SlideShare a Scribd company logo
1 of 23
William Clements, PhD
Mastering the challenges of open source
reinforcement learning
william.clements@indust.ai
indust.ai
Independent
Applied AI
R&D lab
What this talk is about
 RL has huge potential for decision-making and control
 We were one of the first companies to start a dedicated applied
RL lab in mid-2018
 In this talk, we will share what we learned in the process
2indust.ai
Who we are
Solutions in time series analysis, speech recognition, sentiment analysis, …
World-class research in learning from bad data and in decision-making:
 3 publications (including AAAI 2020)
 Academic collaborations: Ecole polytechnique, Oxford, FAIR, …
3indust.ai
Providing medium-size businesses with easy access
to a rare, deep and broad AI resources pool they
need to unlock future growth
A skilled and experienced team, passionate
about applying AI to the core of industrial and
financial services businesses
Overview
 What is reinforcement learning, and what is it good for?
 Use case: building an RL system for smart homes
 Challenges to overcome
 Is open source the answer?
 Our approach
4indust.ai
What is reinforcement learning?
 Create agents that learn to maximize a given objective by interacting with the
environment
 Advances have been made possible by combining reinforcement learning and neural
networks
Environment
Agent
Observation
Reward
Action
Trial and Error
5indust.ai
Reinforcement learning: state of the art
 RL has taken off in the last few years, and achieved impressive results
Go : Learns to beat
humans with no initial
knowledge apart from
the rules of the game
Silver, David, et al.
"Mastering the game of Go
with deep neural networks
and tree search." nature
529.7587 (2016): 484.
Henderson, Peter, et al. "Deep reinforcement learning that matters."
Thirty-Second AAAI Conference on Artificial Intelligence. 2018.
Publications per year
6indust.ai
How RL differs from traditional ML
ML
Requires data
First collect data, then
learn
Mature field
Cannot outperform the
people who label
RL
Requires a simulator
Exploration to collect the right
experience
New field, mostly academic
Can outperform human-specified
systems
7indust.ai
Industrial applications
 Potential applications exist in any system that reacts to external stimulus
Healthcare Recommendation Thermal Control Robotics
A human body reacts
to treatments
A customer reacts
to advertisement
A building reacts to
heating elements
A robot reacts to
motor controls
- Digital doctor
- Dynamic treatments
- Facebook notifications
- Newspaper links
- Google data centers
- 3D printing
- Industrial ovens
- Delivery drones
- Sorting robots
- Robotic tool use
Other fields: logistics, data networks, autonomous driving, portfolio optimization, etc.
Despite this potential, very few real world implementations
8indust.ai
Example Use Case: Smart Home
Role playing: Imagine you’ve been asked to use RL to improve thermal control
How would you do this ?
Observations: thermostat + energy readings
Actions: air conditioning or radiator controls
RL can work here but it won’t be easy
9indust.ai
Why RL is hard – human aspects
 Requires a change of mindset away from traditional ML
 Very little expertise available outside of academia
 RL requires either a simulator or exploration in real life
 Will your customer trust your RL agent?
10indust.ai
Why RL is hard – technical aspects
High compute
requirements
100 years of simulated
experience just to turn
this cube around
Bad at generalizing
A car trained with a green background
fails when the color changes
Hard to specify
« cobra effect » : you may
not get what you want
Equip smart homes
with GPUs?
What will happen in a heat wave? Heat/cost may not be the best
reward for a smart home
11indust.ai
Open Source RL to the rescue?
- In 2018, 72% of companies used open source software, even for critical tasks
(source: Linux Foundation)
- Open source ML tools have been hugely successful (scikit-learn, tensorflow,
pytorch, keras, etc.)
However, open source RL is not the same as open source ML
12indust.ai
 Two choices in RL: build from scratch or build from open source (virtually
no proprietary software yet)
 Why open source:
Open Source RL vs Open Source ML
 Maturity:
 Open source ML has been around for a long time (scikit-learn: 2007), with both industry
and academia in mind
 Open source RL is much more recent (OpenAI gym: 2016), mostly for the academic
community
 Structure:
 Open source ML provides the algorithm, you provide the data
 Open source RL can provide both the algorithm and the simulator
13indust.ai
Open Source Resources
Environments
Spriteworld
OpenAI Gym
BSuite
DM Control
AI Safety Gridworld
PyBullet
OpenSpiel
RLCard
OffWorld Gym
Industrial
Benchmark
DMLab
Algorithms
TFRL
Dopamine
OpenAI
Baselines
Stable
Baselines
ChainerRL
KerasRL
Frameworks
Surreal
Facebook
ReAgent
RLLib
Coach
Simple RL
SLM Lab
14indust.ai
Open Source Resources: Environments
Environments
Spriteworld
OpenAI Gym
BSuite
DM Control
AI Safety Gridworld
PyBullet
OpenSpiel
RLCard
OffWorld Gym
Industrial
Benchmark
DMLab OffWorld GymOpenAI Gym
PyBullet DMLab
15indust.ai
Issues with open source Environments
Environments are generally designed as academic benchmarks, with no
connection to real world applications
You will have to make
your own environment
16indust.ai
Open source Algorithms and Frameworks
https://github.com/openai/baselines
Allow for comparisons between
algorithms
Can support distributed
calculations
Adapted for specific applications
https://surreal.stanford.edu/
https://github.com/facebookresearch/ReAgent
17indust.ai
Issues with open source Algorithms and Frameworks
Open source algorithms tend to be:
- Designed for specific environments
- Not easily customisable
- Written by researchers, not developers
- Not always reproducible
Henderson, Peter, et al. "Deep reinforcement learning that matters."
Thirty-Second AAAI Conference on Artificial Intelligence. 2018.
Example: Different implementations
of the same algorithm
You will have to make
your own framework
18indust.ai
Our approach
19indust.ai
Unreliable and opaque open
source frameworks
We built our own, which we plan
to open source!
- Great for upskilling
Unsuitable open source
environments
We make our own (proprietary)
simulators, working with industry experts
Disconnect between
academia and industry
We started a research activity, in
collaboration with academics
Smart home: our solution
Design a simulator with industry
experts, using OpenAI gym template
Validate simulator using real world data
Build RL framework from
ground up using
Benchmark algorithms on academic environments
Train algorithms on simulator, real world tests, and refinements
20indust.ai
How we tackle the RL human challenge
- Managers often don’t see opportunities: we are evangelists who work
on changing mindsets
- Talents shortage: we do outreach
- “Nuit de l’IA” at Polytechnique
- Lecture at the Machine Learning Summer School 2019
- “Maths et IA” event at Université Paris-Sud 2019
21indust.ai
 Challenges are not just technical:
Conclusion
22indust.ai
 RL has huge potential for impact on industry
 Open source RL does not yet rise to the challenge
 We’ve built a dedicated R&D lab to solve RL for industry
23indust.ai
Want to know more? Drop us a line:
info@indust.ai

More Related Content

What's hot

14 Startups Leading the Artificial Intelligence (AI) Revolution
14 Startups Leading the Artificial Intelligence (AI) Revolution14 Startups Leading the Artificial Intelligence (AI) Revolution
14 Startups Leading the Artificial Intelligence (AI) RevolutionNVIDIA
 
Building Data Ecosystems for Accelerated Discovery
Building Data Ecosystems for Accelerated DiscoveryBuilding Data Ecosystems for Accelerated Discovery
Building Data Ecosystems for Accelerated Discoveryadamkraut
 
AI-SDV 2021: Francisco Webber - Efficiency is the New Precision
AI-SDV 2021: Francisco Webber - Efficiency is the New PrecisionAI-SDV 2021: Francisco Webber - Efficiency is the New Precision
AI-SDV 2021: Francisco Webber - Efficiency is the New PrecisionDr. Haxel Consult
 
Understanding Cognitive Applications: A Framework - Sue Feldman
Understanding Cognitive Applications:  A Framework - Sue FeldmanUnderstanding Cognitive Applications:  A Framework - Sue Feldman
Understanding Cognitive Applications: A Framework - Sue Feldmandiannepatricia
 
Introduction To Data Science
Introduction To Data ScienceIntroduction To Data Science
Introduction To Data ScienceSpotle.ai
 
The Best & Worst Uses of AI in Software Testing
The Best & Worst Uses of AI in Software TestingThe Best & Worst Uses of AI in Software Testing
The Best & Worst Uses of AI in Software TestingEficode
 
Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...
Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...
Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...Tomasz Bednarz
 
Alumni talk-university-of-kachchh
Alumni talk-university-of-kachchhAlumni talk-university-of-kachchh
Alumni talk-university-of-kachchhChetan Khatri
 
National seminar on emergence of internet of things (io t) trends and challe...
National seminar on emergence of internet of things (io t)  trends and challe...National seminar on emergence of internet of things (io t)  trends and challe...
National seminar on emergence of internet of things (io t) trends and challe...Ajay Ohri
 
Introduction to Data Science - Week 4 - Tools and Technologies in Data Science
Introduction to Data Science - Week 4 - Tools and Technologies in Data ScienceIntroduction to Data Science - Week 4 - Tools and Technologies in Data Science
Introduction to Data Science - Week 4 - Tools and Technologies in Data ScienceFerdin Joe John Joseph PhD
 
Introduction of Data Science
Introduction of Data ScienceIntroduction of Data Science
Introduction of Data ScienceJason Geng
 
IBM Academy of Technology & Cognitive Computing
IBM Academy of Technology & Cognitive ComputingIBM Academy of Technology & Cognitive Computing
IBM Academy of Technology & Cognitive ComputingNico Chillemi
 
Ai priti sajja original webinar ai post covid may 2020
Ai priti sajja original webinar ai post covid may 2020Ai priti sajja original webinar ai post covid may 2020
Ai priti sajja original webinar ai post covid may 2020Priti Srinivas Sajja
 
Cognitive computing big_data_statistical_analytics
Cognitive computing big_data_statistical_analyticsCognitive computing big_data_statistical_analytics
Cognitive computing big_data_statistical_analyticsPietro Leo
 
Pycon india-2016-success-story
Pycon india-2016-success-storyPycon india-2016-success-story
Pycon india-2016-success-storyChetan Khatri
 
Internet of things initiative-cskskv
Internet of things   initiative-cskskvInternet of things   initiative-cskskv
Internet of things initiative-cskskvChetan Khatri
 
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017Carol Smith
 
A Pragmatic AI Maturity Model
A Pragmatic AI Maturity ModelA Pragmatic AI Maturity Model
A Pragmatic AI Maturity ModelDATAVERSITY
 
IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...
IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...
IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...Dr. Haxel Consult
 

What's hot (20)

Machine Learning Applications to IoT
Machine Learning Applications to IoTMachine Learning Applications to IoT
Machine Learning Applications to IoT
 
14 Startups Leading the Artificial Intelligence (AI) Revolution
14 Startups Leading the Artificial Intelligence (AI) Revolution14 Startups Leading the Artificial Intelligence (AI) Revolution
14 Startups Leading the Artificial Intelligence (AI) Revolution
 
Building Data Ecosystems for Accelerated Discovery
Building Data Ecosystems for Accelerated DiscoveryBuilding Data Ecosystems for Accelerated Discovery
Building Data Ecosystems for Accelerated Discovery
 
AI-SDV 2021: Francisco Webber - Efficiency is the New Precision
AI-SDV 2021: Francisco Webber - Efficiency is the New PrecisionAI-SDV 2021: Francisco Webber - Efficiency is the New Precision
AI-SDV 2021: Francisco Webber - Efficiency is the New Precision
 
Understanding Cognitive Applications: A Framework - Sue Feldman
Understanding Cognitive Applications:  A Framework - Sue FeldmanUnderstanding Cognitive Applications:  A Framework - Sue Feldman
Understanding Cognitive Applications: A Framework - Sue Feldman
 
Introduction To Data Science
Introduction To Data ScienceIntroduction To Data Science
Introduction To Data Science
 
The Best & Worst Uses of AI in Software Testing
The Best & Worst Uses of AI in Software TestingThe Best & Worst Uses of AI in Software Testing
The Best & Worst Uses of AI in Software Testing
 
Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...
Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...
Platform for Big Data Analytics and Visual Analytics: CSIRO use cases. Februa...
 
Alumni talk-university-of-kachchh
Alumni talk-university-of-kachchhAlumni talk-university-of-kachchh
Alumni talk-university-of-kachchh
 
National seminar on emergence of internet of things (io t) trends and challe...
National seminar on emergence of internet of things (io t)  trends and challe...National seminar on emergence of internet of things (io t)  trends and challe...
National seminar on emergence of internet of things (io t) trends and challe...
 
Introduction to Data Science - Week 4 - Tools and Technologies in Data Science
Introduction to Data Science - Week 4 - Tools and Technologies in Data ScienceIntroduction to Data Science - Week 4 - Tools and Technologies in Data Science
Introduction to Data Science - Week 4 - Tools and Technologies in Data Science
 
Introduction of Data Science
Introduction of Data ScienceIntroduction of Data Science
Introduction of Data Science
 
IBM Academy of Technology & Cognitive Computing
IBM Academy of Technology & Cognitive ComputingIBM Academy of Technology & Cognitive Computing
IBM Academy of Technology & Cognitive Computing
 
Ai priti sajja original webinar ai post covid may 2020
Ai priti sajja original webinar ai post covid may 2020Ai priti sajja original webinar ai post covid may 2020
Ai priti sajja original webinar ai post covid may 2020
 
Cognitive computing big_data_statistical_analytics
Cognitive computing big_data_statistical_analyticsCognitive computing big_data_statistical_analytics
Cognitive computing big_data_statistical_analytics
 
Pycon india-2016-success-story
Pycon india-2016-success-storyPycon india-2016-success-story
Pycon india-2016-success-story
 
Internet of things initiative-cskskv
Internet of things   initiative-cskskvInternet of things   initiative-cskskv
Internet of things initiative-cskskv
 
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
 
A Pragmatic AI Maturity Model
A Pragmatic AI Maturity ModelA Pragmatic AI Maturity Model
A Pragmatic AI Maturity Model
 
IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...
IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...
IC-SDV 2018: Aleksandar Kapisoda (Boehringer) Using Machine Learning for Auto...
 

Similar to #OSSPARIS19 - Overcoming open source challenges in reinforcement learning - WILLIAM CLEMENTS, Stealth

Benefiting from Semantic AI along the data life cycle
Benefiting from Semantic AI along the data life cycleBenefiting from Semantic AI along the data life cycle
Benefiting from Semantic AI along the data life cycleMartin Kaltenböck
 
Yuri Van Geest - Exponential Organizations
Yuri Van Geest - Exponential OrganizationsYuri Van Geest - Exponential Organizations
Yuri Van Geest - Exponential OrganizationsBAQMaR
 
Ai open powermeetupmarch25th_latest
Ai open powermeetupmarch25th_latestAi open powermeetupmarch25th_latest
Ai open powermeetupmarch25th_latestGanesan Narayanasamy
 
Japan 20200724 v13
Japan 20200724 v13Japan 20200724 v13
Japan 20200724 v13ISSIP
 
AI and ML Series - Introduction to Generative AI and LLMs - Session 1
AI and ML Series - Introduction to Generative AI and LLMs - Session 1AI and ML Series - Introduction to Generative AI and LLMs - Session 1
AI and ML Series - Introduction to Generative AI and LLMs - Session 1DianaGray10
 
Tutorial helsinki 20180313 v1
Tutorial helsinki 20180313 v1Tutorial helsinki 20180313 v1
Tutorial helsinki 20180313 v1ISSIP
 
2013 - Yhat - YC app.pdf
2013 - Yhat - YC app.pdf2013 - Yhat - YC app.pdf
2013 - Yhat - YC app.pdfAustin Ogilvie
 
Understanding the New World of Cognitive Computing
Understanding the New World of Cognitive ComputingUnderstanding the New World of Cognitive Computing
Understanding the New World of Cognitive ComputingDATAVERSITY
 
How Decentralized AI can Dominate the Global AI Ecosystem
How Decentralized AI can Dominate the Global AI EcosystemHow Decentralized AI can Dominate the Global AI Ecosystem
How Decentralized AI can Dominate the Global AI EcosystemEficode
 
GENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICS
GENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICSGENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICS
GENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICSNITHYA637064
 
Ai open powermeetupmarch25th
Ai open powermeetupmarch25thAi open powermeetupmarch25th
Ai open powermeetupmarch25thIBM
 
Ai open powermeetupmarch25th
Ai open powermeetupmarch25thAi open powermeetupmarch25th
Ai open powermeetupmarch25thIBM
 
State of AI Report 2022 - ONLINE.pptx
State of AI Report 2022 - ONLINE.pptxState of AI Report 2022 - ONLINE.pptx
State of AI Report 2022 - ONLINE.pptxEithuThutun
 
Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...
Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...
Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...Dataconomy Media
 
Allaboutailuminarylabsjanuary122017 170112151616
Allaboutailuminarylabsjanuary122017 170112151616Allaboutailuminarylabsjanuary122017 170112151616
Allaboutailuminarylabsjanuary122017 170112151616Quang Lê
 
Hype vs. Reality: The AI Explainer
Hype vs. Reality: The AI ExplainerHype vs. Reality: The AI Explainer
Hype vs. Reality: The AI ExplainerLuminary Labs
 
State of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdfState of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdfvizologi
 
Hicss52 20190108 v3
Hicss52 20190108 v3Hicss52 20190108 v3
Hicss52 20190108 v3ISSIP
 

Similar to #OSSPARIS19 - Overcoming open source challenges in reinforcement learning - WILLIAM CLEMENTS, Stealth (20)

Benefiting from Semantic AI along the data life cycle
Benefiting from Semantic AI along the data life cycleBenefiting from Semantic AI along the data life cycle
Benefiting from Semantic AI along the data life cycle
 
Yuri Van Geest - Exponential Organizations
Yuri Van Geest - Exponential OrganizationsYuri Van Geest - Exponential Organizations
Yuri Van Geest - Exponential Organizations
 
Ai open powermeetupmarch25th_latest
Ai open powermeetupmarch25th_latestAi open powermeetupmarch25th_latest
Ai open powermeetupmarch25th_latest
 
Japan 20200724 v13
Japan 20200724 v13Japan 20200724 v13
Japan 20200724 v13
 
AI and ML Series - Introduction to Generative AI and LLMs - Session 1
AI and ML Series - Introduction to Generative AI and LLMs - Session 1AI and ML Series - Introduction to Generative AI and LLMs - Session 1
AI and ML Series - Introduction to Generative AI and LLMs - Session 1
 
Tutorial helsinki 20180313 v1
Tutorial helsinki 20180313 v1Tutorial helsinki 20180313 v1
Tutorial helsinki 20180313 v1
 
2013 - Yhat - YC app.pdf
2013 - Yhat - YC app.pdf2013 - Yhat - YC app.pdf
2013 - Yhat - YC app.pdf
 
Understanding the New World of Cognitive Computing
Understanding the New World of Cognitive ComputingUnderstanding the New World of Cognitive Computing
Understanding the New World of Cognitive Computing
 
How Decentralized AI can Dominate the Global AI Ecosystem
How Decentralized AI can Dominate the Global AI EcosystemHow Decentralized AI can Dominate the Global AI Ecosystem
How Decentralized AI can Dominate the Global AI Ecosystem
 
GENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICS
GENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICSGENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICS
GENERATIVE ARTIFICIAL INTELLIGENCE &DATA ANALYTICS
 
Ai open powermeetupmarch25th
Ai open powermeetupmarch25thAi open powermeetupmarch25th
Ai open powermeetupmarch25th
 
Ai open powermeetupmarch25th
Ai open powermeetupmarch25thAi open powermeetupmarch25th
Ai open powermeetupmarch25th
 
State of AI Report 2022 - ONLINE.pptx
State of AI Report 2022 - ONLINE.pptxState of AI Report 2022 - ONLINE.pptx
State of AI Report 2022 - ONLINE.pptx
 
Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...
Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...
Ajit Jaokar, Data Science for IoT professor at Oxford University “Enterprise ...
 
Semantic AI
Semantic AISemantic AI
Semantic AI
 
Allaboutailuminarylabsjanuary122017 170112151616
Allaboutailuminarylabsjanuary122017 170112151616Allaboutailuminarylabsjanuary122017 170112151616
Allaboutailuminarylabsjanuary122017 170112151616
 
About Machine and real
About Machine and realAbout Machine and real
About Machine and real
 
Hype vs. Reality: The AI Explainer
Hype vs. Reality: The AI ExplainerHype vs. Reality: The AI Explainer
Hype vs. Reality: The AI Explainer
 
State of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdfState of AI Report 2022 - ONLINE.pdf
State of AI Report 2022 - ONLINE.pdf
 
Hicss52 20190108 v3
Hicss52 20190108 v3Hicss52 20190108 v3
Hicss52 20190108 v3
 

More from Paris Open Source Summit

#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...
#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...
#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...Paris Open Source Summit
 
#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...
#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...
#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...Paris Open Source Summit
 
#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...
#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...
#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...Paris Open Source Summit
 
#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino
#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino
#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, ArduinoParis Open Source Summit
 
#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...
#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...
#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...Paris Open Source Summit
 
#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...
#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...
#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...Paris Open Source Summit
 
#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix
#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix
#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, ZabbixParis Open Source Summit
 
#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria
#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria
#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, InriaParis Open Source Summit
 
#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...
#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...
#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...Paris Open Source Summit
 
#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches ...
#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches  ...#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches  ...
#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches ...Paris Open Source Summit
 
#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...
#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...
#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...Paris Open Source Summit
 
#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...
#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...
#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...Paris Open Source Summit
 
#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...
#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...
#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...Paris Open Source Summit
 
#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...
#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...
#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...Paris Open Source Summit
 
#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...
#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...
#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...Paris Open Source Summit
 
#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...
#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...
#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...Paris Open Source Summit
 
#OSSPARIS19 - Table ronde : souveraineté des données
#OSSPARIS19 - Table ronde : souveraineté des données #OSSPARIS19 - Table ronde : souveraineté des données
#OSSPARIS19 - Table ronde : souveraineté des données Paris Open Source Summit
 
#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...
#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...
#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...Paris Open Source Summit
 
#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...
#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...
#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...Paris Open Source Summit
 
#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...
#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...
#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...Paris Open Source Summit
 

More from Paris Open Source Summit (20)

#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...
#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...
#OSSPARIS19 : Control your Embedded Linux remotely by using WebSockets - Gian...
 
#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...
#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...
#OSSPARIS19 : A virtual machine approach for microcontroller programming : th...
 
#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...
#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...
#OSSPARIS19 : RIOT: towards open source, secure DevOps on microcontroller-bas...
 
#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino
#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino
#OSSPARIS19 : The evolving (IoT) security landscape - Gianluca Varisco, Arduino
 
#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...
#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...
#OSSPARIS19: Construire des applications IoT "secure-by-design" - Thomas Gaza...
 
#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...
#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...
#OSSPARIS19 : Detecter des anomalies de séries temporelles à la volée avec Wa...
 
#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix
#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix
#OSSPARIS19 : Supervision d'objets connectés industriels - Eric DOANE, Zabbix
 
#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria
#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria
#OSSPARIS19: Introduction to scikit-learn - Olivier Grisel, Inria
 
#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...
#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...
#OSSPARIS19 - Fostering disruptive innovation in AI with JEDI - André Loesekr...
 
#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches ...
#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches  ...#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches  ...
#OSSPARIS19 : Comment ONLYOFFICE aide à organiser les travaux de recherches ...
 
#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...
#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...
#OSSPARIS19 : MDPH : une solution collaborative open source pour l'instructio...
 
#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...
#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...
#OSSPARIS19 - Understanding Open Source Governance - Gilles Gravier, Wipro Li...
 
#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...
#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...
#OSSPARIS19 : Publier du code Open Source dans une banque : Mission impossibl...
 
#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...
#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...
#OSSPARIS19 : Libre à vous ! Raconter les libertés informatiques à la radio -...
 
#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...
#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...
#OSSPARIS19 - Le logiciel libre : un enjeu politique et social - Etienne Gonn...
 
#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...
#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...
#OSSPARIS19 - Conflits d’intérêt & concurrence : la place de l’éditeur dans l...
 
#OSSPARIS19 - Table ronde : souveraineté des données
#OSSPARIS19 - Table ronde : souveraineté des données #OSSPARIS19 - Table ronde : souveraineté des données
#OSSPARIS19 - Table ronde : souveraineté des données
 
#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...
#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...
#OSSPARIS19 - Comment financer un projet de logiciel libre - LUDOVIC DUBOST, ...
 
#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...
#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...
#OSSPARIS19 - BlueMind v4 : les dessous technologiques de 10 ans de travail p...
 
#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...
#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...
#OSSPARIS19 - Tuto de première installation de VITAM, un système d'archivage ...
 

Recently uploaded

AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAndrey Devyatkin
 
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessAdvantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessPixlogix Infotech
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherRemote DBA Services
 
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps ScriptAutomating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Scriptwesley chun
 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CVKhem
 
Developing An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of BrazilDeveloping An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of BrazilV3cube
 
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityPrincipled Technologies
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationMichael W. Hawkins
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century educationjfdjdjcjdnsjd
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024The Digital Insurer
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc
 
Tech Trends Report 2024 Future Today Institute.pdf
Tech Trends Report 2024 Future Today Institute.pdfTech Trends Report 2024 Future Today Institute.pdf
Tech Trends Report 2024 Future Today Institute.pdfhans926745
 
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...apidays
 
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slideHistor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slidevu2urc
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024The Digital Insurer
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Drew Madelung
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...Neo4j
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationRadu Cotescu
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
 

Recently uploaded (20)

AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of Terraform
 
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessAdvantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your Business
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps ScriptAutomating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Script
 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
 
Developing An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of BrazilDeveloping An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of Brazil
 
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day Presentation
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
Tech Trends Report 2024 Future Today Institute.pdf
Tech Trends Report 2024 Future Today Institute.pdfTech Trends Report 2024 Future Today Institute.pdf
Tech Trends Report 2024 Future Today Institute.pdf
 
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
 
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slideHistor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slide
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 

#OSSPARIS19 - Overcoming open source challenges in reinforcement learning - WILLIAM CLEMENTS, Stealth

  • 1. William Clements, PhD Mastering the challenges of open source reinforcement learning william.clements@indust.ai indust.ai Independent Applied AI R&D lab
  • 2. What this talk is about  RL has huge potential for decision-making and control  We were one of the first companies to start a dedicated applied RL lab in mid-2018  In this talk, we will share what we learned in the process 2indust.ai
  • 3. Who we are Solutions in time series analysis, speech recognition, sentiment analysis, … World-class research in learning from bad data and in decision-making:  3 publications (including AAAI 2020)  Academic collaborations: Ecole polytechnique, Oxford, FAIR, … 3indust.ai Providing medium-size businesses with easy access to a rare, deep and broad AI resources pool they need to unlock future growth A skilled and experienced team, passionate about applying AI to the core of industrial and financial services businesses
  • 4. Overview  What is reinforcement learning, and what is it good for?  Use case: building an RL system for smart homes  Challenges to overcome  Is open source the answer?  Our approach 4indust.ai
  • 5. What is reinforcement learning?  Create agents that learn to maximize a given objective by interacting with the environment  Advances have been made possible by combining reinforcement learning and neural networks Environment Agent Observation Reward Action Trial and Error 5indust.ai
  • 6. Reinforcement learning: state of the art  RL has taken off in the last few years, and achieved impressive results Go : Learns to beat humans with no initial knowledge apart from the rules of the game Silver, David, et al. "Mastering the game of Go with deep neural networks and tree search." nature 529.7587 (2016): 484. Henderson, Peter, et al. "Deep reinforcement learning that matters." Thirty-Second AAAI Conference on Artificial Intelligence. 2018. Publications per year 6indust.ai
  • 7. How RL differs from traditional ML ML Requires data First collect data, then learn Mature field Cannot outperform the people who label RL Requires a simulator Exploration to collect the right experience New field, mostly academic Can outperform human-specified systems 7indust.ai
  • 8. Industrial applications  Potential applications exist in any system that reacts to external stimulus Healthcare Recommendation Thermal Control Robotics A human body reacts to treatments A customer reacts to advertisement A building reacts to heating elements A robot reacts to motor controls - Digital doctor - Dynamic treatments - Facebook notifications - Newspaper links - Google data centers - 3D printing - Industrial ovens - Delivery drones - Sorting robots - Robotic tool use Other fields: logistics, data networks, autonomous driving, portfolio optimization, etc. Despite this potential, very few real world implementations 8indust.ai
  • 9. Example Use Case: Smart Home Role playing: Imagine you’ve been asked to use RL to improve thermal control How would you do this ? Observations: thermostat + energy readings Actions: air conditioning or radiator controls RL can work here but it won’t be easy 9indust.ai
  • 10. Why RL is hard – human aspects  Requires a change of mindset away from traditional ML  Very little expertise available outside of academia  RL requires either a simulator or exploration in real life  Will your customer trust your RL agent? 10indust.ai
  • 11. Why RL is hard – technical aspects High compute requirements 100 years of simulated experience just to turn this cube around Bad at generalizing A car trained with a green background fails when the color changes Hard to specify « cobra effect » : you may not get what you want Equip smart homes with GPUs? What will happen in a heat wave? Heat/cost may not be the best reward for a smart home 11indust.ai
  • 12. Open Source RL to the rescue? - In 2018, 72% of companies used open source software, even for critical tasks (source: Linux Foundation) - Open source ML tools have been hugely successful (scikit-learn, tensorflow, pytorch, keras, etc.) However, open source RL is not the same as open source ML 12indust.ai  Two choices in RL: build from scratch or build from open source (virtually no proprietary software yet)  Why open source:
  • 13. Open Source RL vs Open Source ML  Maturity:  Open source ML has been around for a long time (scikit-learn: 2007), with both industry and academia in mind  Open source RL is much more recent (OpenAI gym: 2016), mostly for the academic community  Structure:  Open source ML provides the algorithm, you provide the data  Open source RL can provide both the algorithm and the simulator 13indust.ai
  • 14. Open Source Resources Environments Spriteworld OpenAI Gym BSuite DM Control AI Safety Gridworld PyBullet OpenSpiel RLCard OffWorld Gym Industrial Benchmark DMLab Algorithms TFRL Dopamine OpenAI Baselines Stable Baselines ChainerRL KerasRL Frameworks Surreal Facebook ReAgent RLLib Coach Simple RL SLM Lab 14indust.ai
  • 15. Open Source Resources: Environments Environments Spriteworld OpenAI Gym BSuite DM Control AI Safety Gridworld PyBullet OpenSpiel RLCard OffWorld Gym Industrial Benchmark DMLab OffWorld GymOpenAI Gym PyBullet DMLab 15indust.ai
  • 16. Issues with open source Environments Environments are generally designed as academic benchmarks, with no connection to real world applications You will have to make your own environment 16indust.ai
  • 17. Open source Algorithms and Frameworks https://github.com/openai/baselines Allow for comparisons between algorithms Can support distributed calculations Adapted for specific applications https://surreal.stanford.edu/ https://github.com/facebookresearch/ReAgent 17indust.ai
  • 18. Issues with open source Algorithms and Frameworks Open source algorithms tend to be: - Designed for specific environments - Not easily customisable - Written by researchers, not developers - Not always reproducible Henderson, Peter, et al. "Deep reinforcement learning that matters." Thirty-Second AAAI Conference on Artificial Intelligence. 2018. Example: Different implementations of the same algorithm You will have to make your own framework 18indust.ai
  • 19. Our approach 19indust.ai Unreliable and opaque open source frameworks We built our own, which we plan to open source! - Great for upskilling Unsuitable open source environments We make our own (proprietary) simulators, working with industry experts Disconnect between academia and industry We started a research activity, in collaboration with academics
  • 20. Smart home: our solution Design a simulator with industry experts, using OpenAI gym template Validate simulator using real world data Build RL framework from ground up using Benchmark algorithms on academic environments Train algorithms on simulator, real world tests, and refinements 20indust.ai
  • 21. How we tackle the RL human challenge - Managers often don’t see opportunities: we are evangelists who work on changing mindsets - Talents shortage: we do outreach - “Nuit de l’IA” at Polytechnique - Lecture at the Machine Learning Summer School 2019 - “Maths et IA” event at Université Paris-Sud 2019 21indust.ai  Challenges are not just technical:
  • 22. Conclusion 22indust.ai  RL has huge potential for impact on industry  Open source RL does not yet rise to the challenge  We’ve built a dedicated R&D lab to solve RL for industry
  • 23. 23indust.ai Want to know more? Drop us a line: info@indust.ai

Editor's Notes

  1. RL is the most computing-power intensive discipline in AI. Below the training times for the most advanced models in RL, NLP and Machine Vision :   BERT (Google, NLP) : 64 GPU, 1 week training Open5 (OpenAI, RL) : 256 GPU, 64000 CPU-core, 1 month training ImageNet (FastAI, CV) : 128 GPU,18 mintues RL is barely taught at University : Out of 42 classes in the leading French teaching program in AI (MVA, ENS Cachan), 2 is dedicated to RL, 16 on Computer Vision (as of Jan-19)
  2. RL is the most computing-power intensive discipline in AI. Below the training times for the most advanced models in RL, NLP and Machine Vision :   BERT (Google, NLP) : 64 GPU, 1 week training Open5 (OpenAI, RL) : 256 GPU, 64000 CPU-core, 1 month training ImageNet (FastAI, CV) : 128 GPU,18 mintues RL is barely taught at University : Out of 42 classes in the leading French teaching program in AI (MVA, ENS Cachan), 2 is dedicated to RL, 16 on Computer Vision (as of Jan-19)