What this 2 minute video seems to show is that neural networks need to see more like humans to detect shapes rather than just textures. Our visual cortex detects lines, edges, and other regions of high contrast in several layers. But we see shapes by moving the fovea along a path connecting these features. That step is missing in the neural networks. Instead it uses the less computationally efficient method of applying the same visual resolution across the entire image
On Sun, Mar 10, 2019, 10:36 AM Jim Bromer <[email protected]> wrote: > This was really interesting. > Jim Bromer > > > On Sat, Mar 9, 2019 at 10:09 AM <[email protected]> wrote: > >> Do Neural Networks Need To Think Like Humans?: >> >> >> https://www.youtube.com/watch?v=YFL-MI5xzgg >> > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + delivery > options <https://agi.topicbox.com/groups/agi/subscription> Permalink > <https://agi.topicbox.com/groups/agi/T4aa81fd9912dbd39-Mad1377e5bc9814ea4c3c2aaa> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T4aa81fd9912dbd39-M01a11f01005e08f35f7ea6a3 Delivery options: https://agi.topicbox.com/groups/agi/subscription
