Deep Learning: Branching into brains
Deep learning is a subfield of machine learning that focuses on training artificial systems to find useful representations of inputs. Recent advances in deep learning have propelled the once arcane field of artificial neural networks into mainstream technology (LeCun et al., 2015). Deep neural networks now regularly outperform humans on difficult problems like face recognition and games such as Go (He et al., 2015; Silver et al., 2017). Traditional neuroscientists have also taken an interest in deep learning because it seemed initially that there were telling analogies between deep networks and the human brain. Nevertheless, there is a growing impression that the field might be approaching a new ‘wall’ and that deep networks and the brain are intrinsically different.
Chief among these differences is the widely held belief that backpropagation, the learning algorithm at the heart of modern artificial neural networks, is biologically implausible. This issue is so central to current thinking about the relationship between artificial and real brains that it has its own name: the credit assignment problem. The error in the output of a neural network (that is, the difference between the output and the 'correct' answer) can be reported or 'backpropagated' to any connection in the network, no matter where it is, to teach the network how to refine the output. But for a biological brain, neurons only receive information from the neurons they are connected to, making credit assignment a real problem. How does the brain blindly adjust the strength of the connections between neurons that are far removed from the output of the network? In the absence of a solution, we may be forced to conclude that deep learning and brains are incompatible after all.
Now, in eLife, Jordan Guerguiev, Timothy Lillicrap and Blake Richards propose a biologically inspired solution to the credit assignment problem (Guerguiev et al., 2017). Central to their model is the structure of the pyramidal neuron, which is the most prevalent cell type in the cortex (the outer layer of the brain). Pyramidal neurons have been a source of aesthetic pleasure and interesting research questions for neuroscientists for decades. Each neuron is shaped like a tree with a trunk reaching up and dividing into branches near the surface of the brain as if extending toward a source of energy or information. Can it be that, while most cells of the body have relatively simple shapes, evolution has seen to it that cortical neurons are so intricately shaped as to be apparently impractical?
Guerguiev et al. – who are based at the University of Toronto, the Canadian Institute for Advanced Research, and DeepMind – report that this impractical shape has an advantage: the long branched structure means that error signals at one end of the neuron and sensory input at the other end are kept separate from each other. These sources of information can then be brought together at the right moment in order to find the best solution to a problem.
As Guerguiev et al. note, many facts about real neurons and the structure of the cortex turn out to be just right to find optimal solutions to problems. For instance, the bottoms of cortical neurons are located just where they need to be to receive signals about sensory input, while the tops of these neurons are well placed to receive feedback error signals (Cauller, 1995; Larkum, 2013). The key to this design principle seems to be to keep these distinct information streams largely independent. At the same time, ion channels under the control of a host of other nearby neurons process and gate the transfer of information within the neuron.
Taking inspiration from these facts Guerguiev et al. implement a deep network with units that have different compartments, just like real neurons, that can separate sensory input from feedback error signals. These units have all the information they need to know in order to nudge the network toward the desired output. Guerguiev et al. prove formally that this approach is mathematically sound. Moreover, their new, biologically plausible deep network is able to perform well on a task to identify handwritten numbers, and does so by creating what are referred to as hierarchical representations. This phenomenon refers to the increasingly complex nature of the responses of the network's layers, commonly found in more traditional deep learning models, and in the sensory cortices of biological brains.
Doubtless, there will be more twists and turns to this story as more biological details are incorporated into the model. For instance the brain also faces a time-based credit assignment problem (Friedrich et al., 2011; Gütig, 2016). Guerguiev et al. admit that this network does not outperform non-biologically derived deep networks – yet. Nevertheless, the model they present paves the way for future work that links biological networks to machine learning. The hope is that this can be a two-way process, in which insights from the brain can be used to improve artificial intelligence, and insights from artificial intelligence can be used to reveal how the brain operates.
References
-
Layer I of primary sensory neocortex: where top-down converges upon bottom-upBehavioural Brain Research 71:163–170.https://doi.org/10.1016/0166-4328(95)00032-1
-
Spatio-temporal credit assignment in neuronal population learningPLoS Computational Biology 7:e1002092.https://doi.org/10.1371/journal.pcbi.1002092
-
ConferenceDelving deep into rectifiers: Surpassing human-level performance on imagenet classificationProceedings of the IEEE International Conference on Computer Vision. pp. 1026–1034.
Article and author information
Author details
Publication history
Copyright
© 2017, Shai et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 6,320
- views
-
- 740
- downloads
-
- 4
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Further reading
-
- Cancer Biology
- Computational and Systems Biology
Effects from aging in single cells are heterogenous, whereas at the organ- and tissue-levels aging phenotypes tend to appear as stereotypical changes. The mammary epithelium is a bilayer of two major phenotypically and functionally distinct cell lineages: luminal epithelial and myoepithelial cells. Mammary luminal epithelia exhibit substantial stereotypical changes with age that merit attention because these cells are the putative cells-of-origin for breast cancers. We hypothesize that effects from aging that impinge upon maintenance of lineage fidelity increase susceptibility to cancer initiation. We generated and analyzed transcriptomes from primary luminal epithelial and myoepithelial cells from younger <30 (y)ears old and older >55 y women. In addition to age-dependent directional changes in gene expression, we observed increased transcriptional variance with age that contributed to genome-wide loss of lineage fidelity. Age-dependent variant responses were common to both lineages, whereas directional changes were almost exclusively detected in luminal epithelia and involved altered regulation of chromatin and genome organizers such as SATB1. Epithelial expression variance of gap junction protein GJB6 increased with age, and modulation of GJB6 expression in heterochronous co-cultures revealed that it provided a communication conduit from myoepithelial cells that drove directional change in luminal cells. Age-dependent luminal transcriptomes comprised a prominent signal that could be detected in bulk tissue during aging and transition into cancers. A machine learning classifier based on luminal-specific aging distinguished normal from cancer tissue and was highly predictive of breast cancer subtype. We speculate that luminal epithelia are the ultimate site of integration of the variant responses to aging in their surrounding tissue, and that their emergent phenotype both endows cells with the ability to become cancer-cells-of-origin and represents a biosensor that presages cancer susceptibility.
-
- Computational and Systems Biology
Degree distributions in protein-protein interaction (PPI) networks are believed to follow a power law (PL). However, technical and study biases affect the experimental procedures for detecting PPIs. For instance, cancer-associated proteins have received disproportional attention. Moreover, bait proteins in large-scale experiments tend to have many false-positive interaction partners. Studying the degree distributions of thousands of PPI networks of controlled provenance, we address the question if PL distributions in observed PPI networks could be explained by these biases alone. Our findings are supported by mathematical models and extensive simulations, and indicate that study bias and technical bias suffice to produce the observed PL distribution. It is, hence, problematic to derive hypotheses about the topology of the true biological interactome from the PL distributions in observed PPI networks. Our study casts doubt on the use of the PL property of biological networks as a modeling assumption or quality criterion in network biology.