There is another Data Science Breakfast at Paris upcoming on this Saturday at 10:00 AM at Strada Cafe, 24 Rue Monge.

Feel free to join, regardless of your level. Everyone is welcome.

Discovering and Classifying In-app Message Intent at Airbnb

An article on how #AirBnB finds intent in conversations to make its service better.

Link: https://medium.com/airbnb-engineering/discovering-and-classifying-in-app-message-intent-at-airbnb-6a55f5400a0c

#NLP #intent

Natural Questions

Google published Natural Questions, a new, large-scale corpus and challenge for training and evaluating open-domain question answering systems, and the first to replicate the end-to-end process in which people find answers to questions.

Link: https://ai.googleblog.com/2019/01/natural-questions-new-corpus-and.html

#Google #NLP #data

Zero-shot transfer across 93 languages: Facebook publishes enhanced LASER library

Link: https://code.fb.com/ai-research/laser-multilingual-sentence-embeddings/

#Facebook #NLP

Dynamic Transfer Learning for Named Entity Recognition

Paper with direct healthcare application by #Amazon interesting proposal to use Dynamic Transfer Network for architecture search

Link: https://arxiv.org/pdf/1812.05288.pdf

#healthcare #NLP

Use Machine Learning to filter messages in the browser

Article on filtering messages during #Twitch streaming with #tensorflow.

Link: https://dev.to/embiem/use-machine-learning-to-filter-messages-in-the-browser-4i19

#DL #NN #NLP #tensorflowjs

VeGAN release

A small library to easily train various existing #GAN using #PyTorch. Based on provided generator and discriminator, VeGANs trains them with a selected GAN algorithm.

Link: https://github.com/unit8co/vegans

StanfordNLP release.

New Python #NLProc package. StanfordNLP provides native, neural (PyTorch) tokenization, POS tagging and dependency parsing for 53 languages based on UD v2 — and a Python CoreNLP interface.

Pip: pip install stanfordnlp
PyPI: https://pypi.org/project/stanfordnlp/
Link: https://stanfordnlp.github.io/stanfordnlp/

#NLP #Stanford

Typical data scientist/machine learning engineer:
> understands that data science is generally engineering most of programmers can deal with
> do want to be overpaid and thus to keep others away from field
> keeps saying rare useless skill he possesses is definitively necessary for data science

Only truth, whole truth and nothing but the truth.

Towards reconstructing intelligible speech from the human auditory cortex



Columbia neuroengineers have created a system that translates thought into intelligible, recognizable speech. By monitoring someone's brain activity, the technology can reconstruct the words a person hears with unprecedented clarity.

Link: https://www.nature.com/articles/s41598-018-37359-z

#BCI #thought2text

​​#vizualization #statistics

​​StNet: Local and Global Spatial-Temporal Modeling for Action Recognition

Baidu in one of #AAAI19 papers, proposed StNet, a novel framework for both local and global spatial-temporal modeling in videos. StNet outperforms several state-of-the-art approaches in action recognition and balances btw accuracy and model complexity.

ArXiV: https://arxiv.org/pdf/1811.01549.pdf

#SOTA #Baidu

​​DeepSlide: A Sliding Window Framework for Classification of High Resolution Microscopy Images (Whole-Slide Images)

DeepSlide, our #DeepLearning library for classification/visualization of high-resolution pathology images is open-source and available on GitHub.

GitHub: https://github.com/BMIRDS/deepslide

#histopathology #healthcare #PyTorch

​​Introducing AresDB: Uber’s GPU-Powered Open Source, Real-time Analytics Engine

Link: https://eng.uber.com/aresdb/

#Uber #analytics #opensource

​​Fair Regression for Health Care Spending

What happens, if fairness built into the objective function for continuous outcomes & see large improvements in group undercompensation?

This is the most interesting & potentially impactful analysis of fairness in #ML for #healthcare, which can lead to significant improvement in the life of millions.

ArXiV: https://arxiv.org/abs/1901.10566
GitHub: https://github.com/zinka88/Fair-Regression

#statistics #regression

Using Nucleus and TensorFlow for DNA Sequencing Error Correction

This is a new #tutorial of the Genomics team in #GoogleBrain. Good place to start with #deeplearning for #genomics using Nucleus and #Tensorflow.

Link: https://medium.com/tensorflow/using-nucleus-and-tensorflow-for-dna-sequencing-error-correction-47f3f7fc1a50
Google colab: https://colab.research.google.com/github/google/nucleus/blob/master/nucleus/examples/dna_sequencing_error_correction.ipynb

#Google #healthcare

​​Bayesian Statistics explained to Beginners in Simple English

Now some #entrylevel material, which still might be useful to review, because repetitio est mater studiorum.

Link: https://www.analyticsvidhya.com/blog/2016/06/bayesian-statistics-beginners-simple-english/

State-of-the-art (SOTA) collection of Paperswithcode

A great site, worth spreading word about: Papers With Code now includes 950+ ML tasks, 500+ evaluation tables (including SOTA results) and 8500+ papers with code. Probably the largest collection of NLP tasks, including 140+ tasks and 100 datasets.

Link: https://paperswithcode.com
Sota link: https://paperswithcode.com/sota

#Meta #collection #sota #useful

Manifold: A Model-Agnostic Visual Debugging Tool for Machine Learning at Uber

Seesm like there is no week without any news from #Uber engineering team. This time Uber built Manifold, a model-agnostic visualization tool for #ML performance diagnosis and model debugging, to facilitate a more informed and actionable model iteration process.

Link: https://ubere.ng/2Hac0O8

#Pipeline #administration

​​Implementation of character based convolutional neural network

A #PyTorch implementation of Character Based ConvNets for text classification published by Yan LeCun in 2015 is open-sourced on. Many training features and hacks are implemented.

Link: https://github.com/ahmedbesbes/character-based-cnn