The recent excitement about neural networks

By F. Crick. 1989

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Large Language Models

Citations in the corpus

What can the brain teach us about building artificial intelligence? (2016, background): “For example, most neural networks use some form of gradient-based (e.g., backpropagation) or Hebbian learning. It has long been argued, however, that backpropagation is not biologically plausible; as Crick (1989) famously pointed out, backpropagation seems to require that information be transmitted ”
Toward an Integration of Deep Learning and Neuroscience (2016, background): “…may also use a host of proposed circuit structures that would allow it to actually perform, in eﬀect, backpropagation of errors through a multi-layer network, using biologically realistic mechanisms—a feat that had once been widely believed to be biologically implausible (Crick, 1989; Stork, 1989).”
Conjunctive representations in learning and memory: principles of cortical and hippocampal function. (2001, background): “However, backpropagation has been widely challenged on the grounds that it lacks a plausible biological mechanism (e.g., Crick, 1989; Zipser & Andersen, 1988).”
A Neural Network Model of Ventriloquism Effect and Aftereffect (2012, unknown): “and descendent gradient algorithm; indeed, recurrent synapses are ubiquitous in the cortex, and Hebbian forms of learning are known to be present in the brain [52], whereas neural implementation of gradient descendent learning is controversial [53].”
An Approximation of the Error Backpropagation Algorithm in a Predictive Coding Network with Local Hebbian Synaptic Plasticity (2017, background): “This led to a common view of the biological implausibility of this algorithm (Crick, 1989)—for example: “Despite the apparent simplicity and elegance of the back-propagation learning rule, it seems quite implausible that something like equations […] are computed in the cortex” (O’Reilly & Munakata,…”
The Division of Labor Between the Neocortex and Hippocampus (2010, methodology): “Although the backpropagation mechanism has been widely challenged as biologically implausible (e.g., Crick, 1989; Zipser & Andersen, 1988), a recent analysis shows that simple biologically-based mechanisms can be used to implement this mechanism (O'Reilly, 1996), so that it is quite reasonable to…”
The remarkable robustness of surrogate gradient learning for instilling complex function in spiking neural networks (2020, methodology): “Although considered biologically implausible (Crick, 1989), we have limited our study to training SNNs with backpropagation, the de facto standard for computing gradients in systems involving recurrence and hidden neurons.”
Assessing the Scalability of Biologically-Motivated Deep Learning Algorithms and Architectures (2018, background): “Historically, there has been significant disagreement about whether BP can tell us anything interesting about learning in the brain [8, 11].”
Synaptic Plasticity Forms and Functions. (2020, background): “…(Rumelhart et al. 1986,LeCun et al.2015),the commonly used back-propagation of error algorithm requires knowledge of downstream weights, the derivatives of output nonlinearities, and other aspects that have led authors to question its direct biological plausibility (Crick 1989, Bengio et al. 2016).”
Equivalence of Backpropagation and Contrastive Hebbian Learning in a Layered Network (2003, methodology): “This work prompted the search for a learning algorithm, which has similar functionality as backpropagation (Crick, 1989; Mazzoni, Andersen, & Jordan, 1991) and at the same time is biologically plausible.”
Dendritic cortical microcircuits approximate the backpropagation algorithm (2018, background): “Interestingly, the workhorse of deep learning is still the classical backpropagation of errors algorithm (backprop; Rumelhart et al., 1986), which has been long dismissed in neuro-science on the grounds of biologically implausibility (Grossberg, 1987; Crick, 1989).”
The neural basis of human error processing: reinforcement learning, dopamine, and the error-related negativity. (2002, background): “In monkeys, the error potentials discovered by Brooks and colleagues in the anterior cingulate sulcus occurred only during the earliest stages of the task as the monkeys learned what was required of them (Brooks, 1986, 1989; Gemba et al., 1986).”
SuperSpike: Supervised Learning in Multilayer Spiking Neural Networks (2017, methodology): “Finally, credit assignment in hidden layers is problematic for two reasons: (i) it is technically challenging because efficient auto-differentiation tools are not available for most event-based spiking neural network frameworks, and (ii) the method of weight updates implemented by the standard backp”