Maarten C Stol

[KDnuggets @kdnuggets]

Google AI chief @JeffDean interview: #Machinelearning trends in 2020
- New AI hardware
- #DeepLearning systems should consider their energy consumption to help combat #climatechange
- expect advances in multitask learning and multimodal learning,
https://t.co/Nv2qGYP46A

[Maarten C Stol @MaartenCStol]

https://t.co/cCWgbCU1Rb

[Maarten C Stol @MaartenCStol]

https://t.co/XA3kgceL7T

[Maarten C Stol @MaartenCStol]

https://t.co/NCgTOnSRKx

[Maarten C Stol @MaartenCStol]

'Europees lab nodig om controle te houden over kunstmatige intelligentie' - https://t.co/FZeGAsnprZ

[Maarten C Stol @MaartenCStol]

https://t.co/HicFYfOj7s

[ODSC (Open Data Science Conference) AI @_odsc]

GDPR: What data scientists and CIOs need to know https://t.co/uCw3dLn1DF #ODSC #DataScience #GDPR

[Karan Goel @krandiash]

Really cool work from the Brunskill lab at Stanford on using model-based RL! https://t.co/s7bbugQq01

[gio @GiorgioPatrini]

Hypothetical: we will discover that there is no practical way to defend neural networks against adversarial examples. What next? https://t.co/MKKCIyijmG

[gio @GiorgioPatrini]

Excited by this direction of formal investigation for adversarial defences: Adversarial examples from computational constraints, Bubeck et al https://t.co/FKUqwuTyE7

[Nando de Freitas @NandoDF]

This is cool. Neural network turns pitch black images into bright colorful photos https://t.co/jzW9d0cKMS

[Max Tegmark @tegmark]

The main risk with advanced #AI isn't malice, but competence: that it accomplishes goals that aren't aligned with ours. Here are some hilarious and cautionary examples of how this can happen accidentally:
https://t.co/5lY5gRA7ND

[Maarten C Stol @MaartenCStol]

https://t.co/g57HIsxI3n

[Maarten C Stol @MaartenCStol]

https://t.co/0VCUZipsZf

[François Chollet @fchollet]

0 is not (necessarily) special -- not all charts need a y axis that starts at 0. Likewise, a ML model with 80% validation accuracy is not "pretty accurate" -- you shouldn't compare to 0%, you should compare to a common-sense baseline (that baseline may be arbitrarily high!) https://t.co/L3CQr2SlR7

[Adam J Calhoun @neuroecology]

New from the lab!

Fast and easy pose tracking and kinematics using deep networks, by @talmop @shaevitz @SamWangPhD https://t.co/KAWimUaW5U

[KDnuggets @kdnuggets]

Top 7 #DataScience Use Cases in #Finance https://t.co/PAe4t7ajAP

[Sebastian Ruder @seb_ruder]

This is a super useful paper that we need more of: Better ImageNet models are not necessarily better feature extractors (ResNet is best); but for fine-tuning, ImageNet performance is strongly correlated with downstream performance. https://t.co/MrkX4yYgHn

[Ferenc Huszár @fhuszar]

Judea Pearl claims all we do in ML is curve fitting. I wrote this post to explain that claim and introduce the basics of causal inference to ML folks.

Machine Learning beyond Curve Fitting: An Intro to Causal Inference and do-Calculus
https://t.co/1osm0VcaaR

[Jeff Dean @JeffDean]

How can you train a Google Clips camera to know when the time is right to capture a special photographic moment? Nice post describing the work that went on behind the scenes to develop a machine learning model that runs entirely on device to do this.

https://t.co/TGLzYQfC9W