#Bayesian Optimization framework can intelligently trade off experiments w/ varying cost & fidelity. We achieve strong regret bounds as well as state-of-the-art performance on multiple real-world #datasets! Preprint: https://arxiv.org/abs/1811.00755v1


#machinelearning

✴️ @AI_Python_EN
🗣 @AI_Python_arXiv
❇️ @AI_Python

The 10 Biggest datasets of 2018

0) Open Images V4 from Google AI on April 30th Contains 15.4M bounding-boxes for 600 categories on 1.9M images.
Paper: https://lnkd.in/fm4xiUm

1) MURA from Stanford University ML Group on May 24 Radiographic image dataset
Paper: https://lnkd.in/fBy5szB

2) BDD100K from BAIR, Georgia Tech, Peking University, Uber AI
on May 30 Self-Driving Car Dataset.
Paper: https://lnkd.in/f-sYj9k

3) SQuAD 2.0 from Stanford
on June 11 QA Dataset.
Paper: https://lnkd.in/fYc6c5W

4) CoQA from Stanford on August 21 QA Dataset
Paper: https://lnkd.in/fKvuTvE

5) Spider 1.0 from Yale Univ on September 24 Cross-domain semantic parsing and text-to-SQL dataset.
Paper: https://lnkd.in/fWyR2x8

6) HototQA from Carnegie, Stanford, and Montreal on September 25 QA Dataset on Wiki
Paper: https://lnkd.in/fTtTgZt

7) Tencent ML Images from Tencent AI Lab on Oct 18 largest open-source multi-label image dataset
Paper: https://lnkd.in/ffV6VD5

8) Tencent AI Lab Embedding Corpus for Chinese words and phrases on Oct 19 Embeddings Dataset
Paper: https://lnkd.in/ffV6VD5

9) fastMRI from NYU and Facebook AI on November 26
Knee MRI Images Dataset
Paper: https://lnkd.in/fQuUDNk

Read: https://lnkd.in/fXU9Kr6

#dataset #datasets

✴️ @AI_Python_EN
🗣 @AI_Python_Arxiv

Shuffling large datasets, have you ever tried that?

Here the author presents an algorithm for shuffling large datasets.
Here you learn the following;

0. why Shuffle in the first place?
1. A 2-pass shuffle algorithm is tested
2. How to deal with oversized piles
3. Parallelization & more

Link to article : https://lnkd.in/dZ8-tyJ
Gist on #Github: for a cool visualization of the shuffle https://lnkd.in/d8iK8fd

#algorithms #github #datasets #deeplearning #machinelearning

❇️ @AI_Python
🗣 @AI_Python_Arxiv
✴️ @AI_Python_EN

Deep Convolutional Sum-Product Networks for Probabilistic Image Representations

Sum-Product Networks (SPNs) are hierarchical probabilistic graphical models capable of fast and exact inference.

Applications of SPNs to real-world data such as large image datasets has been fairly limited in previous literature. Here is a Convolutional Sum-Product Networks (ConvSPNs) which exploit the inherent structure of images in a way similar to deep convolutional neural networks, optionally with weight sharing.
#neuralnetworks #datasets #deeplearning

Paper: https://lnkd.in/ei4Gqjy

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv