G
Gab41
gab41.lab41.org · Company blog · on HN since 2016
Gab41 is the blog of Lab41, a lab that writes about machine learning, deep learning, speech recognition and data tooling.
What HN saysHN commenters link its posts as references, such as feature engineering being easier than expected and committing to GitHub from Jupyter notebooks.
15HN people shared it
3linked it in comments
52stories submitted
181story points
5 years agolast shared
2016people per year2026
Most-shared posts
- 1The 10 Algorithms Machine Learning Engineers Need to Know3 people · 16 points · first shared 2016 · discussion on HN
- 2850k Images in 24 Hours: Automating Deep Learning Dataset Creation3 people · 11 points · first shared 2016 · discussion on HN
- 3SqueezeNet: How Convolutional Neural Network Design Can Reduce Parameter Size2 people · 8 points · first shared 2016 · discussion on HN
- 4Speech Recognition: You Down with CTC?2 people · 3 points · first shared 2016 · discussion on HN
- 5Flexible Image Tagging with Fast0Tag2 people · 2 points · first shared 2016 · discussion on HN
- 6Feature engineering is just easier2 people · 1 points · first shared 2016 · discussion on HN
- 7Artificial Intelligence Needs a Bullshit Meter1 people · 53 points · first shared 2016 · 13 comments on HN
- 8Machine-Learning for Network Security1 people · 12 points · first shared 2016 · discussion on HN
- 9PySEAL: Homomorphic encryption in a user-friendly Python package1 people · 8 points · first shared 2018 · 1 comments on HN
- 10Deep Residual Learning for Image Recognition1 people · 7 points · first shared 2016 · discussion on HN
- 11Deep Learning Reading Group: Skip-Thought Vectors1 people · 6 points · first shared 2016 · discussion on HN
- 12AI Week in Review – Sept 16 Edition1 people · 5 points · first shared 2016 · discussion on HN
- 13Tensorflow 3 Ways1 people · 4 points · first shared 2016 · discussion on HN
- 14Swapout: Learning an Ensemble of Deep Architectures1 people · 4 points · first shared 2016 · discussion on HN
- 15Using machine learning to classify devices on your network1 people · 3 points · first shared 2018 · discussion on HN
What HN people wrote when they linked it
1cvmask · 2021-12-20
Here is a post on why feature engineering is easier in fact:
https://gab41.lab41.org/feature-engineering-is-just-easier-1928d935ed1…
carreau · 2017-04-04
Well you can commit from the UI: https://gab41.lab41.org/commit-and-push-to-github-from-jupyter-noteboo… you just need to install the extension.
deepnet · 2017-07-10
Geoff Hinton refers to thought vectors performing reasoning by analogy using algebra [1] in his Royal Society Lecture.
The other widely reported vector algebras in a semantic space were discovered by Mikolov et al when producing ~300 dimensional vectors for a billion word Wikipedia corpus.
If one performs vector algebra and ~= is near by cosine distance then using Mikolov's Vectors[3].
King - Man + Woman ~= Queen
France - Paris + Gernmany ~= Berlin
Surprisingly this works for other modalities, Chintala, Radford & Metz found a latent semantic space in images, that adds vectors for glasses or smiles to peoples faces. [4] With a generative model new images can be created as outlined in this blog post by Soumith [5]
Karpathy shows trained nets can be assembled like lego across modalities, slice off the classifier to reveal the rich semantic 'thought vector' layer of an Imagenet trained Alexnet, plug in a RNN sentence generator using word2vec and ( some over simplification ... ) you get a convincing image captioner [6].
The thought vectors are akin to high level representations of the world and can cross modalities . Text to Images using thought Vectors ( from hnnews discussion [7] )
So the vectors of though are in some way a an AI mentalese or encoding of a symbolic representation of the world derived from the data and can ( again drastic over simplification ) transfer modalities and even between previously unlinked languages [8]
Also see Anything2Vec…