↳ Chilly

Which GPU(s) to Get for Deep Learning: My Experience and Advice for Using GPUs in Deep Learning

Source: Tim Dettmers Blog …

Deep learning is a field with intense computational requirements and the choice of your GPU will fundamentally determine your deep learning experience. With no GPU this might look like months of waiting for an experiment to finish, or running an experiment for a day or more only to see that the chosen parameters were off.

Read More

↳ Chilly

How cyborg marketers are beating human marketers

Source: Sentient AI

Often times, when you read about artificial intelligence, the boogie man of technological unemployment isn’t far behind. Pessimistic prognosticators have been saying this for quite some time, in fact. But when you think about AI, it’s often more accurate to think of it not as a replacement, but as an aid or some kind of super-charged intern.

Read More

↳ Chilly

Decisions are for making bad outcomes inconsistent

Cheating Death in DamascusNate Soares’ recent decision theory paper with Ben Levinstein, “Cheating Death in Damascus,” prompted some valuable questions and comments from an acquaintance (anonymized here). I’ve put together edited excerpts from the commenter’s email below, with Nate’s responses.

The discussion concerns functional decision theory (FDT), a newly proposed alternative to causal decision theory (CDT) and evidential decision theory (EDT).

Read More

↳ Chilly

Busted: 5 conversion myths

Source: Sentient AI

If you’ve done any conversion rate optimization in the past, you probably know better than to make assumptions. It’s hard to tell what moves the needle. Ideas you’re really excited about often fail and small tweaks can sometimes make a big difference.

It’s also true that every audience is unique.

Read More

↳ Chilly

Artificial Intelligence Weekly – Artificial Intelligence News #58 – Apr 6th 2017

Source: AI Weekly

In the News

The Mobile Internet is over. Baidu goes all in on AI

Baidu spent $2.9M on AI over 2 years, has 1300 AI researchers

bloomberg.com

Inside Facebook's AI class

“I took the AI class Facebookers are literally sprinting to get into”

wired.com

Out of road: driverless vehicles are replacing the trucker

New self-driving technology threatens an American way of life

ft.com

More news…

  • China is making AI a national strategic priority.

Read More

↳ Chilly

Unsupervised Sentiment Neuron

We’ve developed an unsupervised system which learns an excellent representation of sentiment, despite being trained only to predict the next character in the text of Amazon reviews. A linear model using this representation achieves state-of-the-art sentiment analysis accuracy on a small but extensively-studied dataset, the Stanford Sentiment Treebank (we get 91.8% accuracy versus the previous best of 90.2%), and can…

Read More

↳ Chilly

Spam Detection in the Physical World

We’ve created the world’s first Spam-detecting AI trained entirely in simulation and deployed on a physical robot. Our vision system successfully flagging a can of Spam for removal. The vision system is trained entirely in simulation, while the movement policy for grasping and removing the Spam is hard-coded. Our detector is able to avoid other objects, including healthy ones such as…

Read More

↳ Chilly

What does Iris.AI have in common with SpaceX?

Source: Iris AI

Iris.AI gives students, researchers, and engineers alike, the tools to realize and actualize their (spatial) dreams. And in this case, dreams of making aerospace history.

 

Two months after SpaceX’s Dragon launch mission, Iris.AI hosted a science hackathon or Scithon as we like to call it, in partnership with the leading composite materials research institute Swerea SICOMP.

Read More

↳ Chilly

Evolution Strategies as a Scalable Alternative to Reinforcement Learning

We’ve discovered that evolution strategies (ES), an optimization technique that’s been known for decades, rivals the performance of standard reinforcement learning (RL) techniques on modern RL benchmarks (e.g. Atari/MuJoCo), while overcoming many of RL’s inconveniences. In particular, ES is simpler to implement (there is no need for backpropagation), it is easier to scale in a distributed setting, it does not…

Read More