networks
Open accessThe DOI matches its Crossref record; 96 of the 98 references it cites were found in a registry or an index.
Abstract
In artificial intelligence, recent research has demonstrated the remarkable potential of Deep Convolutional Neural Networks (DCNNs), which seem to exceed state-of-the-art performance in new domains weekly, especially on the sorts of very difficult perceptual discrimination tasks that skeptics thought would remain beyond the reach of artificial intelligence. However, it has proven difficult to explain why DCNNs perform so well. In philosophy of mind, empiricists have long suggested that complex cognition is based on information derived from sensory experience, often appealing to a faculty of abstraction. Rationalists have frequently complained, however, that empiricists never adequately explained how this faculty of abstraction actually works. In this paper, I tie these two questions together, to the mutual benefit of both disciplines. I argue that the architectural features that distinguish DCNNs from earlier neural networks allow them to implement a form of hierarchical processing that I call “transformational abstraction”. Transformational abstraction iteratively converts sensory-based representations of category exemplars into new formats that are increasingly tolerant to “nuisance variation” in input. Reflecting upon the way that DCNNs leverage a combination of linear and non-linear processing to efficiently accomplish this feat allows us to understand how the brain is capable of bi-directional travel between exemplars and abstractions, addressing longstanding problems in empiricist philosophy of mind. I end by considering the prospects for future research on DCNNs, arguing that rather than simply implementing 80s connectionism with more brute-force computation, transformational abstraction counts as a qualitatively distinct form of processing ripe with philosophical and psychological significance, because it is significantly better suited to depict the generic mechanism responsible for this important kind of psychological processing in the brain.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“As Buckner [5] argues, contemporary AI systems operate in terms of empiricism by abstracting complex patterns from data bases, in which performance improvements are largely driven not by conceptual breakthroughs in representation or reasoning, but by the scale and character of empirical data.”
From 4E cognition and the coevolution of human–AI interaction (2025) · cites this paper as [5]
Citations
Cites
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
Human-level control through deep reinforcement learning
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. (2015) · Nature · doi:10.1038/nature14236 · Cited by 30,357
Reducing the dimensionality of data with neural networks
Hinton, G. E., & Salakhutdinov, R. R. (2006) · Science · doi:10.1126/science.1127647 · Cited by 20,928
Deep learning in neural networks: an overview
Schmidhuber, J. (2015) · Neural Networks · doi:10.1016/j.neunet.2014.09.003 · Cited by 18,570 · cited 4×
“…was the way it com- 7 In the interests of space, we move quickly over the history here; for more background and discussion, see (Buckner and Garson 2018; Schmidhuber 2015). bined two different types of operation—linear convolutional filtering and nonlinear downsampling—in a single network.”
BackgroundMethodInfluentialMastering the game of Go with deep neural networks and tree search
Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al. (2016) · Nature · doi:10.1038/nature16961 · Cited by 15,612
Approximation by superpositions of a sigmoidal function
Cybenko, G. (1989) · Mathematics of Control, Signals, and Systems · doi:10.1007/bf02551274 · Cited by 13,883 · cited 1×
“…of mathematical abstraction, all neural networks perform pattern recognition or dimensionality reduction, and the universal approximation theory shows that in principle a three-layer perceptron with an unlimited number of nodes can approximate any continuous function (Cybenko 1989; Hornik 1991).”
BackgroundReceptive fields, binocular interaction and functional architecture in the cat’s visual cortex
Hubel, D. H., & Wiesel, T. N. (1962) · The Journal of Physiology · doi:10.1113/jphysiol.1962.sp006837 · Cited by 13,680 · cited 1×
“This traditionwas inspired by an anatomical discovery in cat visual cortex by Hubel and Wiesel (1962).”
MethodBackpropagation Applied to Handwritten Zip Code Recognition
LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989) · Neural Computation · doi:10.1162/neco.1989.1.4.541 · Cited by 11,891 · cited 1×
“Despite some early successes (LeCun et al. 1989), when backpropagation was applied to such deep networks, learning tended to settle into poor localminima or fail to stabilize at all (DeMers andCottrell 1993).”
Background(2016) · Cited by 11,679
Explaining and Harnessing Adversarial Examples
Goodfellow, I. J., Shlens, J., & Szegedy, C. (2014) · arXiv · doi:10.48550/arxiv.1412.6572 · Cited by 10,381 · cited 1×
“Moreover, subsequent research has shown that rather than being an idiosyncratic and easily-eliminated quirk, adversarials are counterintuitively robust: they can be built without god’s-eye access to model parameters, they transfer (with labels) to DCNNs with very different architectures and training sets (and in fact…”
BackgroundPlaying Atari with Deep Reinforcement Learning
Mnih, V., Kavukcuoglu, K., Silver, D., Graves, A., Antonoglou, I., Wierstra, D., & Riedmiller, M. (2013) · arXiv · doi:10.48550/arxiv.1312.5602 · Cited by 9,471 · cited 1×
“This is how some of the most prominent DCNN-based models were trained, such as AlphaGo or DeepMind’s DQN video-game playing system, which can play Atari 2600 games at expert levels of performance (Mnih et al. 2013; Silver et al. 2016).”
MethodMastering the game of Go without human knowledge
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., et al. (2017) · Nature · doi:10.1038/nature24270 · Cited by 9,084
Barsalou, L. W. (1999) · Behavioral and Brain Sciences · doi:10.1017/s0140525x99002149 · Cited by 6,887
Approximation capabilities of multilayer feedforward networks
Hornik, K. (1991) · Neural Networks · doi:10.1016/0893-6080(91)90009-t · Cited by 6,091 · cited 1×
“…of mathematical abstraction, all neural networks perform pattern recognition or dimensionality reduction, and the universal approximation theory shows that in principle a three-layer perceptron with an unlimited number of nodes can approximate any continuous function (Cybenko 1989; Hornik 1991).”
BackgroundImage Style Transfer Using Convolutional Neural Networks
Gatys, L. A., Ecker, A. S., & Bethge, M. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.265 · Cited by 6,025
Fukushima, K. (1980) · Biological Cybernetics · doi:10.1007/bf00344251 · Cited by 5,759
D., C. P., Rosch, E., & Loyd, B. B. (1979) · The American Journal of Psychology · doi:10.2307/1421578 · Cited by 5,264
Intelligence without representation
Brooks, R. A. (1991) · Artificial Intelligence · doi:10.1016/0004-3702(91)90053-m · Cited by 4,853 · cited 1×
“…ambitions will score models that explain a wider range of phenomena as possessing more explanatory power, all other things being equal (Boyd 1999; Buckner 2011; Stinson 2016, 2017; Ylikoski and Kuorikoski 2010), whereas others of more naturalistic inclination leave such matters entirely to the…”
Background(2002) · Foundations of Cognitive Psychology · doi:10.7551/mitpress/3080.003.0017 · Cited by 3,953
(1989) · neural information processing systems · Cited by 3,736
Striving for Simplicity: The All Convolutional Net
Springenberg, J. T., Dosovitskiy, A., Brox, T., & Riedmiller, M. (2014) · arXiv · doi:10.48550/arxiv.1412.6806 · Cited by 3,447 · cited 1×
“For two poles in this debate, see (Patel et al. 2016; Springenberg et al. 2014).”
BackgroundMachamer, P., Darden, L., & Craver, C. F. (2000) · Philosophy of Science · doi:10.1086/392759 · Cited by 3,179
Building machines that learn and think like people
Lake, B. M., Ullman, T. D., Tenenbaum, J. B., & Gershman, S. J. (2017) · Behavioral and Brain Sciences · doi:10.1017/s0140525x16001837 · Cited by 2,550
Building Machines That Learn and Think Like People
Lake, B. M., Ullman, T. D., Tenenbaum, J. B., & Gershman, S. J. (2016) · arXiv · doi:10.48550/arxiv.1604.00289 · Cited by 2,550
How does the brain solve visual object recognition?
DiCarlo, J. J., Zoccolan, D., & Rust, N. C. (2012) · Neuron · doi:10.1016/j.neuron.2012.01.010 · Cited by 1,790 · cited 1×
“Following several other analyses (DiCarlo et al. 2012; Montúfar et al. 2014; Patel et al. 2016; Schmidhuber 2015), I characterize the computational core of these networks in terms of three features: (1) many layers of hierarchical processing which”
MethodAdvances in Neural Information Processing Systems 19 (NIPS 2006)
Simonotto, E., Whalley, H. C., Lawrie, S. M., Murray, L. M., McGonigle, D. J., & Storkey, A. (2006) · Cited by 1,651
Conceptual spaces: the geometry of thought
(2000) · Choice Reviews Online · doi:10.5860/choice.38-1526 · Cited by 1,643
Deconvolution and Checkerboard Artifacts
Odena, A., Dumoulin, V., & Olah, C. (2016) · Distill · doi:10.23915/distill.00003 · Cited by 1,640
Using goal-driven deep learning models to understand sensory cortex
Yamins, D. L. K., & DiCarlo, J. J. (2016) · Nature Neuroscience · doi:10.1038/nn.4244 · Cited by 1,621
Neuroscience-Inspired Artificial Intelligence
Hassabis, D., Kumaran, D., Summerfield, C., & Botvinick, M. (2017) · Neuron · doi:10.1016/j.neuron.2017.06.011 · Cited by 1,557
Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit
Hahnloser, R. H. R., Sarpeshkar, R., Mahowald, M. A., Douglas, R. J., & Seung, H. S. (2000) · Nature · doi:10.1038/35016072 · Cited by 1,293
Rosch, E. (2024) · Cognition and Categorization · doi:10.4324/9781032633275-4 · Cited by 1,231
Deep supervised, but not unsupervised, models may explain IT cortical representation
Khaligh-Razavi, S.-M., & Kriegeskorte, N. (2014) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1003915 · Cited by 1,204
Explanation: a mechanist alternative
Bechtel, W., & Abrahamsen, A. (2005) · Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences · doi:10.1016/j.shpsc.2005.03.010 · Cited by 1,138
A Neurocomputational Perspective: The Nature of Mind and the Structure of Science.
Baker, L. R., & Churchland, P. M. (1992) · The Philosophical Review · doi:10.2307/2185955 · Cited by 1,043 · cited 1×
“…stream visual processing in the mammalian brain by neuroscientists; to articulate this solution, we need to examine the notions of a perceptual similarity space and category manifolds as regions therein (for development and critical examination, see Churchland 1989; Gärdenfors 2004; Gauker 2011).”
BackgroundUntangling invariant object recognition
DiCarlo, J. J., & Cox, D. D. (2007) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2007.06.010 · Cited by 959
Rethinking Mechanistic Explanation
Glennan, S. (2002) · Philosophy of Science · doi:10.1086/341857 · Cited by 867 · cited 1×
“15 This general characterization washes across differences in various canonical accounts of mechanism; see (Bechtel and Abrahamsen 2005; Glennan 2002; Machamer et al. 2000).”
BackgroundA treatise concerning the principles of human knowledge, 1734
Berkeley, G. (1971) · Cited by 841
On the Number of Linear Regions of Deep Neural Networks
Montúfar, G., Pascanu, R., Cho, K., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1402.1869 · Cited by 840 · cited 2×
“Following several other analyses (DiCarlo et al. 2012; Montúfar et al. 2014; Patel et al. 2016; Schmidhuber 2015), I characterize the computational core of these networks in terms of three features: (1) many layers of hierarchical processing which interpolates two different kinds of computational…”
MethodRandom synaptic feedback weights support error backpropagation for deep learning
Lillicrap, T. P., Cownden, D., Tweed, D. B., & Akerman, C. J. (2016) · Nature Communications · doi:10.1038/ncomms13276 · Cited by 812
(2014) · Progress in Molecular Biology and Translational Science · doi:10.1016/c2012-0-00157-3 · Cited by 799
Machery, E. (2009) · Oxford University Press eBooks · doi:10.1093/acprof:oso/9780195306880.001.0001 · Cited by 784 · cited 1×
“…the success of DCNNs requires us to analyze their success at categorization, I do not here attempt to solve the more ambitious problem of showing how they might implement a Fregean notion of concepts, which requires intersubjective agreement and objective standards of correctness (Machery 2009).”
BackgroundWhat Learning Systems do Intelligent Agents Need? Complementary Learning Systems Theory Updated
Kumaran, D., Hassabis, D., & McClelland, J. L. (2016) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2016.05.004 · Cited by 759
(1969) · Ontological Relativity and Other Essays · doi:10.7312/quin92204-004 · Cited by 752
Learning to generate chairs with convolutional neural networks
Dosovitskiy, A., Springenberg, J. T., & Brox, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298761 · Cited by 640
Integrating psychology and neuroscience: functional analyses as mechanism sketches
Piccinini, G., & Craver, C. (2011) · Synthese · doi:10.1007/s11229-011-9898-4 · Cited by 508 · cited 1×
“First, we could say that the DCNNs are “mechanism sketches” of human perceptual neocortex (Kaplan and Craver 2011; Piccinini and Craver 2011), specifically where the synaptic connections of simple and complex cells are black-boxed rather than depicted with structural fidelity.”
BackgroundEquilibrium Propagation: Bridging the Gap between Energy-Based Models and Backpropagation
Scellier, B., & Bengio, Y. (2017) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2017.00024 · Cited by 477 · cited 1×
“…those involved in DCNN research) that randomized error signals can work as well as individuallycomputed ones (Lillicrap et al. 2016) and that biologically-plausible spike-timing dependent plasticity can produce outcomes which are functionally similar to backpropagation (Scellier and Bengio 2017).”
BackgroundSejnowski, T. J., Koch, C., & Churchland, P. S. (1988) · Science · doi:10.1126/science.3045969 · Cited by 458
Microcognition: Philosophy, Cognitive Science, and Parallel Distributed Processing.
Higginbotham, J., & Clark, A. (1994) · The Philosophical Quarterly · doi:10.2307/2220153 · Cited by 453 · cited 1×
“…by the standards of the more classical, rule-basedmethods they aimed to supplant: soft constraints, distributed representation, gradual learning, graceful degradation, and the importance of reproducing subject’s errors as well as their successes (for review and discussion, see Clark 1989).”
BackgroundMicrocognition: Philosophy, Cognitive Science, and Parallel Distributed Processing
Clark, A. (1989) · The MIT Press eBooks · doi:10.7551/mitpress/4597.001.0001 · Cited by 451
Kaplan, D. M., & Craver, C. F. (2011) · Philosophy of Science · doi:10.1086/661755 · Cited by 378
Rajalingham, R., Issa, E. B., Bashivan, P., Kar, K., Schmidt, K., & DiCarlo, J. J. (2018) · The Journal of Neuroscience · doi:10.1523/jneurosci.0388-18.2018 · Cited by 358 · cited 1×
“…similarity judgments and do so using the same sorts of features that are found in late-stage ventral stream processing in V4/5 and IT, they found that these models were not as predictive of high-resolution, image-by-image comparisons in humans as were rhesus monkeys (Rajalingham et al. 2018).”
ResultExplicit information for category-orthogonal object properties increases along the ventral stream
Hong, H., Yamins, D. L. K., Majaj, N. J., & DiCarlo, J. J. (2016) · Nature Neuroscience · doi:10.1038/nn.4247 · Cited by 337
Non-Linear Dimensionality Reduction
DeMers, D., & Cottrell, G. W. (1992) · neural information processing systems · Cited by 309
Letting structure emerge: connectionist and dynamical systems approaches to cognition
McClelland, J. L., Botvinick, M. M., Noelle, D. C., Plaut, D. C., Rogers, T. T., Seidenberg, M. S., & Smith, L. B. (2010) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2010.06.002 · Cited by 305
Ylikoski, P., & Kuorikoski, J. (2010) · Philosophical Studies · doi:10.1007/s11098-008-9324-z · Cited by 273
Emergence of Invariance and Disentanglement in Deep Representations
Achille, A., & Soatto, S. (2018) · 2018 Information Theory and Applications Workshop (ITA) · doi:10.1109/ita.2018.8503149 · Cited by 267 · cited 2×
“6 Achille and Soatto (2017) have recently argued that implicit or explicit regularization is a fourth crucially important feature in generalizing DCNN performance (to prevent them from simply memorizing the mapping for every exemplar in the training set), but since there is significant diversity in…”
BackgroundChatterjee, A. (2010) · Language and Cognition · doi:10.1515/langcog.2010.004 · Cited by 260 · cited 1×
“This provides a visual example of the sort of transformed, pared-down, late-stage representation that has often been cited as the essence of abstraction in artistic and scientific creativity (Camp 2015; Chatterjee 2010).”
BackgroundAbstraction and the Organization of Mechanisms
Levy, A., & Bechtel, W. (2013) · Philosophy of Science · doi:10.1086/670300 · Cited by 257
Parallel Distributed Processing at 25: Further Explorations in the Microstructure of Cognition
Rogers, T. T., & McClelland, J. L. (2014) · Cognitive Science · doi:10.1111/cogs.12148 · Cited by 233
Predicting Deeper into the Future of Semantic Segmentation
Luc, P., Neverova, N., Couprie, C., Verbeek, J., & LeCun, Y. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.77 · Cited by 231 · cited 1×
“…instead of static exemplars, noting that interactions with the same object from slightly different perspectives could play the role of many different training exemplars, and that the attempt to predict the next frames could obviate the need for manually-labeled training data (Luc et al. 2017).”
BackgroundAre More Details Better? On the Norms of Completeness for Mechanistic Explanations
Craver, C. F., & Kaplan, D. M. (2020) · The British Journal for the Philosophy of Science · doi:10.1093/bjps/axy015 · Cited by 225 · cited 2×
“…patterns of filter and max-pooling nodes in a DCNN; but it is a misunderstanding of mechanistic explanations to suppose that they must always aim for this kind of maximal detail irrespective of other factors (Boone and Piccinini 2016; Craver and Kaplan 2018; Levy and Bechtel 2013; Stinson 2017).”
BackgroundThe contribution of spike threshold to the dichotomy of cortical simple and complex cells
Priebe, N. J., Mechler, F., Carandini, M., & Ferster, D. (2004) · Nature Neuroscience · doi:10.1038/nn1310 · Cited by 217 · cited 1×
“Specifically, the bimodality of the firing patterns appears to be explained instead by an interaction between unimodal connectivity patterns and those cells’ nonlinearly thresholded membrane potentials (Priebe et al. 2004).”
BackgroundKinds, Complexity and Multiple realization
Boyd, R. (1999) · Philosophical Studies · doi:10.1023/a:1004511407133 · Cited by 202
Gobet, F., & Simon, H. A. (1996) · Psychological Science · doi:10.1111/j.1467-9280.1996.tb00666.x · Cited by 191
Models and mechanisms in psychological explanation
Weiskopf, D. A. (2011) · Synthese · doi:10.1007/s11229-011-9958-9 · Cited by 185 · cited 1×
“…implement a form of hierarchical abstraction that reduces the complexity of a problem’s feature space (and 8 Note that some functionalists (i.e. Weiskopf 2011a, b) defend the explanatory power of idealized models and so may not think much is gained by restricting DCNNs to more…”
BackgroundConnectionist models and psychological evidence
McClelland, J. L. (1988) · Journal of Memory and Language · doi:10.1016/0749-596x(88)90069-1 · Cited by 173 · cited 1×
“…with actual brains down to the level of individual neurons and their connections, but rather to reveal the brain’s “basic principles of processing” (McClelland 1988, p. 107) by exploring “computational constraints that govern the design of the nervous system” (Sejnowski et al. 1988, p. 1300).”
BackgroundBlundell, C., Uria, B., Pritzel, A., Li, Y., Ruderman, A., Leibo, J. Z., Rae, J., Wierstra, D., & Hassabis, D. (2016) · arXiv · doi:10.48550/arxiv.1606.04460 · Cited by 162
What you learn is what you see: using eye movements to study infant cross‐situational word learning
Yu, C., & Smith, L. B. (2011) · Developmental Science · doi:10.1111/j.1467-7687.2010.00958.x · Cited by 161 · cited 1×
“Some argue that the amount of stimuli to which children and animals are normally exposed is often underestimated, perhaps bolstering their argument with the kind of data collected from head-cams attached to infants by developmental psychologists like Linda Smith (Yu and Smith 2011).”
BackgroundFeature integration and object representations along the dorsal stream visual hierarchy
Perry, C. J., & Fallah, M. (2014) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2014.00084 · Cited by 154
Neocognitron for handwritten digit recognition
Fukushima, K. (2003) · Neurocomputing · doi:10.1016/s0925-2312(02)00614-8 · Cited by 149
Conceptual Spaces - The Geometry of Thought.
Schwering, A. (2008) · Künstliche Intell. · Cited by 109
Boone, W., & Piccinini, G. (2016) · Philosophy of Science · doi:10.1086/687855 · Cited by 107 · cited 2×
“…patterns of filter and max-pooling nodes in a DCNN; but it is a misunderstanding of mechanistic explanations to suppose that they must always aim for this kind of maximal detail irrespective of other factors (Boone and Piccinini 2016; Craver and Kaplan 2018; Levy and Bechtel 2013; Stinson 2017).”
BackgroundA neurocomputational perspective the nature of mind and the structure of science
Churchland, P. M. (1992) · Cited by 104
Reading into neuronal oscillations in the visual system: implications for developmental dyslexia
Vidyasagar, T. R. (2013) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2013.00811 · Cited by 93
Gauker, C. (2011) · Oxford University Press eBooks · doi:10.1093/acprof:oso/9780199599462.001.0001 · Cited by 85 · cited 6×
“One popular idea is that this involves “leaving out” features which are irrelevant to the figures’ triangularity—an approach that Gauker (2011) explicates as “abstraction-as-subtraction”—but explaining how humans and animals can even detect the abstract features in specific particulars has proven…”
InfluentialThe Routledge Handbook of the Computational Mind
Sprevak, M. D., & Colombo, M. (2018) · Routledge eBooks · doi:10.4324/9781315643670 · Cited by 77
A Quinean Critique of Ostrich Nominalism
Pickel, B., & Mantegani, N. (2012) · Edinburgh Research Explorer (University of Edinburgh) · Cited by 76
Concept Nativism and Neural Plasticity
Laurence, S., & Margolis, E. (2015) · The MIT Press eBooks · doi:10.7551/mitpress/9383.003.0010 · Cited by 72 · cited 1×
“Second and relatedly, some of the terms of the great debate will be due for renovation; for example, we should agree here with Laurence and Margolis (2015) that if empiricists must hold that the mind begins as a blank slate lacking any structure whatsoever, then, for quite some time, there shall…”
BackgroundMechanisms in psychology: ripping nature at its seams
Stinson, C. (2015) · Synthese · doi:10.1007/s11229-015-0871-5 · Cited by 72 · cited 1×
“…score models that explain a wider range of phenomena as possessing more explanatory power, all other things being equal (Boyd 1999; Buckner 2011; Stinson 2016, 2017; Ylikoski and Kuorikoski 2010), whereas others of more naturalistic inclination leave such matters entirely to the preference of…”
BackgroundThe Functional Unity of Special Science Kinds
Weiskopf, D. A. (2011) · The British Journal for the Philosophy of Science · doi:10.1093/bjps/axq026 · Cited by 70 · cited 1×
“…implement a form of hierarchical abstraction that reduces the complexity of a problem’s feature space (and 8 Note that some functionalists (i.e. Weiskopf 2011a, b) defend the explanatory power of idealized models and so may not think much is gained by restricting DCNNs to more…”
BackgroundA Probabilistic Framework for Deep LearningPreprint
Patel, A., Nguyen, T. M., & Baraniuk, R. G. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1612.01936 · Cited by 67 · cited 2×
“Following several other analyses (DiCarlo et al. 2012; Montúfar et al. 2014; Patel et al. 2016; Schmidhuber 2015), I characterize the computational core of these networks in terms of three features: (1) many layers of hierarchical processing which interpolates two different kinds of computational…”
BackgroundMethodTwo Approaches to the Distinction between Cognition and ‘Mere Association’
Buckner, C. J. (2011) · International Journal of Comparative Psychology · doi:10.46867/ijcp.2011.24.04.06 · Cited by 64 · cited 1×
“…ambitions will score models that explain a wider range of phenomena as possessing more explanatory power, all other things being equal (Boyd 1999; Buckner 2011; Stinson 2016, 2017; Ylikoski and Kuorikoski 2010), whereas others of more naturalistic inclination leave such matters entirely to the…”
BackgroundSelf-citationLogical Concepts and Associative Characterizations
Camp, E. (2015) · The MIT Press eBooks · doi:10.7551/mitpress/9383.003.0034 · Cited by 63
Building machines that learn and think for themselves
Botvinick, M., Barrett, D. G. T., Battaglia, P., de Freitas, N., Kumaran, D., Leibo, J. Z., Lillicrap, T., Modayil, J., Mohamed, S., Rabinowitz, N. C., et al. (2017) · Behavioral and Brain Sciences · doi:10.1017/s0140525x17000048 · Cited by 54
Cognitive Psychology for Deep Neural Networks: A Shape Bias Case Study
Ritter, S., Barrett, D. G. T., Santoro, A., & Botvinick, M. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1706.08606 · Cited by 54 · cited 1×
“…they produce similar rank-orders for the difficulty of each category to learn and the mutual similarity of typical exemplars, and they even display some of the same systemic learning biases as human children, such as preferring to categorize objects by shape rather than color (Ritter et al. 2017).”
BackgroundAbstraction and the Origin of General Ideas
Laurence, S., & Margolis, E. (2012) · PhilPapers (PhilPapers Foundation) · Cited by 43
Notions of Invariance for Abstraction Principles
Antonelli, G. A. (2010) · Philosophia Mathematica · doi:10.1093/philmat/nkq010 · Cited by 42
Explanation and connectionist models
Stinson, C. (2018) · Routledge eBooks · doi:10.4324/9781315643670-10 · Cited by 36 · cited 1×
“…connectionists typically aim both to reproduce important patterns in the target phenomenon and to do so using biologically plausible constraints—“as though they are deploying an inferential pincer movement” to narrow the space of possible explanations from two directions at once (Stinson 2018).”
BackgroundÜBER LOCKES „ALLGEMEINES DREIECK”
Beth, E. W. (1957) · Kant-Studien · doi:10.1515/kant.1957.48.1-4.361 · Cited by 31
Functional kinds: a skeptical look
Buckner, C. J. (2014) · Synthese · doi:10.1007/s11229-014-0606-z · Cited by 12 · cited 1×
“…feature space (and 8 Note that some functionalists (i.e. Weiskopf 2011a, b) defend the explanatory power of idealized models and so may not think much is gained by restricting DCNNs to more biologically-plausible values (though for rebuttal from a more mechanistic perspective, see Buckner 2015).”
BackgroundSelf-citationDocument Classification with Deep Rectifier Neural Networks and Probabilistic Sampling
Grósz, T., & István Nagy, T. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-10816-2_14 · Cited by 12
Recognition and classification of geometric shapes using neural networks
Spasojevic, S. S., Susic, M. Z., & Đurović, Ž. (2012) · Symposium on Neural Network Applications in Electrical Engineering · doi:10.1109/neurel.2012.6419966 · Cited by 11 · cited 1×
“…to imagine that a couple of children cannot prate together of their sugar-plums and 4 Even three-layer perceptrons have been trained to categorize triangle exemplars with a high degree of accuracy (Spasojević et al. 2012). rattles…till they have first tacked together numberless inconsistencies?””
BackgroundPredicting human adiposity – sometimes – with food insecurity: Broaden the model for better accuracy
Hill, S. E., Leyva, R. P. P., & DelPriore, D. J. (2017) · Behavioral and Brain Sciences · doi:10.1017/s0140525x16001448 · Cited by 5
Back to the Cradle: Mechanism Schemata from Piaget to DNA
Stinson, C. (2017) · The Western Ontario series in philosophy of science · doi:10.1007/978-3-319-52768-0_13 · Cited by 4 · cited 3×
“…patterns of filter and max-pooling nodes in a DCNN; but it is a misunderstanding of mechanistic explanations to suppose that they must always aim for this kind of maximal detail irrespective of other factors (Boone and Piccinini 2016; Craver and Kaplan 2018; Levy and Bechtel 2013; Stinson 2017).”
Background
Research works printed but not linked: 6
- Elsayed, G. F., Shankar, S., Cheung, B., Papernot, N., Kurakin, A., Goodfellow, I., & Sohl-Dickstein, J. (2018). Adversarial examples that fool both human and computer vision. arXiv Preprint arXiv:1802.08195 . · arXiv
- Gray, H. (1918). Anatomy of the human body, rev. and re-edited by Warren H. Lewis. Philadelphia: Lea & Febiger.
- Hochreiter, S. (1991). Untersuchungen zu dynamischen neuronalen Netzen. Diploma thesis for Technische Universität München, München.
- Hume, D. (1739). A treatise on human nature. Oxford: Oxford University Press.
- Marcus, G. (2018). Deep learning: A critical appraisal. arXiv:1801.00631 [cs, Stat]. · arXiv
- Quine, W. V. (1971). Epistemology naturalized. Akten Des XIV. Internationalen Kongresses Für Philosophie, 6, 87–103.
Cited by
Solving the Black Box Problem: A Normative Framework for Explainable Artificial Intelligence
Zednik, C. (2021) · Philosophy & Technology · doi:10.1007/s13347-019-00382-7 · Cited by 419
“Rather, the system is more likely to deploy a taxonomy of abstract categories in which the inputs are combined nonlinearly (see also Buckner 2018).”
BackgroundInfluentialUnderstanding from Machine Learning Models
Sullivan, E. (2022) · The British Journal for the Philosophy of Science · doi:10.1093/bjps/axz035 · Cited by 221
“See Buckner ([2018]) for a philosophical argument that DNNs (CNNs in particular) capture processes of the brain.”
BackgroundDeep learning: A philosophical introduction
Buckner, C. (2019) · Philosophy Compass · doi:10.1111/phc3.12625 · Cited by 152
“…knowledge, a faculty that has frequently been invoked by empiricists to account for learners' transition from basic sensory associations to the abstract representations deployed in higher cognition but has until recently not been adequately explained (c.f. Buckner, 2018; Laurence & Margolis, 2012).”
BackgroundSelf-citationTwo Dimensions of Opacity and the Deep Learning Predicament
Boge, F. J. (2022) · Minds and Machines · doi:10.1007/s11023-021-09569-4 · Cited by 110
“Note that the learning algorithm involved in deriving this model may itself count as yet another model: Buckner (2018) points to the possibility of understanding concept abstraction on the basis of deep convolutional nets, without drawing too close a parallel to either brain processes or most…”
Neurocognitive Mechanisms: Explaining Biological Cognition
Piccinini, G. (2020) · doi:10.1093/oso/9780198866282.001.0001 · Cited by 162
Metadata
- Authors
- Buckner, C.
- Affiliations
- University of Houston (US)
- Corresponding author
- Buckner, Cameron
- ORCID
- Buckner, Cameron
- Venue
- Synthese
- Published
- 2018-09-24
- Volume
- 195
- Issue
- 12
- Pages
- 5339-5372
- ISSN
- 0039-7857 (ISSN-L)
- 1573-0964
- OpenAlex
- W2889853714
- Cited by
- 149
- References
- 98
- Citations by year
- 2026: 5 · 2025: 13 · 2024: 17 · 2023: 28 · 2022: 17 · 2021: 26 · 2020: 33 · 2019: 9 · 2018: 1
- FWCI
- 22.60
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- www.springer.com/tdm (tdm)
- Locations
- Synthese (publishedVersion, cc-by)
- PhilSci-Archive (University of Pittsburgh) (acceptedVersion)
- PhilPapers (PhilPapers Foundation) (submittedVersion)
- Topics
- Social Sciences › Arts and Humanities › Philosophy › Philosophy and Theoretical Science
- Social Sciences › Arts and Humanities › History and Philosophy of Science › Cybernetics and Technology in Society
- Life Sciences › Neuroscience › Cognitive Neuroscience › Embodied and Extended Cognition
- Keywords
- deep convolutional neural networks, empiricism, rationalism, hierarchical processing, perceptual discrimination, connectionism, philosophy of mind, neural network architecture
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.