Deep Learning Indaba 2018

by AI Journal

Deep Learning Indaba is a gathering to strengthen Machine Learning in Africa. We still lack diversity in AI and is currently being practiced in only certain regions actively. This is one of many initiatives to bring more diversity to AI. More than 500 students and researchers from Africa and around the world visited this event.

Such initiatives greatly help in fostering local ML communities and making research knowledge more accessible. People are more exposed to cutting edge research and it gives them opportunities to visit conferences like NIPS (travel grant). Here is the open letter to the Indaba Team which highlights the achievement of this initiative.

I truly believe that the benefits of AI can truly be realized only when it's not limited to some corporations/communities but when it's made publicly accessible (democratization). This is the only way to maximize the positive impact of AI.

This event was organized great set of researchers which includes Nando De Freitas (DeepMind), Jeff Dean (Chief of Google Brain), David Silver (DeepMind), Shakir Mohamed(DeepMind), Kyunghyun Cho (NYU) et al. as speakers

Jeff Dean talked greatly about exposing oneself to more diverse research. As per him, it's better to read more papers and gather more wider knowledge and inspiration rather than reading one paper in depth.

David Silver gave a talk on 10 Principles for Deep Reinforcement Learning. Sebastian Ruder has provided a nice summary of each slide in his this thread .

His talk revolved around how our RL agent should scale and this can be applied to common ML as well.

A list of Practicals also have been provided, Here . You can directly run them on Google Cloud which is super convenient.

How to write a great research paper  ?

Nando  de Freitas, Ulrich Paquet, Stephan Gouws, Martin Arjovsky, and  Kyunghyun Cho gave a session on how to How to write a great research  paper. They discussed Simon Peyton Jones’ 7 simple suggestions:

  • Don’t wait: write.
  • Identify your key idea
  • Tell one story.
  • Nail your contributions to the mast.
  • Related work: later.
  • Put your readers first.
  • Listen to your readers.

Here are some photos of the events

Jeff Dean giving his talk
If you liked this, you may also like AI Journal . Don’t forget to subscribe . Stay connected with us on Twitter to stay updated in AI Research. Please support me on Patreon