In [3]:
HTML("""
<h1 class="title">Decoding of feature selectivity in neural activity</h1>
<h1>Concrete applications in visual data</h1>
<h2>Laurent U Perrinet, INT - Wahiba Taouali, INMED</h2>
<img src="figures/troislogos.png" width=61%/>
""")
Out[3]:

Decoding of feature selectivity in neural activity

Concrete applications in visual data

Laurent U Perrinet, INT - Wahiba Taouali, INMED

  • Course on Computational Neuroscience, Marseille, December 8th, 2015

Acknowledgements:

  • PhD program: Anna Montagnini, Frédéric Chavane, Nadia Pittet, INT, Marseille
  • Material: Peggy Series, Christopher Olah
  • NeuroPhysiology: Giacomo Benvenuti, Frédéric Chavane, INT, Marseille

http://invibe.net/LaurentPerrinet/Presentations/2015-12-08_cours-NeuroComp

In [4]:
HTML("""
    <!-- <h2>Problem statement</h2> -->
    <video controls src="{1}" poster="{0}" width=100%/> 
    """.format(figpath + 'scientists.jpg', figpath + 'ComplexDirSelCortCell250_title.webm')
)
Out[4]:

Bayes : travelling back to the feature space

(see http://colah.github.io/posts/2015-09-Visual-Information/)

In [5]:
Image('figures/prob-1D-rain.png', height=height)
Out[5]:
In [6]:
Image('figures/prob-1D-coat.png', height=height)
Out[6]:
In [7]:
Image('figures/prob-2D-independent-rain.png', height=height)
Out[7]:
In [8]:
Image('figures/prob-2D-dependant-rain-squish.png', height=height)
Out[8]:
In [9]:
Image('figures/prob-2D-factored-rain-arrow.png', height=height)
Out[9]:
In [10]:
Image('figures/prob-2D-factored1-clothing-B.png', height=height)
Out[10]:

examples of Bayesian mechanism in perception

In [11]:
Image('figures/jov-2-6-6-fig002.jpeg', height=height)
# http://jov.arvojournals.org/article.aspx?articleid=2121565
Out[11]:
In [12]:
Image('figures/dotsconvex.jpeg', height=height)
Out[12]:
In [13]:
Image('figures/dotsconcave.jpeg', height=height)
Out[13]:
In [14]:
Image('figures/hollow_mask.jpg', height=height)
Out[14]:

summary

In [15]:
Image('figures/encoding_problem.png', height=height)
Out[15]: