Unsupervised Natural Visual Experience Rapidly Reshapes Size Invariant Object Representation in Inferior Temporal Cortex

By Nuo Li, James J. DiCarlo. 2010

143 citations recorded by the source corpus.

Summary We easily recognize objects and faces across a myriad of retinal images produced by each object. One hypothesis is that this tolerance (a.k.a. “invariance”) is learned by relying on the fact that object identities are temporally stable. While we previously found neuronal evidence supporting this idea at the top of the non-human primate ventral visual stream (inferior temporal cortex, IT), we here test if this is a general tolerance learning mechanism. First, we found that the same type of unsupervised experience that reshaped IT position tolerance also predictably reshaped IT size tolerance, and the magnitude of reshaping was quantitatively similar. Second, this tolerance reshaping can be induced under naturally occurring dynamic visual experience, even without eye movements. Third, unsupervised temporal contiguous experience can build new neuronal tolerance. These results suggest that the ventral visual stream uses a general unsupervised tolerance learning (UTL) algorithm to build its invariant object representation.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition

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 Computer Vision and Pattern Recognition

Citations in the corpus

View-tolerant face recognition and Hebbian learning imply mirror-symmetric neural tuning to head orientation (2016, positive): “There is considerable evidence from physiology and psychophysics that the brain employs a temporal-association strategy of this sort [35, 56, 12, 30, 57, 31].”
How does the brain solve visual object recognition? (2012, positive): “…and sizes can rapidly and strongly reshape the position and size tolerance of IT neurons—destroying existing tolerance and building new tolerance, depending on the provided visual experience statistics (Li and DiCarlo, 2008, 2010), and predictably modifying object perception (Cox et al., 2005).”
Slowness and Sparseness Have Diverging Effects on Complex Cell Learning (2014, background): “Li & DiCarlo [39,40] found neural correlates of the learning of invariances by manipulating the statistics of the presented stimuli.”
Unsupervised learning of invariant representations (2016, methodology): “In that paper we modeled Li and DiCarlo’s ”invariance disruption” experiments in which they showed that a temporal association paradigm can induce individual IT neurons to change their stimulus preferences under specific transformation conditions [54, 58].”
Learning Invariance from Natural Images Inspired by Observations in the Primary Visual Cortex (2012, background): “Such properties are not rigidly encoded in the visual system; they are a product of learning and adaption in the visual system (Cox, Meier, Oertelt, & DiCarlo, 2005; Li & DiCarlo, 2010).”
Unsupervised learning of invariant representations with low sample complexity: the magic of sensory cortex or a new framework for machine learning? (2013, methodology): “In that paper we modeled Li and DiCarlo’s ”invariance disruption” experiments in which they showed that a temporal association paradigm can induce individual IT neurons to change their stimulus preferences under speciﬁc transformation conditions [54, 58].”
iLab-20M: A Large-Scale Controlled Object Dataset to Investigate Deep Learning (2016, background): “In humans and monkeys, this is believed to be exploited by the neural representations [37], though c) d) camera”
Adaptive Resonance Theory: How a brain learns to consciously attend, learn, and recognize a changing world (2013, methodology): “…have used the positional ARTSCAN (pARTSCAN) extension of the ARTSCAN model to simulate neurophysiological data of Li and DiCarlo (2008; see also Li and DiCarlo (2010)) showing that views from different objects can be merged within inferotemporal categories when monkeys are presented with an…”
Learning and disrupting invariance in visual recognition (2011, methodology): “A subsequent physiology experiment by Li and DiCarlo using a similar paradigm showed that individual neurons in primate anterior inferotemporal cortex (AIT) change their selectivity in a position-dependent manner after less than an hour of exposure to the altered visual environment [4].”
Deep Hierarchies in the Primate Visual Cortex: What Can We Learn for Computer Vision? (2013, background): “However, the measurable changes at the single-cell level induced by learning appear to be much smaller at earlier levels in the visual hierarchy such as V1 [154] compared to later stages such as V4 [139] or IT [104].”
Unsupervised Learning of Invariant Representations in Hierarchical Architectures (2013, methodology): “In that paper we modeled Li and DiCarlo’s ”invariance disruption” experiments in which they showed that a temporal association paradigm can induce individual IT neurons to change their stimulus preferences under specific transformation conditions [60, 30].”
Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognition (2014, background): “IT cortex have been observed [52–54], even during passive viewing [55].”