SlideShare a Scribd company logo
1 of 16
Farzaneh S. Fard
Machine Learning Scientist at Fluent.ai
March 8th, 2019 – at Element AI
1Farzaneh.fard@fluent.ai
Ubiquitous Machine Learning
• In this talk:
• Cognitive Robotics
• Speech recognition
2Farzaneh.fard@fluent.ai
COGNITIVE ROBOTICS
• Aspects of cognitive robotics:
• Flexible, safe, and adaptive robots
• Obtaining better understanding of the brain
3Farzaneh.fard@fluent.ai
BRAIN CONTROL SYSTEMS
• Deliberative planning
• Needs the model of the environment to plan ahead
• Model-based reinforcement learning / supervised learning
• Forward and inverse models
• Habitual controller
• Learns the reward function by repeating past experiences that are most
rewarding
• Model-free reinforcement learning
• Actor-critic
Integrating two controllers?
4Farzaneh.fard@fluent.ai
PROS & CONS
5
• Deliberative planning
• Time consuming
• Computationally expensive
• Very accurate
• Habitual controller
• Fast
• Less computationally expensive
• Needs many examples to train
Arbitrated Predictive Actor-Critic (APAC)
Farzaneh.fard@fluent.ai
SIMULATING HUMAN TARGET REACHING
• One step mapping
• Multi-step reaching
• Conditions
• Static/ changing kinematics
• Static/ changing target
6Farzaneh.fard@fluent.ai
APAC for TARGET REACHING
• Deliberative planning
• Habitual controller
• With arbitrator
7Farzaneh.fard@fluent.ai
APAC – MAIN RESULTS
• Adaptive to changes both
in Kinematics and the
environment
• Performs as accurate as
planning while relying
more on habits
8Farzaneh.fard@fluent.ai
APAC – MAIN RESULTS
• Automatically shifting to
habitual system from
planning after learning
9Farzaneh.fard@fluent.ai
SPEECH RECOGNITION
• Keyword spotting
● Providing users a fully hands-free experience
● Distinguishing wake phrase from ordinary talks
● Always listening - power consumption
● Classification: wake phrase or filler
• Challenges
• Noise robust
• Far field: signal attenuation, reverberation/echo
• Small footprint
• Speaker identification
10Farzaneh.fard@fluent.ai
TIME DELAY NEURAL NETWORK
11
Speech input Spectrogram
Farzaneh.fard@fluent.ai
EXPERIMENTS
12Farzaneh.fard@fluent.ai
FLUENT.AI LATEST PRODUCT
13Farzaneh.fard@fluent.ai
• AI on the edge
• On device if on the cloud is not necessary
• smart lock
• Universal Speech recognition
• In any language or multi language
• 3 languages: English/German/Korean
14Farzaneh.fard@fluent.ai
WE ARE HIRING
Machine Learning Scientist & Interns
www.fluent.ai/careers
15Farzaneh.fard@fluent.ai
16Farzaneh.fard@fluent.ai

More Related Content

Similar to Ubiquitous Machine Learning: Lessons from DeepRL in Robotics and Speech, by Farzaneh S. Fard, Ph.D., Machine Learning Scientist at Fluent.ai

Machine Learning for Designers
Machine Learning for DesignersMachine Learning for Designers
Machine Learning for DesignersMemi Beltrame
 
APMP - Automation Support for Proposal Development
APMP - Automation Support for Proposal DevelopmentAPMP - Automation Support for Proposal Development
APMP - Automation Support for Proposal DevelopmentVisibleThread
 
Types of Artificial Intelligence.ppt
Types of Artificial Intelligence.pptTypes of Artificial Intelligence.ppt
Types of Artificial Intelligence.pptGEETHAS668001
 
Nimble Framework - Software architecture and design in agile era - PSQT Template
Nimble Framework - Software architecture and design in agile era - PSQT TemplateNimble Framework - Software architecture and design in agile era - PSQT Template
Nimble Framework - Software architecture and design in agile era - PSQT Templatetjain
 
Summer IT Training in Chandigarh by JK Soft Tech Solutions
Summer IT Training in Chandigarh by JK Soft Tech SolutionsSummer IT Training in Chandigarh by JK Soft Tech Solutions
Summer IT Training in Chandigarh by JK Soft Tech SolutionsNaresh Dhamija
 
The Un-researched Persona
The Un-researched PersonaThe Un-researched Persona
The Un-researched Personanlemonier
 
Agile Software Development
Agile Software DevelopmentAgile Software Development
Agile Software DevelopmentSynerzip
 
Synerzip Agile Software Development
Synerzip Agile Software DevelopmentSynerzip Agile Software Development
Synerzip Agile Software DevelopmentSynerzip
 
Artificial Intelligence and The Complexity
Artificial Intelligence and The ComplexityArtificial Intelligence and The Complexity
Artificial Intelligence and The ComplexityHendri Karisma
 
NLP,expert,robotics.pptx
NLP,expert,robotics.pptxNLP,expert,robotics.pptx
NLP,expert,robotics.pptxAmanBadesra1
 
Machine Learning Applications in E-learning - Bias, Risks, and Mitigations
Machine Learning Applications in E-learning - Bias, Risks, and MitigationsMachine Learning Applications in E-learning - Bias, Risks, and Mitigations
Machine Learning Applications in E-learning - Bias, Risks, and MitigationsStella Lee
 
Enterprise Agile Transformation
Enterprise Agile TransformationEnterprise Agile Transformation
Enterprise Agile TransformationPooja Wandile
 
Lesson learned in developing UI and mobile apps blibli.com
Lesson learned in developing UI and mobile apps blibli.comLesson learned in developing UI and mobile apps blibli.com
Lesson learned in developing UI and mobile apps blibli.comifnu bima
 
Agile Software Development and DevOps 21092019
Agile Software Development and DevOps 21092019Agile Software Development and DevOps 21092019
Agile Software Development and DevOps 21092019Ahmed Misbah
 
Agile Offsharing: Using Pair Work to Overcome Nearshoring Difficulties
Agile Offsharing: Using Pair Work to OvercomeNearshoring DifficultiesAgile Offsharing: Using Pair Work to OvercomeNearshoring Difficulties
Agile Offsharing: Using Pair Work to Overcome Nearshoring DifficultiesMobileSolutionsDTAG
 
Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...
Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...
Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...3Play Media
 
Agile software development for startups
Agile software development for startupsAgile software development for startups
Agile software development for startupsHemant Elhence
 
Good Help is Hard to Find
Good Help is Hard to FindGood Help is Hard to Find
Good Help is Hard to FindElaine Meyer
 

Similar to Ubiquitous Machine Learning: Lessons from DeepRL in Robotics and Speech, by Farzaneh S. Fard, Ph.D., Machine Learning Scientist at Fluent.ai (20)

Machine Learning for Designers
Machine Learning for DesignersMachine Learning for Designers
Machine Learning for Designers
 
APMP - Automation Support for Proposal Development
APMP - Automation Support for Proposal DevelopmentAPMP - Automation Support for Proposal Development
APMP - Automation Support for Proposal Development
 
Types of Artificial Intelligence.ppt
Types of Artificial Intelligence.pptTypes of Artificial Intelligence.ppt
Types of Artificial Intelligence.ppt
 
Nimble Framework - Software architecture and design in agile era - PSQT Template
Nimble Framework - Software architecture and design in agile era - PSQT TemplateNimble Framework - Software architecture and design in agile era - PSQT Template
Nimble Framework - Software architecture and design in agile era - PSQT Template
 
Summer IT Training in Chandigarh by JK Soft Tech Solutions
Summer IT Training in Chandigarh by JK Soft Tech SolutionsSummer IT Training in Chandigarh by JK Soft Tech Solutions
Summer IT Training in Chandigarh by JK Soft Tech Solutions
 
The Un-researched Persona
The Un-researched PersonaThe Un-researched Persona
The Un-researched Persona
 
Agile Software Development
Agile Software DevelopmentAgile Software Development
Agile Software Development
 
Synerzip Agile Software Development
Synerzip Agile Software DevelopmentSynerzip Agile Software Development
Synerzip Agile Software Development
 
Artificial Intelligence and The Complexity
Artificial Intelligence and The ComplexityArtificial Intelligence and The Complexity
Artificial Intelligence and The Complexity
 
NLP,expert,robotics.pptx
NLP,expert,robotics.pptxNLP,expert,robotics.pptx
NLP,expert,robotics.pptx
 
Machine Learning Applications in E-learning - Bias, Risks, and Mitigations
Machine Learning Applications in E-learning - Bias, Risks, and MitigationsMachine Learning Applications in E-learning - Bias, Risks, and Mitigations
Machine Learning Applications in E-learning - Bias, Risks, and Mitigations
 
Enterprise Agile Transformation
Enterprise Agile TransformationEnterprise Agile Transformation
Enterprise Agile Transformation
 
Lesson learned in developing UI and mobile apps blibli.com
Lesson learned in developing UI and mobile apps blibli.comLesson learned in developing UI and mobile apps blibli.com
Lesson learned in developing UI and mobile apps blibli.com
 
Mwalimu
MwalimuMwalimu
Mwalimu
 
Agile Software Development and DevOps 21092019
Agile Software Development and DevOps 21092019Agile Software Development and DevOps 21092019
Agile Software Development and DevOps 21092019
 
Code reviews
Code reviewsCode reviews
Code reviews
 
Agile Offsharing: Using Pair Work to Overcome Nearshoring Difficulties
Agile Offsharing: Using Pair Work to OvercomeNearshoring DifficultiesAgile Offsharing: Using Pair Work to OvercomeNearshoring Difficulties
Agile Offsharing: Using Pair Work to Overcome Nearshoring Difficulties
 
Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...
Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...
Usability and Accessibility Have a Conversation: How Accessibility and UI/UX ...
 
Agile software development for startups
Agile software development for startupsAgile software development for startups
Agile software development for startups
 
Good Help is Hard to Find
Good Help is Hard to FindGood Help is Hard to Find
Good Help is Hard to Find
 

More from WiMLDSMontreal

The Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
The Five Ws of Funding, by Sahar Ansary, Partner, R&D PartnersThe Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
The Five Ws of Funding, by Sahar Ansary, Partner, R&D PartnersWiMLDSMontreal
 
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...WiMLDSMontreal
 
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...WiMLDSMontreal
 
How to build a data science project in a corporate setting, by Soraya Christi...
How to build a data science project in a corporate setting, by Soraya Christi...How to build a data science project in a corporate setting, by Soraya Christi...
How to build a data science project in a corporate setting, by Soraya Christi...WiMLDSMontreal
 
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...WiMLDSMontreal
 
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...WiMLDSMontreal
 
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...WiMLDSMontreal
 
Artistic Applications of AI, by Luba Elliott, AI Curator
Artistic Applications of AI, by Luba Elliott, AI CuratorArtistic Applications of AI, by Luba Elliott, AI Curator
Artistic Applications of AI, by Luba Elliott, AI CuratorWiMLDSMontreal
 
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...WiMLDSMontreal
 
Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...
Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...
Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...WiMLDSMontreal
 
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...WiMLDSMontreal
 

More from WiMLDSMontreal (11)

The Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
The Five Ws of Funding, by Sahar Ansary, Partner, R&D PartnersThe Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
The Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
 
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
 
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
 
How to build a data science project in a corporate setting, by Soraya Christi...
How to build a data science project in a corporate setting, by Soraya Christi...How to build a data science project in a corporate setting, by Soraya Christi...
How to build a data science project in a corporate setting, by Soraya Christi...
 
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
 
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
 
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
 
Artistic Applications of AI, by Luba Elliott, AI Curator
Artistic Applications of AI, by Luba Elliott, AI CuratorArtistic Applications of AI, by Luba Elliott, AI Curator
Artistic Applications of AI, by Luba Elliott, AI Curator
 
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
 
Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...
Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...
Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head ...
 
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...
 

Recently uploaded

Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...HostedbyConfluent
 
How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?XfilesPro
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...shyamraj55
 
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...Neo4j
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
Azure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAzure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAndikSusilo4
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Alan Dix
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):comworks
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
IAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI SolutionsIAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI SolutionsEnterprise Knowledge
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphNeo4j
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxnull - The Open Security Community
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountPuma Security, LLC
 
Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraDeakin University
 

Recently uploaded (20)

Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
 
How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
 
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food Manufacturing
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
Azure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAzure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & Application
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptxE-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
 
IAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI SolutionsIAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI Solutions
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path Mount
 
Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning era
 

Ubiquitous Machine Learning: Lessons from DeepRL in Robotics and Speech, by Farzaneh S. Fard, Ph.D., Machine Learning Scientist at Fluent.ai

  • 1. Farzaneh S. Fard Machine Learning Scientist at Fluent.ai March 8th, 2019 – at Element AI 1Farzaneh.fard@fluent.ai
  • 2. Ubiquitous Machine Learning • In this talk: • Cognitive Robotics • Speech recognition 2Farzaneh.fard@fluent.ai
  • 3. COGNITIVE ROBOTICS • Aspects of cognitive robotics: • Flexible, safe, and adaptive robots • Obtaining better understanding of the brain 3Farzaneh.fard@fluent.ai
  • 4. BRAIN CONTROL SYSTEMS • Deliberative planning • Needs the model of the environment to plan ahead • Model-based reinforcement learning / supervised learning • Forward and inverse models • Habitual controller • Learns the reward function by repeating past experiences that are most rewarding • Model-free reinforcement learning • Actor-critic Integrating two controllers? 4Farzaneh.fard@fluent.ai
  • 5. PROS & CONS 5 • Deliberative planning • Time consuming • Computationally expensive • Very accurate • Habitual controller • Fast • Less computationally expensive • Needs many examples to train Arbitrated Predictive Actor-Critic (APAC) Farzaneh.fard@fluent.ai
  • 6. SIMULATING HUMAN TARGET REACHING • One step mapping • Multi-step reaching • Conditions • Static/ changing kinematics • Static/ changing target 6Farzaneh.fard@fluent.ai
  • 7. APAC for TARGET REACHING • Deliberative planning • Habitual controller • With arbitrator 7Farzaneh.fard@fluent.ai
  • 8. APAC – MAIN RESULTS • Adaptive to changes both in Kinematics and the environment • Performs as accurate as planning while relying more on habits 8Farzaneh.fard@fluent.ai
  • 9. APAC – MAIN RESULTS • Automatically shifting to habitual system from planning after learning 9Farzaneh.fard@fluent.ai
  • 10. SPEECH RECOGNITION • Keyword spotting ● Providing users a fully hands-free experience ● Distinguishing wake phrase from ordinary talks ● Always listening - power consumption ● Classification: wake phrase or filler • Challenges • Noise robust • Far field: signal attenuation, reverberation/echo • Small footprint • Speaker identification 10Farzaneh.fard@fluent.ai
  • 11. TIME DELAY NEURAL NETWORK 11 Speech input Spectrogram Farzaneh.fard@fluent.ai
  • 14. • AI on the edge • On device if on the cloud is not necessary • smart lock • Universal Speech recognition • In any language or multi language • 3 languages: English/German/Korean 14Farzaneh.fard@fluent.ai
  • 15. WE ARE HIRING Machine Learning Scientist & Interns www.fluent.ai/careers 15Farzaneh.fard@fluent.ai

Editor's Notes

  1. ML has wide range of application.. In almost everything you do during a normal day you are using a machine learning algotrithms. Only on our phone we use many of these algorithms, such as pattern recognition, weather forecast, spam checking for your emails, shopping recommenders and many more… Here I will talk briefly about two of these applications. So let’s jump into this.
  2. Cognitive robotics is a multi disciplinary area that has two major aspects. One aspect is developing robots that can mimic human behavior, which leads us to a more flexible and adaptive robots which are safe to interact and cooperate with human. The other aspect is obtaining better understanding of the brain and try to help people with brain disorders, by studying and testing the theories of how the brain works onto robots and analyze robot's behavior.
  3. There are many models and algorithms that have been proposed in robotics for each control system. Deliberative planning is time consuming to predict the best action at the time and needs a predefined model. The habitual controller is fast but it is useless in dynamic environment with changing reward functions.
  4. - The learning happens through integration of these two controller
  5. Since the main reason to have a brain is being able to move, we chose the target reaching task. We simulated a 2D robot arm and we applied the model to learn the nonlinear function of arm movement to reach the target located in the environment (shown in the picture).
  6. Integrator could be Kalman-filter.
  7. We learn a new task with deliberately planning and then it becomes a habit! This work is not similar to Barto&Sutton’s work (supervised AC and Dyna-Q)