Octavian.ai

Machine learning and reasoning research

[Octavian.ai @Octavian_ai]

Octavian’s mission is to create software that can reason over and understand information as well as a human can. https://t.co/Jmx0lDCnkq

[Octavian.ai @Octavian_ai]

The story of a year in Octavian: https://t.co/lFBYoJPByg

[Octavian.ai @Octavian_ai]

Excited to share with you the second exercise in our Machine Learning on Graphs course: Graph Convolutional Networks https://t.co/N1nkj6YplW

[Octavian.ai @Octavian_ai]

Language could be an important part of artificial general intelligence

[David Mack @DavidHHMack]

https://t.co/esMpWoz6JX

[Octavian.ai @Octavian_ai]

We've updated our Machine Learning on Graphs course webpage with the upcoming schedule: https://t.co/haPvzOn77Y

[Octavian.ai @Octavian_ai]

The first part of our machine learning on graphs course is ready! A warmup exercise on TensorFlow and graph data: https://t.co/7HkWcdSMwe https://t.co/mOKUCDIcrT

[Octavian.ai @Octavian_ai]

Finally, after 1 month of waiting on @Mailchimp support, our course welcome email is out to you all!
https://t.co/hqxH8P13ki

[Octavian.ai @Octavian_ai]

The pre-requisites to our summer machine learning on graphs course: https://t.co/nUhh4GAAdU

[Octavian.ai @Octavian_ai]

We're hosting a free online course "Machine Learning on Graphs" this summer - sign up for the pilot: https://t.co/haPvzOn77Y

[Octavian.ai @Octavian_ai]

How to use the Learning Rate Finder in TensorFlow by Ashwath Salimath https://t.co/3jF9IYI2DM

[Octavian.ai @Octavian_ai]

Eerily good timing, this is the perfect dataset for item 2: https://t.co/gBhtIylbvJ

[Google DeepMind @GoogleDeepMind]

Today we're releasing a large-scale extendable dataset of mathematical questions, for training (and evaluating the abilities of) neural models that can reason algebraically.

Paper: https://t.co/D8g477gcQ4
Code and data: https://t.co/QvR2WkK7j2

[Octavian.ai @Octavian_ai]

Research ideas: https://t.co/gBhtIylbvJ

[Octavian.ai @Octavian_ai]

@Lancophone Having githubs/notebooks as papers would be awesome!

[MIT CSAIL @MIT_CSAIL]

Happy Pi Day from all of us at MIT!

Let's celebrate with a pizza of radius "z" and depth "a".

Volume = pi * z * z * a

#happypiday

[Octavian.ai @Octavian_ai]

Interested?

[David Mack @DavidHHMack]

I’m working on a TPU implementation of BigGANs, if you’re interested in collaborating I’d love to hear from you. https://t.co/HNqJNa9bdi

[Octavian.ai @Octavian_ai]

A simple explanation of the Inception Score https://t.co/7uqBqHAqgP

[Octavian.ai @Octavian_ai]

@chenoehart Let your love and passion for them shine through

[Andy Brock @ajmooch]

BigGAN has been accepted for oral presentation at ICLR2019! We've uploaded a revision of the paper with an improved architecture, BigGAN-deep: 4x the depth, 50% *fewer* parameters, and even better performance.

https://t.co/JkHOcisihx

[Octavian.ai @Octavian_ai]

Handy overview of different GAN loss functions written as TensorFlow
https://t.co/IEf4GMs35T

[Octavian.ai @Octavian_ai]

A good guide to how TensorFlow uses GPU memory, and how you can save 10x space https://t.co/c64tFs2vqe Thanks to @OpenAI

[Octavian.ai @Octavian_ai]

Pretzels are a good test of how well GANs can integrate non-local information. [SA/Big]GAN has one self-attention layer, and has mixed success: