Using goal-driven deep learning models to understand sensory cortex

By Daniel L. K. Yamins, James J. DiCarlo. 2016

1,765 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Neuroscience

How researchers see this paper

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

In Cognitive Neuroscience

In Neuroscience

Citations in the corpus

Lifelong learning of human actions with deep neural network self-organization (2017, background): “…recurrent GWR networks followed by MAX-pooling shape a functional hierarchy that shares lower-level features for composing higher-level representations, without aiming to model the response of neurons in any specific brain area (e.g. Stevens, Law, Antolík, & Bednar, 2013; Yamins & DiCarlo, 2016).”
Critical Learning Periods in Deep Neural Networks (2017, background): “It is also not our goal to suggest that, since they both exhibit critical periods, ANNs are necessarily a valid model of neurobiological information processing, although recent work has emphasized this aspect (15, 54, 55).”
Multiagent cooperation and competition with deep reinforcement learning (2015, background): “Convolutional neural networks have become the much needed high-level computational framework against which to contrast data-driven hypotheses of visual processing in the brain [22, 23].”
Toward an Integration of Deep Learning and Neuroscience (2016, background): “…been done to aid the interpretation of the neural dynamics underlying decision making in the PFC (Sussillo, 2014), working memory in the posterior parietal cortex (Rajan et al., 2016) and object or action representation in the visual system (Tacchetti et al., 2016; Yamins and DiCarlo, 2016a,b).”
The remarkable robustness of surrogate gradient learning for instilling complex function in spiking neural networks (2020, neutral): “…(Maheswaranathan et al., 2018; McClure & Kriegesko-rte, 2016; McIntosh, Maheswaranathan, Nayebi, Ganguli, & Baccus, 2016; Pospisil, Pasupathy, & Bair, 2018; Tanaka et al., 2019; Yamins & DiCarlo, 2016; Yamins et al., 2014) and serve to formulate hypotheses about their mechanistic underpinnings.”
The Rhetoric and Reality of Anthropomorphism in Artificial Intelligence (2019, methodology): “DNNs have been used to model information processing at various stages of the visual cortex of human and nonhuman primates (Cichy et al. 2016; Kriegeskorte 2015; Yamins and DiCarlo 2016), achieving state of the art predictive performance while simultaneously suggesting novel subcortical functions.”
Opening the Black Box of Financial AI with CLEAR-Trade: A CLass-Enhanced Attentive Response Approach for Explaining and Visualizing Deep Learning-Driven Stock Market Prediction (2017, background): “However, deep learning, to large extent, have essentially been unexplainable "black boxes", with no clear explanation as to how they reach particular decisions [6].”
If deep learning is the answer, what is the question? (2020, background): “1a) , whereas in deeper layers, the distributions and similarity structure of neural representations for objects and categories resemble those in the primate ventral stream 12–14,19 ( Fig.”
Reinforcement Learning, Fast and Slow. (2019, neutral): “These developments have attracted growing interest from psychologists, psycholinguists and neuroscientists, curious about whether developments in AI might suggest new hypotheses concerning human cognition and brain function [7, 8, 9, 10, 11].”
A parametric texture model based on deep convolutional features closely matches texture appearance for humans (2017, methodology): “In this paper we synthesise textures using the ﬁrst convolutional layer from each stack after the max pooling. is nearly imperceptible to a human can cause a deep network to misclassify an object with high conﬁdence (Szegedy et al., 2013, see Yamins and DiCarlo (2016) for additional discussion).”
Dendritic cortical microcircuits approximate the backpropagation algorithm (2018, background): “…growing evidence demonstrates that deep neural networks outperform alternative frameworks in accurately reproducing activity patterns observed in the cortex (Lillicrap and Scott, 2013; Yamins et al., 2014; Khaligh-Razavi and Kriegeskorte, 2014; Yamins and DiCarlo, 2016; Kell et al., 2018).”
Deep Learning With Spiking Neurons: Opportunities and Challenges (2018, background): “Recently discovered analogies of real neural representations in the cortex to those learned in deep networks (Kriegeskorte, 2015; Yamins and DiCarlo, 2016) have increased the interest of the neuroscience community in deep learning, and deep SNNs could become an interesting tool to study the…”
Theories of Error Back-Propagation in the Brain (2019, background): “…neural networks learn to perform tasks such as image classi ﬁ cation or navigation, the neurons in their layers develop representations similar to those seen in brain areas involved in these tasks, such as receptive ﬁ elds across the visual hierarchy or grid cells in the entorhinal cortex [5 – 7].”