Innovation


Object Detection Demo – Detectron Drone Flight

Neural Networks are getting better and better. Recently Facebook has published “detectron” – an open source software to detect objects and draw segmenation-masks or keypoint annotations around them. Detectron uses a Mask R-CNN architecture and is trained on the MS COCO dataset. It is quite easy to use – you can just start a docker […]


machines that learn better than people with deep learning

Machines are starting to learn and understand better than humans. Technologist Jeremy Howard shows new developments in the fast-moving field of deep learning. The TED talk below shows both wonderful and terrifying implications.

Today machines are able to learn from massive amounts of data. And we already know that companies are using this to recommend products, find friends in social networks and optimize advertising campaigns.

But there is a lot of other development going on. Companies are inventing self driving cars that can recognize a street sign or a human being and that can make reasonable decisions on events like that.  They are inventing systems that can understand speech and translate them in real time – and they do not write a translating program – instead they let the system learn the languages from massive amounts of data. Companies and researchers are using the deep learning algorithm to solve a variety of problems that couldn’t be solved before.

  • Computers are able to learn.
  • Computers start to get better than humans in many fields.
  • They can listen and see.
  • Computers can read and write.
  • They can understand!

The input is not restricted to machine readable text any more. It can be images, audio and video as well. Machines will be able to beat human performance in learning and understanding very soon … and they will not be limited to a single brain without “RAM extension slots”.

See more TED talks on www.ted.com


see invisible motion and hear silent sounds using videoscope – a motion microscope

Michael Rubinstein has developed the “motion microscope” that can highlight tiny movements in videos taken by normal cameras. You can watch a baby breathe, see the pulse of an actor or see a glass vibrating from sound waves.

In his talk he recreates a conversation by amplifying the movements from sound waves bouncing off a bag of chips … quite impressive:

 


Carol Dweck researches “growth mindset” — the idea that we can grow our brain’s capacity … a great introduction to this field.