Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Watson ecosystem and portfolio

Ronan
July 01, 2019

Watson ecosystem and portfolio

A presentation of IBM Watson ecosystems : Services, Studio, Solutions. And how IBM will help you succeed in you AI journey.

Ronan

July 01, 2019
Tweet

More Decks by Ronan

Other Decks in Technology

Transcript

  1. A path to Artificial Intelligence (AI) Watson Studio to go

    further and develop your own models PowerAI to benefit from all power you need to develop and use your own models Watson Services to discover all AI features and power
  2. 5 Watson Services AI Assistant Watson Assistant Knowledge Discovery Natural

    Language Understanding Knowledge Studio Vision Visual Recognition Speech Speech to Text Text to Speech Language Language Translator Natural Language Classifier Empathy Personality Insight Tone Analyzer https://www.ibm.com/watson/products-services/
  3. - Extensive use of API - Configuration GUI for some

    tools - Play with the tools, many docs and hands on - https://ibmcloud-watson-day.mybluemix.net/ - Regular evolutions of tools - Configuration, training 6 Watson commons
  4. - To create « chatbot » but it’s not a

    chatbot - Same Conversation engine for many chats (Slack, Messenger, Web, …) - Use through API or channels - Configuration through GUI - New Skills and Assistants - Can be used for voice chat with Watson Speech - https://watson-assistant- demo.ng.bluemix.net/ 7 Watson Assistant (WA)
  5. - Core of Language understanding - Analyze sentences and extract

    informations - Entities : cars, trees, … - Relationship in between entities : is part of, depends on, … - Keywords : detect keywords in the text - Semantic roles : subject, verb, … - Standard Context - Can be improved by WKS - Pure API - https://natural-language-understanding-demo.ng.bluemix.net 8 Watson Natural Language Understanding (NLU)
  6. - Create a new topology for a new context -

    Specific entities - Specific links - Intensive configuration with 100+ documents to anotate - No training, configuration can be used for many projects - http://knowledge-studio- demo.ng.bluemix.net 9 Watson Knowledge Studio (WKS)
  7. - Analyze document with NLU - Ingest and Index documents

    in WDS - Base of IBM W3 Watson search - Can be improved with WKS - Understand 4 types of documents (HTML, PDF, Word, JSON) 10 Watson Discovery Service (WDS) - Capabilities - Smart Document Understanding - Training - Document Segmentation - Conversion / Enrichment / Normalization - Keywords, Sentiments, Concept, Entities, Categories, Semantic Role, Emotion - Query Language / Natural Language - Passage Retrieval - https://discovery-news-demo.ng.bluemix.net
  8. - Speech To Text (STT) : Transcript Speech into Text

    - Text To Speech (TTS) : Speak out a written Text - Ongoing improvement in French - Can be improve by client - Pure API - Demos - STT : https://speech-to-text-demo.ng.bluemix.net - TTS : https://text-to-speech-demo.ng.bluemix.net 11 Speech
  9. - Personality Insight : Predict personality characteristics, needs and values

    through written text. - https://personality-insights-demo.ng.bluemix.net/ - Tone Analyzer:Understand emotions and communication style in text. - Tones : anger, joy, fear, sadness, analytical, confidence, dare - Point out what part of text triggers a tone - https://tone-analyzer-demo.ng.bluemix.net - Pure API 12 Empathy
  10. - Detect entities in images - Can detect many entities

    - Training for new entities - Pretrained models - Face - Food - Explicit - https://www.ibm.com/watson/s ervices/visual- recognition/demo/#demo 13 Visual Recognition
  11. - Language Translator : Translate from one language to another

    - Using custom model - Document translation (beta) - https://www.ibm.com/watson/services/language-translator/#demo - Natural Language Classifier : Detect intentions used in sentence - Multi-lingual support - Training - https://natural-language-classifier-demo.ng.bluemix.net - Pure API 14 Language
  12. 16 Watson Business Solutions Voice of the Customer Improve management

    of customer feedback Customer Care Virtual Agent Help Customer Services Agents Expert Assist Find the right information for your client professionals Compliance Assist Ensure you comply with all rules and laws
  13. • Fully functional, foundational production applications that use 1 or

    more Watson APIs • Built around specific, repeatable use cases • Configurable and trainable using customer’s data in four weeks • Owned by the client at the end of an engagement • Intended to be sold quickly with a set scope • Able to be leveraged in custom solutions • An IBM Do it for me Services engagement, not a Product 17 Watson Business Solutions
  14. • Customer complaints analyzes • Watson Discovery and Knowledge Studio

    • https://wbs-voice-of-the-customer.mybluemix.net 18 Voice of the Customer Voice of the Customer Improve management of customer feedback
  15. • Chatbot with customer for advice on credit card abroad

    • Watson Assistant and Discovery • https://ccva-abc-bank.mybluemix.net/ 19 Customer Care Virtual Agent Customer Care Virtual Agent Help Customer Services Agents
  16. • Help expert answering client questions • Watson Assistant, Discovery

    and Knowledge Studio • Better demo with expert assist « listening » to the dialog and bringing up information • https://ccva-abc-bank.mybluemix.net/ 20 Expert Assist Expert Assist Find the right information for your client professionals
  17. • Help companies maintaining legal compliance in their documents •

    Watson Compare and Comply (Watson Discovery Service with Element Classification) • https://compliance-assist-prod.us- south.containers.mybluemix.net/collections 21 Compliance Assist Compliance Assist Ensure you comply with all rules and laws
  18. 23 Watson Studio and Data Management Data Watson Studio Watson

    Machine Learning Watson Knowledge Catalog Watson OpenScale https://www.ibm.com/watson/products-services/ Watson Machine Learning Accelerator
  19. - Studio GUI to use common tools - Jupyter Notebooks

    and R Studio for model development - Data preparation and exploration - Workflow to manage projects - Templates 24 Watson Studio
  20. - Machine Learning and Deep Learning - Run Models -

    Integrated with Watson Studio - Machine Learning Accelerator with Power Systems (PowerAI) 25 Watson Machine Learning
  21. - Connect to all datasource in enterprise - Structured and

    Unstructured - SQL / NoSQL - … - Integrated with Watson Studio - Organize your data 26 Watson Knowledge Catalog
  22. - Give trust and transparency on how models work -

    Explain how results are found - Improve models - Bias detection 27 Watson Openscale
  23. A lot of materials to start easily AI : Watson

    Services to start and Watson Studio to go further Infrastructure : VM, containers, kubernetes, baremetal, … Development : Languages, IDE, source and ticket management, tests, … - Docs : https://cloud.ibm.com/developer/watson/documentation - Starter Kits : https://www.ibm.com/watson/developer-resources/ - Huge list of projects in https://developer.ibm.com - Some hands on: https://ibmcloud-watson-day.mybluemix.net/
  24. - IBM Cloud - Public - Dedicated - IBM Cloud

    Private for Data (ICP4D) - On Premise - Other clouds - Watson Studio Desktop - On your laptop ! 30 Where can you find Watson ? Everywhere !
  25. - Start for Free in IBM Cloud ! - Free

    account, no credit card - Lite plan unlimited in time - Plenty of documentation and labs - Docs : https://cloud.ibm.com/developer/watson/documentation - Starter Kits : https://www.ibm.com/watson/developer-resources/ - Huge list of projects in https://developer.ibm.com - Some hands on: https://ibmcloud-watson-day.mybluemix.net/ 31 How to start with Watson ?
  26. Start easily and simply evolve to a full environment Specific

    offer for startups https://developer.ibm.com/startups/ Pay Per Use : You pay only for what you use Free offer for everyone in https://cloud.ibm.com/
  27. - Complexity in Data !!! - Training and testing -

    Preprocessing - Security, data masking - Continuous Relevancy training - Understand Data and process - Subject Matter Experts (SME) 33 Importance of Data management
  28. Watson Customer Engagement Model 35 Increasing IBM involvement in project

    outcomes • You want to engage with Watson in a variety of ways… through digital self service, enablement services and/or deep services partnership • Watson strives make you feel ready and empowered at each level • The Watson Expert Services suite of offerings enable you to do more with Watson capabilities with less reliance on IBM Digital Self Service - Documentation - On-Line Communities - Watson Academy IBM Do It for Me IBM developed, complex production projects IBM Help Me Do It
  29. 36 Watson Expert Services 4 weeks Build your prototype with

    IBM guidance Modify and refine your implementation plan for pilot level deployment Build Plan 5 day workshop Develop your solution architecture and viability assessment Create your development roadmap for cognitive solutions Learn 2-4 day workshop Gain hands-on experience with Watson APIs to understand capabilities Explore your high-level implementation methodology 8 weeks Refine your prototype to production pilot scaling and deployment Gain Watson skill and methodology experience in an IBM supported setting Deploy Design 2 day workshop Design in a user centric manner Develop your Watson solutions using the IBM Design Thinking framework On Demand Consulting 3 month pack of SME assistance Provides consulting ranging from online "Ask an Expert" to onsite SME time Augment any offering above or use as stand alone
  30. - 4% of enterprise are doing AI in production -

    96% are looking at AI - AI is disruptive and enterprise looking at AI will have a competitive advantage 39
  31. - Data are key for success of projects - Watson

    Services are simple to use - Watson Studio for more complex cases - Watson Everywhere (IBM Clouds, Clouds, Local) - Start Free - IBM will help you succeed in your AI journey 40 Key take aways
  32. 41 Group Name / DOC ID / May 23-25, 2018

    / © 2018 IBM Corporation