<div dir="ltr"><div class="gmail_default" style="font-size:large"><p class="gmail-m_6635963219871096802gmail-p1" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px">Announcing a 3-year postdoctoral research position in Denis Pelli's lab at NYU in Psychology, affiliated with Center for Neural Science, supported by NIH grant to Denis Pelli (PI), Jon Winawer, & Yann LeCun, in computational modeling of human object recognition in clutter. Crowding is a computational bottleneck in human object recognition. Is crowding just a nuisance, or is it the key to understanding how an object is recognized by a limited number of neurons? A trained ConvNet mimics many aspects of human letter recognition within a single crowding unit (<a href="http://f1000.com/posters/browse/summary/1095738" target="_blank"><span class="gmail-m_6635963219871096802gmail-s1" style="color:rgb(220,161,13)">Ziskind et al. 2014</span></a>). However, some things cannot be recognized by people unless they extend across several crowding units, and human learning generalizes across scale and other dimensions. We’d like to extend our modeling of object recognition to explain that. Ongoing results on efficiency (contrast thresholds in noise) and the need for several crowding units provide modeling clues (<a href="https://jov.arvojournals.org/article.aspx?articleid=2699020&resultClick=1" target="_blank"><span class="gmail-m_6635963219871096802gmail-s1" style="color:rgb(220,161,13)">Pelli 2018</span></a>). fMRI collaboration with Winawer’s lab is documenting conservation of crowding distance in area V4 (<a href="https://jov.arvojournals.org/article.aspx?articleid=2699845&resultClick=1" target="_blank"><span class="gmail-m_6635963219871096802gmail-s1" style="color:rgb(220,161,13)">Zhou et al. 2018</span></a>).<span class="gmail-m_6635963219871096802gmail-Apple-converted-space"> </span></p><p class="gmail-m_6635963219871096802gmail-p2" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px;min-height:14px"><br></p><p class="gmail-m_6635963219871096802gmail-p1" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px">Pelli’s lab is in the Psychology Dept., on the same elevator as the Center for Neural Science and our 3T magnet, two blocks from Courant and Facebook. Salary will be $48,432 or higher, based on experience.</p><p class="gmail-m_6635963219871096802gmail-p1" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px">For informal inquiries, email <span class="gmail-m_6635963219871096802gmail-s2" style="text-decoration-line:underline"><a href="mailto:denis.pelli@nyu.edu" target="_blank">denis.pelli@nyu.edu</a></span></p><p class="gmail-m_6635963219871096802gmail-p3" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px;color:rgb(220,161,13)"><span class="gmail-m_6635963219871096802gmail-s3" style="color:rgb(0,0,0)">To apply, please visit: <a href="https://apply.interfolio.com/53970" target="_blank"><span class="gmail-m_6635963219871096802gmail-s2">https://apply.interfolio.com/53970</span></a></span></p><p class="gmail-m_6635963219871096802gmail-p2" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px;min-height:14px"><br></p><p class="gmail-m_6635963219871096802gmail-p1" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px">Ziskind, A.J., Hénaff, O., LeCun, Y., & Pelli, D.G. (2014) The bottleneck in human letter recognition: A computational model. Vision Sciences Society, St. Pete Beach, Florida, May 16-21, 2014, 56.583. <a href="http://f1000.com/posters/browse/summary/1095738" target="_blank"><span class="gmail-m_6635963219871096802gmail-s1" style="color:rgb(220,161,13)">http://f1000.com/posters/browse/summary/1095738</span></a></p><p class="gmail-m_6635963219871096802gmail-p2" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px;min-height:14px"><br></p><p class="gmail-m_6635963219871096802gmail-p1" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px">D. Pelli (2018) Despite a 100-fold drop in cortical magnification, a fixed-size letter is recognized equally well at eccentricities of 0 to 20 deg. How can this be?. <i>Journal of Vision</i> 2018;18(10):26. <a href="https://jov.arvojournals.org/article.aspx?articleid=2699020&resultClick=1" target="_blank"><span class="gmail-m_6635963219871096802gmail-s1" style="color:rgb(220,161,13)">https://jov.arvojournals.org/article.aspx?articleid=2699020&resultClick=1</span></a></p><p class="gmail-m_6635963219871096802gmail-p2" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px;min-height:14px"><br></p><p class="gmail-m_6635963219871096802gmail-p1" style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:12px;line-height:normal;font-family:"Helvetica Neue";margin:0px">J. Zhou, N. Benson, J. Winawer, D. Pelli (2018) Conservation of crowding distance in human V4. <i>Journal of Vision</i>2018;18(10):856. <a href="https://jov.arvojournals.org/article.aspx?articleid=2699845&resultClick=1" target="_blank"><span class="gmail-m_6635963219871096802gmail-s1" style="color:rgb(220,161,13)">https://jov.arvojournals.org/article.aspx?articleid=2699845&resultClick=1</span></a><span class="gmail-m_6635963219871096802gmail-Apple-converted-space"> </span></p></div></div>