Advice for AI startups Issue 61: CognitionX Data Science, AI and Machine Learning Briefing

VC money is pouring in but the reality from the AI trenches is different.” Matt Turck of First Mark and Peter Brodsky CEO HyperScience advise AI startups to stick to vertical solutions rather than horizontal in order to position away from the big 4

Check out their slides from the O’Reilly Conference: Building an AI Startup: Realities & Tactics.

Tabitha ‘Until The Bots Take Over’ Goldstaub

Feed your mind over lunch

Get it wrong once it’s your fault, get it wrong twice it must be the data

Dr June Andrews, Data Scientist at Pinterest slides “Replication in Data Science – A Dance Between Data Science & Machine Learning” provides good insight into Iterative Supervised Clustering. Intresting to see that although simplicity can unlock great power she shows the shocking result of how hard it is to replicated data science conclusions.

Ethics question for the day

Data For Good Exchange

Check out the papers from Bloomberg’s yearly conference on applications of data science to problems related to the social good. 2016 theme was “how data science can be applied to solve public interest problems.”

Data visualisation

A Dramatic Tour of Python’s Data Visualisation Landscape

Another quality and humorous instalment of Regress to Impress. Dan Saber, UCLA Data Scientist takes a look at competing Python data visualisation packages by way of a two act play!

And in other data visualisation news…

The new release of ggplot2 (2.20) is set for early November and a pre release version is now available

Business Impact of AI

Is deep learning being commoditised?

Another example of robots taking human jobs

Robot does as Robot sees

The google research team explain the three ways to teach a robot motion skills;
learning directly from experience, learning internal models of physics, and learning with human assistance.

Teaching robots how to open a door, I knew we’d all have jobs available to us in the future!

Pure unadulterated research

Deep Reinforcement Learning for Robotic Manipulation

For those who are hungry for an elaboration on topics discussed in the google research blog article… two of its authors published a paper on it yesterday!

Education, training and advice we rate

Want some advice on building an AI Startup?

As was said in the intro, I was clearly over excited about this blog post…
Check out Matt Turck of First Mark and Peter Brodsky CEO HyperScience slides listing the “5P” for success: Positioning (finding blue ocean), Product, Petabytes (data), Process (social engineering) and People.

Published in