Home / Course catalog / Build NLU solutions with the expert.ai N... (EPP01)


Get this course

expert.ai NL Platform eLearning → Expert.ai NL Platform: practical courses

Build NLU solutions with the expert.ai NL Platform (EPP01)


Description
This course gets you hands-on with Corpora, Categorization, Extraction projects, Symbolic models and Workflow APIs in the expert.ai NL Platform. This course is specific to the Winter 2022 release of the NL Platform.

No particular technical skills and knowledge are required to take this course though, a basic understanding of NLP, ML, Symbolic and the NL Platform is suggested.

Target audience: computational linguists, data scientists, software developers, knowledge engineers, taxonomists, data analysts, professional services directors, project managers.

Trainers: Samuel Algherini, Gianmarco Saretto, Silvia Ronchi, Caterina Zapparoli.

Content
  • !!! Download the training material FIRST !!!
  • Unit 0 - Corpora Projects
  • Dataset deep dive with Corpora
  • Unit 1 - Categorization projects
  • Categorization projects
  • NLP approaches
  • Categorization workflow
  • Log in and project creation
  • Taxonomy
  • Demo of taxonomy creation
  • How to import a taxonomy
  • Magic Taxonomy
  • Training and test libraries overview
  • Library import
  • Introduction to search functions and analytics
  • Managing libraries
  • How to annotate and share a project
  • High-level algorithms
  • How to launch an experiment
  • Metrics
  • Results analysis: micro, macro, sample and weighted average
  • Test Unit 1
  • Unit 2 - extraction projects
  • Introduction to extraction projects and extraction classes
  • Extraction workflow, project creation and extraction classes
  • Library import
  • How to annotate for extraction
  • Good annotation practices and validation
  • Project sharing and annotation
  • High-level algorithms
  • Explainable extraction
  • How to launch an extraction experiment
  • Results analysis
  • Improvement strategies
  • Test Unit 2
  • UNIT 2.1 - Advanced features for EXTRACTION
  • Active Learning
  • Matching Strategies
  • UNIT 3 - PRO TIPS ON SETTINGS AND HYPERPARAMETERS
  • Categorization and extraction: high-level settings
  • Categorization hyperparameters settings and explainable categorization settings
  • UNIT 4 - KNOWLEDGE GRAPH ENRICHMENT
  • Customizing the Knowledge Graph
  • Unit 5 - Workflow
  • Runtime Core and API KEY Creation
  • Workflow: parallel creation (URL) through wizard
  • Workflow: sequential creation (Tika Converter), publishing, testing & monitoring
  • Javascript hexagon
  • Additional operator: tunnel
  • Leveraging Knowledge Models in your Workflows
  • How to publish the workflow and conduct URL test
  • The map block
  • Postman collection
  • Test Unit 5
  • UNIT 6 - SYMBOLIC MODELS DEVELOPMENT
  • The Symbolic approach in a nutshell
  • Expert.ai Studio: a closer look
  • UNIT 6.1 - SEMANTICS AT EXPERT.AI
  • Semantics at expert.ai pt.1
  • Browsing the Knowledge Graph
  • Semantics at expert.ai pt.2
  • Semantics at expert.ai pt.3
  • Hands-on with disambiguation
  • Disambiguation: PRO tips
  • UNIT 6.2 - SETTING UP A PROJECT IN STUDIO
  • Setting up your first Studio project
  • PRO TIP: how to soft-wrap your documents
  • UNIT 6.3 - CATEGORIZATION PROJECTS
  • How Categorization works with Symbolic
  • How Categorization rules work
  • Creating/importing a taxonomy
  • Dealing with build errors
  • Annotating your documents (Categorization)
  • Creating a rules file
  • The structure of a Categorization rule
  • Apply rules comments
  • Rules writing: TIPS & TRICKS
  • KEYWORD and LEMMA attributes
  • LEMMA: PRO tips
  • The SYNCON attribute
  • The LIST attribute
  • Boolean operators: AND & AND NOT
  • Rules scopes: PRO tips
  • UNIT 6.4 - EXTRACTION PROJECTS
  • How Extraction works with Symbolic
  • Annotating your documents (Extraction)
  • How Extraction rules work
  • PRO tips: ANCESTOR and TYPE attributes
  • Combining attributes and SYNCON UNKNOWN
  • Leveraging regexes in your rules: PATTERN attribute
  • Using positional sequence operators
  • Combine negations and positional sequence operators
  • Hybrid Categorization and Extraction rules
  • UNIT 6.5 - QUALITY ASSURANCE
  • Quality Assurance: measuring your Symbolic model's quality
  • Fine tuning strategies for Symbolic models
  • Tools to support fine tuning techniques
  • Comparative Analysis: comparing your dataset analyses
  • UNIT 6.6 - WORKING IN TEAM
  • Working in team: BEST PRACTICES
  • UNIT 6.7 - STUDIO-NL PLATFORM INTEGRATION
  • Assessing models quality leveraging the expert.ai NL Platform
  • Connect Studio to the expert.ai NL Platform
  • UNIT 6.8 - MODEL DEPLOYMENT & API
  • How to deploy a Symbolic model: the procedure
  • Deploy your Symbolic model to a Workflow API
  • CERTIFICATION TEST
  • Final remarks
  • Course Feedback Survey
  • Certification Test
Completion rules
  • All units must be completed
  • Leads to a certificate with a duration: 1 year
Back