SlideShare uma empresa Scribd logo
1 de 45
Baixar para ler offline
Research Updates from Rasa:
Transformers in NLU and Dialogue
Alan Nichol
Co-Founder & CTO, Rasa
We’ll cover two recent research projects from Rasa
● Why we do research at Rasa
● DIET: new NLU architecture
● TED: new dialogue policy
● Q&A
● More resources
Why we do research
OUR MISSION
Empower all makers to create AI
assistants that work for everyone
To do that, we’re building the standard infrastructure for conversational AI
@alanmnichol
Open Source Community Applied Research
*Cumulative Pypi and Github downloads
of Rasa open source tools
Downloads
2M+ 8,000+
Forum Members
300+
Contributors
Rasa X: downloaded in 135 countries
Downloads
Our community is friendly, global, and growing fast
RASA COMMUNITY
Check out rasa.com/research to see some of the projects we’re working on
Today’s topics
Conversational AI requires NLU and Dialogue management
@alanmnichol
We’ll talk about the role of transformer architectures in both of these tasks
Dual Intent and Entity
Transformer (DIET)
DIET is our new neural network architecture for NLU
💡 To understand how DIET works, check
our YouTube channel
What is DIET?
● New state of the art neural network architecture for NLU
● Predicts intents and entities together
● Plug and play pretrained language models
How to use DIET in your Rasa project
Here’s an example config.yml
Before the DIET model, you can specify any
featurizer.
In our experiments, we use:
● Sparse features (aka no pre-trained model)
● GloVe (word vectors)
● BERT (large language model)
● ConveRT (pre-trained encoder for
conversations)
Experiments on the NLU-benchmark dataset
● Repo is on github
● Domain: human-robot interaction (smart home setting)
● 64 different intents
● 54 different entity types
● ~26k labelled examples
Previous state of the art:
● HERMIT NLU (Vanzo, Bastianelli, and Lemon @ SIGdial 2019)
● uses ELMo embeddings
Result 1: DIET outperforms SotA even without any pretrained embeddings
Previous state of the art: intent: 87.55 entities: 84.74
@alanmnichol
Result 2: GloVe embeddings perform better than BERT
Result 3: ConveRT embeddings perform best on the NLU-benchmark dataset
Result 4: DIET outperforms fine-tuning BERT
Which featurizer is best depends on your dataset, so try different ones!
At Rasa, we don’t believe in “one size fits all”
machine learning
● We aim to provide sensible defaults and
suggestions
● BUT even more important that Rasa models
are easy to customize
Share your results and compare notes with 8000+
Rasa developers at forum.rasa.com
Transformer
Embedding Dialogue
policy (TED)
Conversational AI requires NLU and Dialogue management
@alanmnichol
Happy paths are best described in code
@alanmnichol
But real conversations don’t follow the happy path
@alanmnichol
Users will always surprise you
@alanmnichol
And will revisit topics as they please
@alanmnichol
You can’t anticipate all the ways users will act
@alanmnichol
Can we build a model that handles this?
People typically use a recurrent neural net (RNN) to model dialogue
h1
h2
h3
y1
y2
y3
W
W
W
W
W
W
W
W
@alanmnichol
But not all input should be treated equally
@alanmnichol
https://ai.googleblog.com/2017/08/transformer-novel-neural-network.html
Transformers (AKA self-attention) are now state of the art for many tasks
https://distill.pub/2016/augmented-rnns/
@alanmnichol
We found out that the Transformer Embedding Dialogue policy can untangle
sub-dialogues
@alanmnichol
paper
TED is available in Rasa 1.3 and up
The embedding policy (TED)
● better at handling unseen edge cases
● less likely to get confused when users
behave in highly unexpected ways
● used in combination with other policies
● Becoming the new default ML policy
(replacing KerasPolicy)
With all contextual assistants, please write tests!
@alanmnichol
So we now have the algorithms to handle this
@alanmnichol
But you also need training data!
@alanmnichol
Review conversations and
improve your assistant based
on what you learn
Collect
conversations
between users and
your assistant
Ship updates using
continuous
integration &
deployment
Build minimum
viable assistant Improve by
talking to the
assistant
Improve using
conversations
with real users
Improve using
conversations
with test users
Quality of assistant
Rasa Open Source (Local)
Rasa X (Server)
Rasa Open Source is an open
source framework for natural
language understanding, dialogue
management, and integrations.
Rasa X is a toolset used
to improve a contextual
assistant built using
Rasa Open Source.
Deploy your minimum viable assistant on a server and improve it using Rasa X
Rasa X: downloaded in 135 countries
Q&A
How can the transitions be effectively tested in a large
dialogue tree, to ensure that the policy works as expected?
Will Rasa provide a way to select the best policy based on my
use case and training data?
Does Rasa support multi-label classification for intents and
entities?
Is there a way to do cross domain transfer learning using
Rasa? (For instance, a healthcare assistant trained on
healthcare terminology to an IT help desk assistant)
Resources
To get started, watch the Rasa Masterclass on YouTube
● Unpacking the TED Policy in Rasa Open Source ( Rasa Blog)
● Introducing DIET: state-of-the-art architecture that outperforms fine-tuning BERT
and is 6X faster to train (Rasa Blog)
● Rasa Algorithm Whiteboard - Diet Architecture 1: How it Works (YouTube)
● Rasa Algorithm Whiteboard - Diet Architecture 2: Design Decisions (YouTube)
Further Reading
https://forum.rasa.com
Alan Nichol
Co-founder & CTO
alan@rasa.com
@alanmnichol
Email me! →
alan@rasa.com

Mais conteúdo relacionado

Mais procurados

Mais procurados (20)

Build Mandarin AI Conversational Agent with Rasa
Build Mandarin AI Conversational Agent with RasaBuild Mandarin AI Conversational Agent with Rasa
Build Mandarin AI Conversational Agent with Rasa
 
End-to-end dialogue systems, or a feature which wasn’t meant to happen | Rasa...
End-to-end dialogue systems, or a feature which wasn’t meant to happen | Rasa...End-to-end dialogue systems, or a feature which wasn’t meant to happen | Rasa...
End-to-end dialogue systems, or a feature which wasn’t meant to happen | Rasa...
 
Conversational AI with Transformer Models
Conversational AI with Transformer ModelsConversational AI with Transformer Models
Conversational AI with Transformer Models
 
What Is GPT-3 And Why Is It Revolutionizing Artificial Intelligence?
What Is GPT-3 And Why Is It Revolutionizing Artificial Intelligence?What Is GPT-3 And Why Is It Revolutionizing Artificial Intelligence?
What Is GPT-3 And Why Is It Revolutionizing Artificial Intelligence?
 
Deprecating the state machine: building conversational AI with the Rasa stack...
Deprecating the state machine: building conversational AI with the Rasa stack...Deprecating the state machine: building conversational AI with the Rasa stack...
Deprecating the state machine: building conversational AI with the Rasa stack...
 
Deep learning for NLP and Transformer
 Deep learning for NLP  and Transformer Deep learning for NLP  and Transformer
Deep learning for NLP and Transformer
 
GPT-2: Language Models are Unsupervised Multitask Learners
GPT-2: Language Models are Unsupervised Multitask LearnersGPT-2: Language Models are Unsupervised Multitask Learners
GPT-2: Language Models are Unsupervised Multitask Learners
 
An introduction to the Transformers architecture and BERT
An introduction to the Transformers architecture and BERTAn introduction to the Transformers architecture and BERT
An introduction to the Transformers architecture and BERT
 
CHATGPT.pptx
CHATGPT.pptxCHATGPT.pptx
CHATGPT.pptx
 
GPT for polyglots and language learners.pptx
GPT for polyglots and language learners.pptxGPT for polyglots and language learners.pptx
GPT for polyglots and language learners.pptx
 
Nlp and transformer (v3s)
Nlp and transformer (v3s)Nlp and transformer (v3s)
Nlp and transformer (v3s)
 
[Paper Reading] Attention is All You Need
[Paper Reading] Attention is All You Need[Paper Reading] Attention is All You Need
[Paper Reading] Attention is All You Need
 
Natural language processing PPT presentation
Natural language processing PPT presentationNatural language processing PPT presentation
Natural language processing PPT presentation
 
How does ChatGPT work: an Information Retrieval perspective
How does ChatGPT work: an Information Retrieval perspectiveHow does ChatGPT work: an Information Retrieval perspective
How does ChatGPT work: an Information Retrieval perspective
 
Natural language processing and transformer models
Natural language processing and transformer modelsNatural language processing and transformer models
Natural language processing and transformer models
 
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
 
OpenAI-Copilot-ChatGPT.pptx
OpenAI-Copilot-ChatGPT.pptxOpenAI-Copilot-ChatGPT.pptx
OpenAI-Copilot-ChatGPT.pptx
 
‘Big models’: the success and pitfalls of Transformer models in natural langu...
‘Big models’: the success and pitfalls of Transformer models in natural langu...‘Big models’: the success and pitfalls of Transformer models in natural langu...
‘Big models’: the success and pitfalls of Transformer models in natural langu...
 
Natural language processing
Natural language processingNatural language processing
Natural language processing
 
Generative Models for General Audiences
Generative Models for General AudiencesGenerative Models for General Audiences
Generative Models for General Audiences
 

Semelhante a Research Updates from Rasa: Transformers in NLU and Dialogue

Semelhante a Research Updates from Rasa: Transformers in NLU and Dialogue (20)

Scalable state of-the-art conversational AI
Scalable state of-the-art conversational AIScalable state of-the-art conversational AI
Scalable state of-the-art conversational AI
 
Top 7 Frameworks for Integration AI in App Development
Top 7 Frameworks for Integration AI in App DevelopmentTop 7 Frameworks for Integration AI in App Development
Top 7 Frameworks for Integration AI in App Development
 
SDL Trados Studio 2017, Jocelyn He (SDL)
SDL Trados Studio 2017, Jocelyn He (SDL)SDL Trados Studio 2017, Jocelyn He (SDL)
SDL Trados Studio 2017, Jocelyn He (SDL)
 
Hadoop training in mumbai
Hadoop training in mumbaiHadoop training in mumbai
Hadoop training in mumbai
 
You already have an LxP, you just don't know it
You already have an LxP, you just don't know itYou already have an LxP, you just don't know it
You already have an LxP, you just don't know it
 
The Latest Advances in Generative AI_ Exploring New Technology for Data Integ...
The Latest Advances in Generative AI_ Exploring New Technology for Data Integ...The Latest Advances in Generative AI_ Exploring New Technology for Data Integ...
The Latest Advances in Generative AI_ Exploring New Technology for Data Integ...
 
SIM RTP Meeting - So Who's Using Open Source Anyway?
SIM RTP Meeting - So Who's Using Open Source Anyway?SIM RTP Meeting - So Who's Using Open Source Anyway?
SIM RTP Meeting - So Who's Using Open Source Anyway?
 
Realizing AI Conversational Bot
Realizing AI Conversational BotRealizing AI Conversational Bot
Realizing AI Conversational Bot
 
OS Accelerate London - 09/16/15
OS Accelerate London - 09/16/15OS Accelerate London - 09/16/15
OS Accelerate London - 09/16/15
 
Architecting for analytics
Architecting for analyticsArchitecting for analytics
Architecting for analytics
 
Keynote at Converge 2019
Keynote at Converge 2019Keynote at Converge 2019
Keynote at Converge 2019
 
Trikonf 2015 - Community, Studio and the OpenExchange
Trikonf 2015  - Community, Studio and the OpenExchangeTrikonf 2015  - Community, Studio and the OpenExchange
Trikonf 2015 - Community, Studio and the OpenExchange
 
Training Chatbots and Conversational Artificial Intelligence Agents with Amaz...
Training Chatbots and Conversational Artificial Intelligence Agents with Amaz...Training Chatbots and Conversational Artificial Intelligence Agents with Amaz...
Training Chatbots and Conversational Artificial Intelligence Agents with Amaz...
 
NTLM - Open Source Language AI Tools
NTLM - Open Source Language AI ToolsNTLM - Open Source Language AI Tools
NTLM - Open Source Language AI Tools
 
Copy of OTel Me All About OpenTelemetry The Current & Future State, Navigatin...
Copy of OTel Me All About OpenTelemetry The Current & Future State, Navigatin...Copy of OTel Me All About OpenTelemetry The Current & Future State, Navigatin...
Copy of OTel Me All About OpenTelemetry The Current & Future State, Navigatin...
 
Dev learn: Our notes
Dev learn: Our notesDev learn: Our notes
Dev learn: Our notes
 
Presentation 4 1 REDY ok.pptx
Presentation 4 1  REDY ok.pptxPresentation 4 1  REDY ok.pptx
Presentation 4 1 REDY ok.pptx
 
Tensorflow a brief introduction (1).pptx
Tensorflow a brief introduction (1).pptxTensorflow a brief introduction (1).pptx
Tensorflow a brief introduction (1).pptx
 
Internet of Things Brings On Development Demands That DevOps Manages, Say Exp...
Internet of Things Brings On Development Demands That DevOps Manages, Say Exp...Internet of Things Brings On Development Demands That DevOps Manages, Say Exp...
Internet of Things Brings On Development Demands That DevOps Manages, Say Exp...
 
xAPI: The Landscape
xAPI: The LandscapexAPI: The Landscape
xAPI: The Landscape
 

Mais de Rasa Technologies

Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...
Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...
Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...
Rasa Technologies
 

Mais de Rasa Technologies (20)

Six Steps to Conversation Driven Development
Six Steps to Conversation Driven DevelopmentSix Steps to Conversation Driven Development
Six Steps to Conversation Driven Development
 
Beyond Sentiment Analysis: Creating Engaging Conversational Experiences throu...
Beyond Sentiment Analysis: Creating Engaging Conversational Experiences throu...Beyond Sentiment Analysis: Creating Engaging Conversational Experiences throu...
Beyond Sentiment Analysis: Creating Engaging Conversational Experiences throu...
 
Using Rasa to Power an Immersive Multimedia Conversational Experience | Rasa ...
Using Rasa to Power an Immersive Multimedia Conversational Experience | Rasa ...Using Rasa to Power an Immersive Multimedia Conversational Experience | Rasa ...
Using Rasa to Power an Immersive Multimedia Conversational Experience | Rasa ...
 
How to Effectively Test Your Chatbot | Rasa Summit
How to Effectively Test Your Chatbot  | Rasa SummitHow to Effectively Test Your Chatbot  | Rasa Summit
How to Effectively Test Your Chatbot | Rasa Summit
 
Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...
Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...
Voice First: Ready Your Content to Serve 50% of Global Searches | Rasa Summit...
 
The missing link: How AI can help create a safer society and better businesse...
The missing link: How AI can help create a safer society and better businesse...The missing link: How AI can help create a safer society and better businesse...
The missing link: How AI can help create a safer society and better businesse...
 
Boss - Bringing More Diversity to Tech | Rasa Summit
Boss - Bringing More Diversity to Tech | Rasa SummitBoss - Bringing More Diversity to Tech | Rasa Summit
Boss - Bringing More Diversity to Tech | Rasa Summit
 
How Our Team Uses Rasa to Learn from Real Conversations | Rasa Summit
How Our Team Uses Rasa to Learn from Real Conversations | Rasa SummitHow Our Team Uses Rasa to Learn from Real Conversations | Rasa Summit
How Our Team Uses Rasa to Learn from Real Conversations | Rasa Summit
 
Applying Conversational AI in the Enterprise
Applying Conversational AI in the EnterpriseApplying Conversational AI in the Enterprise
Applying Conversational AI in the Enterprise
 
Ai = your data | Rasa Summit 2021
Ai = your data | Rasa Summit 2021Ai = your data | Rasa Summit 2021
Ai = your data | Rasa Summit 2021
 
Supercharging User Interfaces with Rasa | Rasa Summit 2021
Supercharging User Interfaces with Rasa | Rasa Summit 2021Supercharging User Interfaces with Rasa | Rasa Summit 2021
Supercharging User Interfaces with Rasa | Rasa Summit 2021
 
STAR: A Schema-Guided Dialog Dataset for Transfer Learning | Rasa Summit 2021
STAR: A Schema-Guided Dialog Dataset for Transfer Learning | Rasa Summit 2021 STAR: A Schema-Guided Dialog Dataset for Transfer Learning | Rasa Summit 2021
STAR: A Schema-Guided Dialog Dataset for Transfer Learning | Rasa Summit 2021
 
Continuous Improvement of Conversational AI in Production | Rasa Summit
Continuous Improvement of Conversational AI in Production | Rasa SummitContinuous Improvement of Conversational AI in Production | Rasa Summit
Continuous Improvement of Conversational AI in Production | Rasa Summit
 
Ethnobots: Reimagining Chatbots as Ethnographic Research Tools | Rasa Summit ...
Ethnobots: Reimagining Chatbots as Ethnographic Research Tools | Rasa Summit ...Ethnobots: Reimagining Chatbots as Ethnographic Research Tools | Rasa Summit ...
Ethnobots: Reimagining Chatbots as Ethnographic Research Tools | Rasa Summit ...
 
The State of Conversation Design - Designing for the Conversational Future
The State of Conversation Design - Designing for the Conversational FutureThe State of Conversation Design - Designing for the Conversational Future
The State of Conversation Design - Designing for the Conversational Future
 
Rasa Open Source - What's next?
Rasa Open Source - What's next?Rasa Open Source - What's next?
Rasa Open Source - What's next?
 
Building an AI Assistant Factory - Rasa Summit 2021
Building an AI Assistant Factory - Rasa Summit 2021Building an AI Assistant Factory - Rasa Summit 2021
Building an AI Assistant Factory - Rasa Summit 2021
 
Building an End-to-End Test Automation Pipeline for Conversational AI | Rasa ...
Building an End-to-End Test Automation Pipeline for Conversational AI | Rasa ...Building an End-to-End Test Automation Pipeline for Conversational AI | Rasa ...
Building an End-to-End Test Automation Pipeline for Conversational AI | Rasa ...
 
Deploy your Rasa Chatbots like a Boss with DevOps | Rasa Summit 2021
Deploy your Rasa Chatbots like a Boss with DevOps | Rasa Summit 2021Deploy your Rasa Chatbots like a Boss with DevOps | Rasa Summit 2021
Deploy your Rasa Chatbots like a Boss with DevOps | Rasa Summit 2021
 
What’s next in CDD: Intent Clashes and Selective Confidence | Rasa Summit 2021
What’s next in CDD: Intent Clashes and Selective Confidence | Rasa Summit 2021What’s next in CDD: Intent Clashes and Selective Confidence | Rasa Summit 2021
What’s next in CDD: Intent Clashes and Selective Confidence | Rasa Summit 2021
 

Último

Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Joaquim Jorge
 

Último (20)

Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?
 
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
 
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
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
 
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
 
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
 
Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)
 
GenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdfGenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdf
 
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
 
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
 
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone ProcessorsExploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processors
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024
 
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
 
Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...
 

Research Updates from Rasa: Transformers in NLU and Dialogue

  • 1. Research Updates from Rasa: Transformers in NLU and Dialogue Alan Nichol Co-Founder & CTO, Rasa
  • 2. We’ll cover two recent research projects from Rasa ● Why we do research at Rasa ● DIET: new NLU architecture ● TED: new dialogue policy ● Q&A ● More resources
  • 3. Why we do research
  • 4. OUR MISSION Empower all makers to create AI assistants that work for everyone
  • 5. To do that, we’re building the standard infrastructure for conversational AI @alanmnichol Open Source Community Applied Research
  • 6. *Cumulative Pypi and Github downloads of Rasa open source tools Downloads 2M+ 8,000+ Forum Members 300+ Contributors Rasa X: downloaded in 135 countries Downloads Our community is friendly, global, and growing fast RASA COMMUNITY
  • 7. Check out rasa.com/research to see some of the projects we’re working on
  • 9. Conversational AI requires NLU and Dialogue management @alanmnichol We’ll talk about the role of transformer architectures in both of these tasks
  • 10. Dual Intent and Entity Transformer (DIET)
  • 11. DIET is our new neural network architecture for NLU 💡 To understand how DIET works, check our YouTube channel What is DIET? ● New state of the art neural network architecture for NLU ● Predicts intents and entities together ● Plug and play pretrained language models
  • 12. How to use DIET in your Rasa project Here’s an example config.yml Before the DIET model, you can specify any featurizer. In our experiments, we use: ● Sparse features (aka no pre-trained model) ● GloVe (word vectors) ● BERT (large language model) ● ConveRT (pre-trained encoder for conversations)
  • 13. Experiments on the NLU-benchmark dataset ● Repo is on github ● Domain: human-robot interaction (smart home setting) ● 64 different intents ● 54 different entity types ● ~26k labelled examples Previous state of the art: ● HERMIT NLU (Vanzo, Bastianelli, and Lemon @ SIGdial 2019) ● uses ELMo embeddings
  • 14. Result 1: DIET outperforms SotA even without any pretrained embeddings Previous state of the art: intent: 87.55 entities: 84.74 @alanmnichol
  • 15. Result 2: GloVe embeddings perform better than BERT
  • 16. Result 3: ConveRT embeddings perform best on the NLU-benchmark dataset
  • 17. Result 4: DIET outperforms fine-tuning BERT
  • 18. Which featurizer is best depends on your dataset, so try different ones! At Rasa, we don’t believe in “one size fits all” machine learning ● We aim to provide sensible defaults and suggestions ● BUT even more important that Rasa models are easy to customize Share your results and compare notes with 8000+ Rasa developers at forum.rasa.com
  • 20. Conversational AI requires NLU and Dialogue management @alanmnichol
  • 21. Happy paths are best described in code @alanmnichol
  • 22. But real conversations don’t follow the happy path @alanmnichol
  • 23. Users will always surprise you @alanmnichol
  • 24. And will revisit topics as they please @alanmnichol
  • 25. You can’t anticipate all the ways users will act @alanmnichol
  • 26. Can we build a model that handles this?
  • 27. People typically use a recurrent neural net (RNN) to model dialogue h1 h2 h3 y1 y2 y3 W W W W W W W W @alanmnichol
  • 28. But not all input should be treated equally @alanmnichol https://ai.googleblog.com/2017/08/transformer-novel-neural-network.html
  • 29. Transformers (AKA self-attention) are now state of the art for many tasks https://distill.pub/2016/augmented-rnns/ @alanmnichol
  • 30. We found out that the Transformer Embedding Dialogue policy can untangle sub-dialogues @alanmnichol paper
  • 31. TED is available in Rasa 1.3 and up The embedding policy (TED) ● better at handling unseen edge cases ● less likely to get confused when users behave in highly unexpected ways ● used in combination with other policies ● Becoming the new default ML policy (replacing KerasPolicy) With all contextual assistants, please write tests! @alanmnichol
  • 32. So we now have the algorithms to handle this @alanmnichol
  • 33. But you also need training data! @alanmnichol Review conversations and improve your assistant based on what you learn Collect conversations between users and your assistant Ship updates using continuous integration & deployment
  • 34. Build minimum viable assistant Improve by talking to the assistant Improve using conversations with real users Improve using conversations with test users Quality of assistant Rasa Open Source (Local) Rasa X (Server) Rasa Open Source is an open source framework for natural language understanding, dialogue management, and integrations. Rasa X is a toolset used to improve a contextual assistant built using Rasa Open Source. Deploy your minimum viable assistant on a server and improve it using Rasa X
  • 35. Rasa X: downloaded in 135 countries
  • 36. Q&A
  • 37. How can the transitions be effectively tested in a large dialogue tree, to ensure that the policy works as expected?
  • 38. Will Rasa provide a way to select the best policy based on my use case and training data?
  • 39. Does Rasa support multi-label classification for intents and entities?
  • 40. Is there a way to do cross domain transfer learning using Rasa? (For instance, a healthcare assistant trained on healthcare terminology to an IT help desk assistant)
  • 42. To get started, watch the Rasa Masterclass on YouTube
  • 43. ● Unpacking the TED Policy in Rasa Open Source ( Rasa Blog) ● Introducing DIET: state-of-the-art architecture that outperforms fine-tuning BERT and is 6X faster to train (Rasa Blog) ● Rasa Algorithm Whiteboard - Diet Architecture 1: How it Works (YouTube) ● Rasa Algorithm Whiteboard - Diet Architecture 2: Design Decisions (YouTube) Further Reading
  • 45. Alan Nichol Co-founder & CTO alan@rasa.com @alanmnichol Email me! → alan@rasa.com