A Lyapunov-based Approach to Safe Reinforcement Learning

Research that develops RL algorithms based on the concept of Lyapunov functions. This work represents a step toward applying RL to real-world problems.

ArXiV: https://arxiv.org/abs/1805.07708

#RL #Lyapunov #facebook

Innovative AI Breath Analyzer Diagnoses Diseases by “Smell”
http://bit.ly/2ROBaSy

AI apparel eye. Deep learning based approach to recognize attributes of apparel.
http://bit.ly/2RNDkll

A Must-Read NLP Tutorial on Neural Machine Translation – The Technique Powering Google Translate
http://bit.ly/2WMfOZL

Recipes for Popular Cocktails
https://tabsoft.co/2ROBdhc

AI is a Big Fat Lie
http://bit.ly/2RNf2rE

Text-based game to feel yourself in not so distant future where real AI exists.

We all now, how DS and #AI products affect our life and how they actually influence our lifestyle. But in the future our relationship will become more tense. It is rather nice experience to play this game, not only because it reminds of old-school games, but also because it looks shookingly possible.

Link: https://www.theverge.com/2019/1/31/18140796/wake-word-algorithm-text-game-ai-artificial-intelligence

#interactive #Meta

#medicine_hitecher

The CRISPR/Cas system is based on the bacterial immune system that protects bacteria from viruses. Scientists have managed to improve and supplement the natural defense mechanism with a toolkit that provides them with a quick and inexpensive way of correcting DNA in plants, animals and humans. These changes occur only in specific parts of the genome without affecting other areas.

Read on Hitecher:

▫️ CRISPR development history
▫️ How CRISPR can be applied in agriculture
▫️ How human DNA works

Boom! What to read on the weekend, Issue #49

Mayo Clinic research uses artificial intelligence to develop inexpensive, widely available early detector of silent heart disease
https://mayocl.in/2tdQO0b

How one brewery tapped into big data to improve its ale
http://bit.ly/2tdqzXu

Jupyter Notebook: An Introduction
http://bit.ly/2teDecH

How Many Kids We Have and When We Have Them
http://bit.ly/2UPLyvc

Data Science and the Art of Persuasion
http://bit.ly/2UPLBHo

Virtual creators aren’t AI — but AI is coming for them
http://bit.ly/2tdqB1y

History of the Periodic Table
https://tabsoft.co/2UPLBqS

Build a Recurrent Neural Network from Scratch in Python – An Essential Read for Data Scientists
http://bit.ly/2UPqDZk

Your AI skills are worth less than you think
http://bit.ly/2thE0FL

Making face recognition less biased doesn’t make it less scary
http://bit.ly/2tbzyIV

Boom! Weekly selection from DSB. Issue #40
The TOP 10 most popular links for the week

1. Monetizing the Math — are you ready?
http://bit.ly/2G9HNxB

2. Papers with code: Browse state-of-the-art
http://bit.ly/2DbD4aS

3. AI is a Big Fat Lie
http://bit.ly/2RNf2rE

4. A Self-Learning, Modern Computer Science Curriculum
http://bit.ly/2WMkPSb

5. The state of the octoverse: machine learning
http://bit.ly/2DbD0Ia

6. Busted! 11 Myths Data Science Transitioners Need to Avoid
http://bit.ly/2RMskVw

7. Data Portals. A Comprehensive List of Open Data Portals from Around the World
http://bit.ly/2G9HQth

8. Elements of Artificial Intelligence free online course
http://bit.ly/2t5FNha

9. A Must-Read NLP Tutorial on Neural Machine Translation — The Technique Powering Google Translate
http://bit.ly/2WMfOZL

10. Your AI skills are worth less than you think
http://bit.ly/2thE0FL

Busted! 11 Myths Data Science Transitioners Need to Avoid
http://bit.ly/2RMskVw

Common license types for datasets
http://bit.ly/2WORh6D

A Self-Learning, Modern Computer Science Curriculum
http://bit.ly/2WMkPSb

Which European cities have gotten warmer? (Spoiler: All of them.)
http://bit.ly/2WMkPBF

Foundations Built for a General Theory of Neural Networks
http://bit.ly/2RKxdhE

Boom! Weekly selection from DSB. Issue #39
The TOP 10 most popular links for the week

1. Top 15 Best Machine Learning Courses for Free in 2019
http://bit.ly/2SjA2Lj

2. How to solve 90% of NLP problems: a step-by-step guide
http://bit.ly/2Rs8COw

3. Practical Deep Learning for Coders 2019
http://bit.ly/2Rq2lD7

4. English soccer team hires predictive AI to coach a game. Yes, really.
http://bit.ly/2SgsmJI

5. Snapshot: data scientist salaries and jobs in Europe
http://bit.ly/2Rw0wEQ

6. Film genre popularity 1910-2018, vizualization by tableau
https://tabsoft.co/2HAPvSI

7. End To End Guide For Machine Learning Projects
http://bit.ly/2ScqBwN

8. Year in Review by The World Bank: 2018 in 14 Charts
http://bit.ly/2WnWMc8

9. Deep learning for NLP Crash Course
http://bit.ly/2WgPfvw

10. We summarized 14 NLP research breakthroughs you can apply to your business
http://bit.ly/2SbmS2

The state of the octoverse: machine learning
http://bit.ly/2DbD0Ia

Data Portals. A Comprehensive List of Open Data Portals from Around the World
http://bit.ly/2G9HQth

Papers with code: Browse state-of-the-art
http://bit.ly/2DbD4aS

Oscar Watch 2019
https://tabsoft.co/2G9HPFJ

Monetizing the Math - are you ready?
http://bit.ly/2G9HNxB

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

U.S., China take the lead in race for artificial intelligence
https://reut.rs/2HUrRAV

Data For Everyone. Here are some of our favorite open datasets created on the Figure Eight platform.
http://bit.ly/2t56eDH

Elements of Artificial Intelligence free online course
http://bit.ly/2t5FNha

The Lifespan of News Stories. How the news enters (and exits) the public consciousness
http://bit.ly/2HTdmNQ

A step closer to self-aware machines—engineers create a robot that can imagine itself
http://bit.ly/2HQJscQ

EE-559 – DEEP LEARNING (SPRING 2019)

Deep learning course covering the main deep learning tools and theoretical results, with examples in the #PyTorch framework.

Taught by François Fleuret from École Polytechnique Fédérale de Lausanne, Switzerland.

Link: https://fleuret.org/ee559/

#DL #course #learnhardgopro

New study pinpoints the "neurons” in machine learning systems that capture specific features in language-translation tasks: https://t.co/wq30WGQzex

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

Boom! What to read on the weekend, Issue #48

Snapshot: data scientist salaries and jobs in Europe
http://bit.ly/2Rw0wEQ

Artificial Intelligence for Investment Research and Trading Insight
http://bit.ly/2WDT4uO

Goodbye, Mid-Range Shot
http://bit.ly/2WBPeCE

Introduction to Monte Carlo Tree Search: The Game-Changing Algorithm behind DeepMind AlphaGo
http://bit.ly/2WJsfWq

Why chatbots haven’t (yet) lived up to their hype
http://bit.ly/2WAcAIG

Enterprise adoption of AI has grown 270 percent over the past four years
https://zd.net/2Rw0VHm

Diamond Pro’s AI spots imperfection in gems
http://bit.ly/2WB6mZi

A new predictive model for more accurate electrical grid mapping
http://bit.ly/2WDwzXg

What the US-China tech and AI arms race means for the world
http://bit.ly/2WBmKsD

streenart-notstreatart: Train ResNet based model to detect if images is street art or not
http://bit.ly/2RymEhB

​​#стартапы@hitecher

Beijing-based Megvii project focuses on the development of facial recognition technology. Its Face ++ data base boasts solutions developed by 300 thousand programmers. The application performs a widest range of tasks: from identity checks of taxi drivers working for aggregator services to catching criminals. By way of example, it took the application just 7 minutes to verify a journalist’s identity.

This is how the program works: a user's picture is uploaded into the system, it integrates facial biometric data simultaneously tracking 83 key points on the face. It enables identification of a person’s face with 99% accuracy. In addition to identity verification, the program determines the person’s age, race, gender, and emotions. The application provides access to private property: for example, it allows you to track people’s movements inside a building. This capability makes the technology extremely useful for shopping centers and banks to prevent robberies.

Face ++ has a mixed reputation: human rights activists consider it to be an instrument of public control. The situation around the application is aggravated by the fact that Face ++ is part of the smart video surveillance programme in PRC. The program is also used by the Chinese police helping them to identify criminals much more efficiently than through fingerprint analysis or the viewing of videos recorded by surveillance cameras.

The startup is currently working with several companies, such as Ant Financial Alibaba — Alipay units. Face++ can be used to pay for purchases in China by just smiling into the camera. The programme is already used by more than 120 million people in China.

English soccer team hires predictive AI to coach a game. Yes, really.
http://bit.ly/2SgsmJI

Top 10 Presentations from rstudio:conf 2019 – The Best R Conference of the Year!
http://bit.ly/2Sm2koc

Why Ice Cream Is Linked to Shark Attacks – Correlation/Causation Smackdown
http://bit.ly/2Gb5KDX

Poker-Playing AI Tapped For Military Use In $10M Pentagon Deal
http://bit.ly/2GcILs5

Introduction to Pandas
http://bit.ly/2GiEGml

Nobel prizes & Laureates (1901-2018)
https://tabsoft.co/2SjMElm

FiveThirtyEight is sharing the data and code behind some of the articles and graphics.
https://53eig.ht/2HBWWZY

Artificial Intelligence and the Evolution of the Sales Process
http://bit.ly/2HHiapx

Neural Network Models in R
http://bit.ly/2Shksj3

Must-Read Tutorial to Learn Sequence Modeling
http://bit.ly/2Si1UPN

Practical Deep Learning for Coders 2019
http://bit.ly/2Rq2lD7

How to solve 90% of NLP problems: a step-by-step guide
http://bit.ly/2Rs8COw

Policy, process changes needed to safely integrate AI into clinical workflows
http://bit.ly/2Rs8AGo

Physicists use supercomputers and AI to create the most accurate model yet of black hole mergers
http://bit.ly/2WvJKcn

Conversational AI in 2018: An arms race of new products, acquisitions, and more
http://bit.ly/2WBH2lY

Homenum Revelio! AI will destroy sperm donor anonymity
http://bit.ly/2VfGfFC

Do not Miss this Comprehensive 7 Step Process to Ace Data Science Interviews!
http://bit.ly/2EsMHUi

Troubleshooting Deep Neural Networks. A Field Guide to Fixing Your Model
http://bit.ly/2EuRnsV

The 10 Highest Grossing Actors of All Time
https://tabsoft.co/2Etpdy8

Text Analysis for Online Marketers
http://bit.ly/2EswKgW