SlideShare a Scribd company logo
Knowledge Graph
Convolutional Networks
James Fletcher, Principal Researcher
Follow us @GraknLabs
Machines aren’t as smart as we’d like
Highly domain constrained,
generalisation and model-wise
Reliant on big data
Require complex learning models
Only use flat arrays of input data
Don’t exploit data context
Follow us @GraknLabs
Human Capabilities
Long-term memory
Basic body function
Basic emotions
Appropriate social
response
Skilled movement
coordination
Language
Vision
Sensory processing
Planning
Motivation Motor control
Working memory
Reasoning
Abstract thought
Hearing
Personality
Learning
Follow us @GraknLabs
Human Capabilities
Long-term memory
Reasoning
Learning
Follow us @GraknLabs
Human Capabilities
Long-term memory
Reasoning
Learning
Follow us @GraknLabs
Learning
Follow us @GraknLabs
Learning
Follow us @GraknLabs
Human Capabilities
Long-term memory
Reasoning
Learning
Follow us @GraknLabs
Example: CITES Animal Trade Dataset
Convention on International Trade in
Endangered Species of Wild Fauna and Flora
• Lists exchanges (imports & exports) of items
between countries
• Forms a sparse graph of interconnected data
Scenario: We’re looking for suspicious trade to
enforce protection of the most endangered
animals
Follow us @GraknLabs
?
Reasoning
Not popular
This allows us to infer
additional information
Species: Balaenoptera
acutorostrata
Genus: Balaenoptera
Family: Balaenopteridae How can we get machines to do
reasoning for us?
Symbolic Automation
Requires structured data
storage
taxonomic-ranking sub rule,
when {
(super-taxon: $grandparent, sub-taxon: $parent) isa taxonomic-hierarchy;
(super-taxon: $parent, sub-taxon: $child) isa taxonomic-hierarchy;
}, then {
(super-taxon: $grandparent, sub-taxon: $child) isa taxonomic-hierarchy;
};
$grandparent
$parent
$child
t.-hierarchy
t.-hierarchy
t.-hierarchy
Follow us @GraknLabs
Human Capabilities
Long-term memory
Reasoning
Follow us @GraknLabs
Long-term Memory
Sounds like a database,
but with crazy complexity
But how do we humans store memories?
More of a distributed web - a graph
Hard to search without governing structure
(how many times did you forget what you went there for?)
Follow us @GraknLabs
Structured Long-term Memory
We can define a structure for our data
This way, it’s:
• Easy to search
• Possible to automate reasoning
Model: Entity, Relation, Attribute (nodes)
Roles (edges)
Demo:
match $s isa taxonomic-species, has name $n;
$n "Balaenoptera acutorostrata";
$f isa taxonomic-family;
$h($s, $f) isa taxonomic-hierarchy;
get;
Follow us @GraknLabs
Human Capabilities
Long-term memory
Reasoning
Learning
Follow us @GraknLabs
High-Level Architecture
Learner
Reasoner
Knowledge Graph
Learning
Reasoning
Long-term Memory
A learner talks to a knowledge graph
via a reasoner
Follow us @GraknLabs
High-Level Architecture
KGCN
Reasoner
Knowledge Graph
Graql
Learner
Grakn
Follow us @GraknLabs
Query for a batch of concepts
Learning
Task
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
Embedding for a single concept
Use as the Knowledge-set for
training/testing/prediction
KGCNmatch $x isa traded-item; get;
KGCNs - Objective
Create a bridge to ML
Follow us @GraknLabs
Node Connection Inferred Connection
.
.
.
.
.
.
.
.
.
.
.
.
Knowledge Graph
with information for
each node .
.
.
.
.
.
.
.
.
.
.
.
Aggregate
.
.
.
.
.
.
Information at Node
Combine .
.
.
.One vector
represents
whole context
KGCNs – How do they work?
Follow us @GraknLabs
.
.
.
.
.
.
.
.
.
.
..
.
.
.
.
.
.
.
.
.
.
.
Chained
result
KGCNs
Follow us @GraknLabs
.
.
.
.
.
.
.
.
.
.
.
.
Maxpool
Corresponding outputs
KGCNs: Aggregation
Representations of
connected current Nodes
Follow us @GraknLabs
Concat
KGCNs: Combination
Aggregated
context
representation
Representation
of current Node
Combined
representation
of current Node
and its context
.
.
.
.
Follow us @GraknLabs
Supervised
Learner
Application
KGCN
Grakn
Query for
example Things
Retrieve
example Things
Get input data Call train()
Build feed dicts
Query for
neighbours
Build context
arrays
Find neighbour
Things
Run training
loop
Generate
embeddings
Calculate
loss
Optimise
weights
Make
predictions
Report Loss
< max_steps
> max_steps
KGCNs
Follow us @GraknLabs
2-hop arrays 1-hop arrays
Aggregate 2-hop
Combine with 1-hop
Aggregate 1-hop
Combine with example
Example (0-hop) arrays
Embeddings
labels
Dense Layer
Class predictions 0 0 1
Softmax Cross Entropy
Loss
KGCNs
Follow us @GraknLabs
Example: CITES Animal Trade Dataset
Convention on International Trade in
Endangered Species of Wild Fauna and Flora
• Lists exchanges (imports & exports) of items
between countries
• Forms a sparse graph of interconnected data
Scenario: We’re looking for suspicious trade to
enforce protection of the most endangered
animals
Follow us @GraknLabs
Predicting endangerment-level associated with traded-items
20 9 1
7 22 1
5 3 22
Confusion
Matrix
Actual
Class
Predicted Class
1
2
3
1 2 3
Follow us @GraknLabs
Ingesting the results
Follow us @GraknLabs
Hunting down suspicious activity
Follow us @GraknLabs
Machine Learning over Reasoned Knowledge
Long-term memory
Reasoning
Learning
Follow us @GraknLabs
What does Grakn do?
Knowledge schema
Flexible Entity-Relationship
concept-level schema to
build knowledge models
Model complex
domains
Automated Reasoning
Automated deductive
reasoning of data points
during runtime (OLTP)
Derive implicit facts &
simplification
Distributed Analytics
Automated distributed
algorithms (BSP) as a
language (OLAP)
Automated large scale
analytics
Higher-Level Language
Strong abstraction over low-
level constructs and
complex relationships
Easier to work with
complex data
Follow us @GraknLabs
What does KGCN do?
Data Agnostic
Generic method with simple
network architecture; use
directly or customise
Easy to plug-and-play for
immediate results
Inductive Approach
Only concepts of interest are
considered, not the whole
graph
Scales up for large graphs
Downstream Learning
Use embeddings for your
own learning pipeline via
sup/unsup training
Leverage KGCNs to fulfil your
own objectives
Combines human skills
Performs learning and
reasoning over a hypergraph
A step change in cognitive
system capabilities
Follow us @GraknLabs
Pre-release out now!
kglib repo
Twitter: @jmsfltchr
Slack: grakn.ai/slack
Follow us @GraknLabs
Selfie Time
Knowledge Graph
Convolutional Networks
James Fletcher, Principal Researcher
Follow us @GraknLabs
Full Architecture
Reasoner
Knowledge Graph
Graql
Grakn
KGCN
Interface
TensorFlow Model
Neighbour Trees To Arrays
Grakn Python
Client
Neighbour Traverser
Application
Grakn Python
Client
Data Import
match $x isa animal-item; get; concepts
performance results / predictions
Follow us @GraknLabs
Full Architecture
Reasoner
Knowledge Graph
Graql
Grakn
KGCN
Interface
TensorFlow Model
Neighbour Trees To Arrays
Grakn Python
Client
Neighbour Traverser
Application
Grakn Python
Client
Data Import
match $x isa animal-item; get; concepts
performance results / predictions

More Related Content

What's hot

Cheat sheets for data scientists
Cheat sheets for data scientistsCheat sheets for data scientists
Cheat sheets for data scientists
Ajay Ohri
 
Data Science With Python | Python For Data Science | Python Data Science Cour...
Data Science With Python | Python For Data Science | Python Data Science Cour...Data Science With Python | Python For Data Science | Python Data Science Cour...
Data Science With Python | Python For Data Science | Python Data Science Cour...
Simplilearn
 
Final Presentation
Final PresentationFinal Presentation
Final Presentation
Love Tyagi
 

What's hot (20)

GRAKN.AI: The Hyper-Relational Database for Knowledge-Oriented Systems
GRAKN.AI: The Hyper-Relational Database for Knowledge-Oriented SystemsGRAKN.AI: The Hyper-Relational Database for Knowledge-Oriented Systems
GRAKN.AI: The Hyper-Relational Database for Knowledge-Oriented Systems
 
Big Data Analytics with Grakn
Big Data Analytics with GraknBig Data Analytics with Grakn
Big Data Analytics with Grakn
 
Knowledge base completion presentation
Knowledge base completion presentationKnowledge base completion presentation
Knowledge base completion presentation
 
Logical Inference in a Hyper-Relational Database
Logical Inference in a Hyper-Relational DatabaseLogical Inference in a Hyper-Relational Database
Logical Inference in a Hyper-Relational Database
 
Introduction to Knowledge Graphs with Grakn and Graql
Introduction to Knowledge Graphs with Grakn and Graql Introduction to Knowledge Graphs with Grakn and Graql
Introduction to Knowledge Graphs with Grakn and Graql
 
Natural Language Processing with Neo4j
Natural Language Processing with Neo4jNatural Language Processing with Neo4j
Natural Language Processing with Neo4j
 
Natural language search using Neo4j
Natural language search using Neo4jNatural language search using Neo4j
Natural language search using Neo4j
 
The Art of Social Media Analysis with Twitter & Python
The Art of Social Media Analysis with Twitter & PythonThe Art of Social Media Analysis with Twitter & Python
The Art of Social Media Analysis with Twitter & Python
 
The Hitchhiker's Guide to Machine Learning with Python & Apache Spark
The Hitchhiker's Guide to Machine Learning with Python & Apache SparkThe Hitchhiker's Guide to Machine Learning with Python & Apache Spark
The Hitchhiker's Guide to Machine Learning with Python & Apache Spark
 
Cheat sheets for data scientists
Cheat sheets for data scientistsCheat sheets for data scientists
Cheat sheets for data scientists
 
A Blended Approach to Analytics at Data Tactics Corporation
A Blended Approach to Analytics at Data Tactics CorporationA Blended Approach to Analytics at Data Tactics Corporation
A Blended Approach to Analytics at Data Tactics Corporation
 
Introduction To Data Science With Python
Introduction To Data Science With PythonIntroduction To Data Science With Python
Introduction To Data Science With Python
 
Data Science With Python | Python For Data Science | Python Data Science Cour...
Data Science With Python | Python For Data Science | Python Data Science Cour...Data Science With Python | Python For Data Science | Python Data Science Cour...
Data Science With Python | Python For Data Science | Python Data Science Cour...
 
Graph Networks for Object Recognition
Graph Networks for Object RecognitionGraph Networks for Object Recognition
Graph Networks for Object Recognition
 
Basic Sentiment Analysis using Hive
Basic Sentiment Analysis using HiveBasic Sentiment Analysis using Hive
Basic Sentiment Analysis using Hive
 
The Art of Social Media Analysis with Twitter & Python-OSCON 2012
The Art of Social Media Analysis with Twitter & Python-OSCON 2012The Art of Social Media Analysis with Twitter & Python-OSCON 2012
The Art of Social Media Analysis with Twitter & Python-OSCON 2012
 
Final Presentation
Final PresentationFinal Presentation
Final Presentation
 
Fasta
FastaFasta
Fasta
 
Does Google still need links? - SearchLove San Diego 2017
Does Google still need links? - SearchLove San Diego 2017Does Google still need links? - SearchLove San Diego 2017
Does Google still need links? - SearchLove San Diego 2017
 
Fasta
FastaFasta
Fasta
 

Similar to Knowledge graph convolutional networks - Berlin 2019

Cassandra summit-2013
Cassandra summit-2013Cassandra summit-2013
Cassandra summit-2013
dfilppi
 

Similar to Knowledge graph convolutional networks - Berlin 2019 (20)

Deutsche Telecom Expert System - Router Troubleshooting
Deutsche Telecom Expert System - Router TroubleshootingDeutsche Telecom Expert System - Router Troubleshooting
Deutsche Telecom Expert System - Router Troubleshooting
 
Opportunities for X-Ray science in future computing architectures
Opportunities for X-Ray science in future computing architecturesOpportunities for X-Ray science in future computing architectures
Opportunities for X-Ray science in future computing architectures
 
Writing Machines: Detection and Stylometric Profiling
Writing Machines: Detection and Stylometric ProfilingWriting Machines: Detection and Stylometric Profiling
Writing Machines: Detection and Stylometric Profiling
 
Ehtsham Elahi, Senior Research Engineer, Personalization Science and Engineer...
Ehtsham Elahi, Senior Research Engineer, Personalization Science and Engineer...Ehtsham Elahi, Senior Research Engineer, Personalization Science and Engineer...
Ehtsham Elahi, Senior Research Engineer, Personalization Science and Engineer...
 
DITA's New Thang: Going Mapless!
DITA's New Thang: Going Mapless!DITA's New Thang: Going Mapless!
DITA's New Thang: Going Mapless!
 
EgoSystem: Presentation to LITA, American Library Association, Nov 8 2014
EgoSystem: Presentation to LITA, American Library Association, Nov 8 2014EgoSystem: Presentation to LITA, American Library Association, Nov 8 2014
EgoSystem: Presentation to LITA, American Library Association, Nov 8 2014
 
eScience: A Transformed Scientific Method
eScience: A Transformed Scientific MethodeScience: A Transformed Scientific Method
eScience: A Transformed Scientific Method
 
Cassandra summit-2013
Cassandra summit-2013Cassandra summit-2013
Cassandra summit-2013
 
Cloud com foster december 2010
Cloud com foster december 2010Cloud com foster december 2010
Cloud com foster december 2010
 
Computing Outside The Box June 2009
Computing Outside The Box June 2009Computing Outside The Box June 2009
Computing Outside The Box June 2009
 
Multiplatform Spark solution for Graph datasources by Javier Dominguez
Multiplatform Spark solution for Graph datasources by Javier DominguezMultiplatform Spark solution for Graph datasources by Javier Dominguez
Multiplatform Spark solution for Graph datasources by Javier Dominguez
 
How Graph Databases used in Police Department?
How Graph Databases used in Police Department?How Graph Databases used in Police Department?
How Graph Databases used in Police Department?
 
Computation and Knowledge
Computation and KnowledgeComputation and Knowledge
Computation and Knowledge
 
Useful Techniques in Artificial Intelligence
Useful Techniques in Artificial IntelligenceUseful Techniques in Artificial Intelligence
Useful Techniques in Artificial Intelligence
 
AI/ML session by GDSC ZHCET AMU, ALIGARH
AI/ML session by GDSC ZHCET AMU, ALIGARHAI/ML session by GDSC ZHCET AMU, ALIGARH
AI/ML session by GDSC ZHCET AMU, ALIGARH
 
Big Data Tutorial V4
Big Data Tutorial V4Big Data Tutorial V4
Big Data Tutorial V4
 
Spark Meetup @ Netflix, 05/19/2015
Spark Meetup @ Netflix, 05/19/2015Spark Meetup @ Netflix, 05/19/2015
Spark Meetup @ Netflix, 05/19/2015
 
Scaling PyData Up and Out
Scaling PyData Up and OutScaling PyData Up and Out
Scaling PyData Up and Out
 
Deep Content Learning in Traffic Prediction and Text Classification
Deep Content Learning in Traffic Prediction and Text ClassificationDeep Content Learning in Traffic Prediction and Text Classification
Deep Content Learning in Traffic Prediction and Text Classification
 
Visualization of 3D Genome Data
Visualization of 3D Genome DataVisualization of 3D Genome Data
Visualization of 3D Genome Data
 

More from Vaticle

Building Biomedical Knowledge Graphs for In-Silico Drug Discovery
Building Biomedical Knowledge Graphs for In-Silico Drug DiscoveryBuilding Biomedical Knowledge Graphs for In-Silico Drug Discovery
Building Biomedical Knowledge Graphs for In-Silico Drug Discovery
Vaticle
 
A Data Modelling Framework to Unify Cyber Security Knowledge
A Data Modelling Framework to Unify Cyber Security KnowledgeA Data Modelling Framework to Unify Cyber Security Knowledge
A Data Modelling Framework to Unify Cyber Security Knowledge
Vaticle
 
Unifying Space Mission Knowledge with NLP & Knowledge Graph
Unifying Space Mission Knowledge with NLP & Knowledge GraphUnifying Space Mission Knowledge with NLP & Knowledge Graph
Unifying Space Mission Knowledge with NLP & Knowledge Graph
Vaticle
 
Knowledge Graphs for Supply Chain Operations.pdf
Knowledge Graphs for Supply Chain Operations.pdfKnowledge Graphs for Supply Chain Operations.pdf
Knowledge Graphs for Supply Chain Operations.pdf
Vaticle
 
TypeDB Academy | Modelling Principles
TypeDB Academy | Modelling PrinciplesTypeDB Academy | Modelling Principles
TypeDB Academy | Modelling Principles
Vaticle
 
Intro to TypeDB and TypeQL | A strongly-typed database
Intro to TypeDB and TypeQL | A strongly-typed databaseIntro to TypeDB and TypeQL | A strongly-typed database
Intro to TypeDB and TypeQL | A strongly-typed database
Vaticle
 
Graph Databases vs TypeDB | What you can't do with graphs
Graph Databases vs TypeDB | What you can't do with graphsGraph Databases vs TypeDB | What you can't do with graphs
Graph Databases vs TypeDB | What you can't do with graphs
Vaticle
 

More from Vaticle (20)

Building Biomedical Knowledge Graphs for In-Silico Drug Discovery
Building Biomedical Knowledge Graphs for In-Silico Drug DiscoveryBuilding Biomedical Knowledge Graphs for In-Silico Drug Discovery
Building Biomedical Knowledge Graphs for In-Silico Drug Discovery
 
Loading Huge Amounts of Data
Loading Huge Amounts of DataLoading Huge Amounts of Data
Loading Huge Amounts of Data
 
Natural Language Interface to Knowledge Graph
Natural Language Interface to Knowledge GraphNatural Language Interface to Knowledge Graph
Natural Language Interface to Knowledge Graph
 
A Data Modelling Framework to Unify Cyber Security Knowledge
A Data Modelling Framework to Unify Cyber Security KnowledgeA Data Modelling Framework to Unify Cyber Security Knowledge
A Data Modelling Framework to Unify Cyber Security Knowledge
 
Unifying Space Mission Knowledge with NLP & Knowledge Graph
Unifying Space Mission Knowledge with NLP & Knowledge GraphUnifying Space Mission Knowledge with NLP & Knowledge Graph
Unifying Space Mission Knowledge with NLP & Knowledge Graph
 
The Next Big Thing in AI - Causality
The Next Big Thing in AI - CausalityThe Next Big Thing in AI - Causality
The Next Big Thing in AI - Causality
 
Building a Cyber Threat Intelligence Knowledge Graph
Building a Cyber Threat Intelligence Knowledge GraphBuilding a Cyber Threat Intelligence Knowledge Graph
Building a Cyber Threat Intelligence Knowledge Graph
 
Knowledge Graphs for Supply Chain Operations.pdf
Knowledge Graphs for Supply Chain Operations.pdfKnowledge Graphs for Supply Chain Operations.pdf
Knowledge Graphs for Supply Chain Operations.pdf
 
Building a Distributed Database with Raft.pdf
Building a Distributed Database with Raft.pdfBuilding a Distributed Database with Raft.pdf
Building a Distributed Database with Raft.pdf
 
Enabling the Computational Future of Biology.pdf
Enabling the Computational Future of Biology.pdfEnabling the Computational Future of Biology.pdf
Enabling the Computational Future of Biology.pdf
 
TypeDB Academy | Inference with Rules
TypeDB Academy | Inference with RulesTypeDB Academy | Inference with Rules
TypeDB Academy | Inference with Rules
 
TypeDB Academy | Modelling Principles
TypeDB Academy | Modelling PrinciplesTypeDB Academy | Modelling Principles
TypeDB Academy | Modelling Principles
 
Beyond SQL - Comparing SQL to TypeQL
Beyond SQL - Comparing SQL to TypeQLBeyond SQL - Comparing SQL to TypeQL
Beyond SQL - Comparing SQL to TypeQL
 
TypeDB Academy- Getting Started with Schema Design
TypeDB Academy- Getting Started with Schema DesignTypeDB Academy- Getting Started with Schema Design
TypeDB Academy- Getting Started with Schema Design
 
Comparing Semantic Web Technologies to TypeDB
Comparing Semantic Web Technologies to TypeDBComparing Semantic Web Technologies to TypeDB
Comparing Semantic Web Technologies to TypeDB
 
Reasoner, Meet Actors | TypeDB's Native Reasoning Engine
Reasoner, Meet Actors | TypeDB's Native Reasoning EngineReasoner, Meet Actors | TypeDB's Native Reasoning Engine
Reasoner, Meet Actors | TypeDB's Native Reasoning Engine
 
Intro to TypeDB and TypeQL | A strongly-typed database
Intro to TypeDB and TypeQL | A strongly-typed databaseIntro to TypeDB and TypeQL | A strongly-typed database
Intro to TypeDB and TypeQL | A strongly-typed database
 
Graph Databases vs TypeDB | What you can't do with graphs
Graph Databases vs TypeDB | What you can't do with graphsGraph Databases vs TypeDB | What you can't do with graphs
Graph Databases vs TypeDB | What you can't do with graphs
 
Pandora Paper Leaks With TypeDB
 Pandora Paper Leaks With TypeDB Pandora Paper Leaks With TypeDB
Pandora Paper Leaks With TypeDB
 
Strongly Typed Data for Machine Learning
Strongly Typed Data for Machine LearningStrongly Typed Data for Machine Learning
Strongly Typed Data for Machine Learning
 

Recently uploaded

Essentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with ParametersEssentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with Parameters
Safe Software
 

Recently uploaded (20)

Key Trends Shaping the Future of Infrastructure.pdf
Key Trends Shaping the Future of Infrastructure.pdfKey Trends Shaping the Future of Infrastructure.pdf
Key Trends Shaping the Future of Infrastructure.pdf
 
Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...
 
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...
 
Knowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and backKnowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and back
 
FIDO Alliance Osaka Seminar: Passkeys and the Road Ahead.pdf
FIDO Alliance Osaka Seminar: Passkeys and the Road Ahead.pdfFIDO Alliance Osaka Seminar: Passkeys and the Road Ahead.pdf
FIDO Alliance Osaka Seminar: Passkeys and the Road Ahead.pdf
 
ODC, Data Fabric and Architecture User Group
ODC, Data Fabric and Architecture User GroupODC, Data Fabric and Architecture User Group
ODC, Data Fabric and Architecture User Group
 
"Impact of front-end architecture on development cost", Viktor Turskyi
"Impact of front-end architecture on development cost", Viktor Turskyi"Impact of front-end architecture on development cost", Viktor Turskyi
"Impact of front-end architecture on development cost", Viktor Turskyi
 
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdfSmart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
 
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualitySoftware Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
 
Essentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with ParametersEssentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with Parameters
 
FIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdfFIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdf
 
Demystifying gRPC in .Net by John Staveley
Demystifying gRPC in .Net by John StaveleyDemystifying gRPC in .Net by John Staveley
Demystifying gRPC in .Net by John Staveley
 
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
 
Empowering NextGen Mobility via Large Action Model Infrastructure (LAMI): pav...
Empowering NextGen Mobility via Large Action Model Infrastructure (LAMI): pav...Empowering NextGen Mobility via Large Action Model Infrastructure (LAMI): pav...
Empowering NextGen Mobility via Large Action Model Infrastructure (LAMI): pav...
 
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
 
Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........
 
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
 
To Graph or Not to Graph Knowledge Graph Architectures and LLMs
To Graph or Not to Graph Knowledge Graph Architectures and LLMsTo Graph or Not to Graph Knowledge Graph Architectures and LLMs
To Graph or Not to Graph Knowledge Graph Architectures and LLMs
 
JMeter webinar - integration with InfluxDB and Grafana
JMeter webinar - integration with InfluxDB and GrafanaJMeter webinar - integration with InfluxDB and Grafana
JMeter webinar - integration with InfluxDB and Grafana
 
IoT Analytics Company Presentation May 2024
IoT Analytics Company Presentation May 2024IoT Analytics Company Presentation May 2024
IoT Analytics Company Presentation May 2024
 

Knowledge graph convolutional networks - Berlin 2019

  • 1. Knowledge Graph Convolutional Networks James Fletcher, Principal Researcher
  • 2. Follow us @GraknLabs Machines aren’t as smart as we’d like Highly domain constrained, generalisation and model-wise Reliant on big data Require complex learning models Only use flat arrays of input data Don’t exploit data context
  • 3. Follow us @GraknLabs Human Capabilities Long-term memory Basic body function Basic emotions Appropriate social response Skilled movement coordination Language Vision Sensory processing Planning Motivation Motor control Working memory Reasoning Abstract thought Hearing Personality Learning
  • 4. Follow us @GraknLabs Human Capabilities Long-term memory Reasoning Learning
  • 5. Follow us @GraknLabs Human Capabilities Long-term memory Reasoning Learning
  • 8. Follow us @GraknLabs Human Capabilities Long-term memory Reasoning Learning
  • 9. Follow us @GraknLabs Example: CITES Animal Trade Dataset Convention on International Trade in Endangered Species of Wild Fauna and Flora • Lists exchanges (imports & exports) of items between countries • Forms a sparse graph of interconnected data Scenario: We’re looking for suspicious trade to enforce protection of the most endangered animals
  • 10. Follow us @GraknLabs ? Reasoning Not popular This allows us to infer additional information Species: Balaenoptera acutorostrata Genus: Balaenoptera Family: Balaenopteridae How can we get machines to do reasoning for us? Symbolic Automation Requires structured data storage taxonomic-ranking sub rule, when { (super-taxon: $grandparent, sub-taxon: $parent) isa taxonomic-hierarchy; (super-taxon: $parent, sub-taxon: $child) isa taxonomic-hierarchy; }, then { (super-taxon: $grandparent, sub-taxon: $child) isa taxonomic-hierarchy; }; $grandparent $parent $child t.-hierarchy t.-hierarchy t.-hierarchy
  • 11. Follow us @GraknLabs Human Capabilities Long-term memory Reasoning
  • 12. Follow us @GraknLabs Long-term Memory Sounds like a database, but with crazy complexity But how do we humans store memories? More of a distributed web - a graph Hard to search without governing structure (how many times did you forget what you went there for?)
  • 13. Follow us @GraknLabs Structured Long-term Memory We can define a structure for our data This way, it’s: • Easy to search • Possible to automate reasoning Model: Entity, Relation, Attribute (nodes) Roles (edges) Demo: match $s isa taxonomic-species, has name $n; $n "Balaenoptera acutorostrata"; $f isa taxonomic-family; $h($s, $f) isa taxonomic-hierarchy; get;
  • 14. Follow us @GraknLabs Human Capabilities Long-term memory Reasoning Learning
  • 15. Follow us @GraknLabs High-Level Architecture Learner Reasoner Knowledge Graph Learning Reasoning Long-term Memory A learner talks to a knowledge graph via a reasoner
  • 16. Follow us @GraknLabs High-Level Architecture KGCN Reasoner Knowledge Graph Graql Learner Grakn
  • 17. Follow us @GraknLabs Query for a batch of concepts Learning Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Embedding for a single concept Use as the Knowledge-set for training/testing/prediction KGCNmatch $x isa traded-item; get; KGCNs - Objective Create a bridge to ML
  • 18. Follow us @GraknLabs Node Connection Inferred Connection . . . . . . . . . . . . Knowledge Graph with information for each node . . . . . . . . . . . . Aggregate . . . . . . Information at Node Combine . . . .One vector represents whole context KGCNs – How do they work?
  • 20. Follow us @GraknLabs . . . . . . . . . . . . Maxpool Corresponding outputs KGCNs: Aggregation Representations of connected current Nodes
  • 21. Follow us @GraknLabs Concat KGCNs: Combination Aggregated context representation Representation of current Node Combined representation of current Node and its context . . . .
  • 22. Follow us @GraknLabs Supervised Learner Application KGCN Grakn Query for example Things Retrieve example Things Get input data Call train() Build feed dicts Query for neighbours Build context arrays Find neighbour Things Run training loop Generate embeddings Calculate loss Optimise weights Make predictions Report Loss < max_steps > max_steps KGCNs
  • 23. Follow us @GraknLabs 2-hop arrays 1-hop arrays Aggregate 2-hop Combine with 1-hop Aggregate 1-hop Combine with example Example (0-hop) arrays Embeddings labels Dense Layer Class predictions 0 0 1 Softmax Cross Entropy Loss KGCNs
  • 24. Follow us @GraknLabs Example: CITES Animal Trade Dataset Convention on International Trade in Endangered Species of Wild Fauna and Flora • Lists exchanges (imports & exports) of items between countries • Forms a sparse graph of interconnected data Scenario: We’re looking for suspicious trade to enforce protection of the most endangered animals
  • 25. Follow us @GraknLabs Predicting endangerment-level associated with traded-items 20 9 1 7 22 1 5 3 22 Confusion Matrix Actual Class Predicted Class 1 2 3 1 2 3
  • 27. Follow us @GraknLabs Hunting down suspicious activity
  • 28. Follow us @GraknLabs Machine Learning over Reasoned Knowledge Long-term memory Reasoning Learning
  • 29. Follow us @GraknLabs What does Grakn do? Knowledge schema Flexible Entity-Relationship concept-level schema to build knowledge models Model complex domains Automated Reasoning Automated deductive reasoning of data points during runtime (OLTP) Derive implicit facts & simplification Distributed Analytics Automated distributed algorithms (BSP) as a language (OLAP) Automated large scale analytics Higher-Level Language Strong abstraction over low- level constructs and complex relationships Easier to work with complex data
  • 30. Follow us @GraknLabs What does KGCN do? Data Agnostic Generic method with simple network architecture; use directly or customise Easy to plug-and-play for immediate results Inductive Approach Only concepts of interest are considered, not the whole graph Scales up for large graphs Downstream Learning Use embeddings for your own learning pipeline via sup/unsup training Leverage KGCNs to fulfil your own objectives Combines human skills Performs learning and reasoning over a hypergraph A step change in cognitive system capabilities
  • 31. Follow us @GraknLabs Pre-release out now! kglib repo Twitter: @jmsfltchr Slack: grakn.ai/slack
  • 33. Knowledge Graph Convolutional Networks James Fletcher, Principal Researcher
  • 34. Follow us @GraknLabs Full Architecture Reasoner Knowledge Graph Graql Grakn KGCN Interface TensorFlow Model Neighbour Trees To Arrays Grakn Python Client Neighbour Traverser Application Grakn Python Client Data Import match $x isa animal-item; get; concepts performance results / predictions
  • 35. Follow us @GraknLabs Full Architecture Reasoner Knowledge Graph Graql Grakn KGCN Interface TensorFlow Model Neighbour Trees To Arrays Grakn Python Client Neighbour Traverser Application Grakn Python Client Data Import match $x isa animal-item; get; concepts performance results / predictions