How Graph Technology is Changing AI

Graph enhancements to Artificial Intelligence and Machine Learning are changing the landscape of intelligent applications. Beyond improving accuracy and modeling speed, graph technologies make building AI solutions more accessible. Join us to hear about 4 areas at the forefront of graph enhanced AI and ML, and find out which techniques are commonly used today and which hold the potential for disrupting industries. We’ll provide examples and specifically look how: – Graphs provide better accuracy through connected feature extraction – Graphs provide better performance through contextual model optimization – Graphs provide context through knowledge graphs – Graphs add explainability to neural networks
展开查看详情

1.WIFI SSID:SparkAISummit | Password: UnifiedAnalytics

2.How Graph Technology is Changing AI Jake Graham & Alicia Frame, Neo4j #UnifiedAnalytics #SparkAISummit

3.

4.Where Do Graphs Matter? FinCrime Detection Drug Discovery Recommendations Customer Segmentation Cybersecurity Churn Prediction Predictive Maintenance Search/MDM

5.Labeled Property Graphs name: “Dan” Nodes born: May 29, 1970 name: “Ann” twitter: “@dan” born: Dec 5, 1975 • Can have Labels to classify nodes • Labels have native indexes MARRIED TO Relationships PERSON PERSON LIVES WITH • Relate nodes by type and direction RID Properties NS VE OW • Attributes of Nodes & Relationships since: S Jan 10, 2011 • Stored as Name/Value pairs brand: “Volvo” • Can have indexes and composite indexes CAR model: “V70” Latitude: 37.5629900° Longitude: -122.3255300° 5

6.Graphs provide more accurate predictions With the data you already have o Current data science models ignore network structure and complex relationships o Graph models add highly predictive features to existing ML models MACHINE LEARNING LIBRARY

7. Lest you think the authors think they've got it all figured out, the paper The idea lists someislingering that graph networks are shortcomings. bigger than Battaglia et al.any one pose machine-learning the big question, approach. Graphs bring an ability to generalize about structure that the "Where do the graphsindividual come from that neural graph nets don'tnetworks have. operate over?”

8.Graph Model Building SparkCypher & Neo4j Neo4j Graph SparkGraph Morpheus Platform Cypher 9 in Spark to Cypher 10 over Native Graph create non- Spark for seamless Algorithms, persistent graphs Neo4j integration Processing, and Storage

9. Explore Graphs Build Graphs in in o Massively scalable o Persistent, dynamic graphs o Powerful data pipelining o Graph native query and algorithm o Robust ML Libraries performance o Non-persistent, non-native graphs o Constantly growing list of graph algorithms and embeddings

10.The Steps of Graph Data Science Knowledge Graph Graph Graph Feature Native Engineering Learning Complexity Graph Neural Networks Graph Embeddings Graph Algorithm Query Based Feature Feature Engineering Query Based Engineering Knowledge Graph Delivery Timeline Neo4J for Graph Persistence

11.The Steps of Graph Data Science Knowledge Graph Graph Graph Feature Native Engineering Learning Complexity Graph Neural Networks Graph Embeddings Graph Algorithm Query Based Feature Feature Engineering Query Based Engineering Knowledge Graph Delivery Timeline Neo4J for Graph Persistence

12.Connecting the Dots at NASA “Using Neo4j someone from our Orion project found information from the Apollo project that prevented an issue, saving well over two years of work and one million dollars of taxpayer funds.” David Meza, Chief Knowledge Architect – NASA 2015

13.The Steps of Graph Data Science Knowledge Graph Graph Graph Feature Native Engineering Learning Complexity Graph Neural Networks Graph Embeddings Graph Algorithm Query Based Feature Feature Engineering Query Based Engineering Knowledge Graph Delivery Timeline Neo4J for Graph Persistence

14.Mining Knowledge Graphs for Drug Discovery • HetioNet is a knowledge graph integrating over 50 years of biomedical data • Leveraged to predict new uses for drugs by using the graph topology to create features to predict new links 14

15.Knowledge Graphs - het.io • HetioNet is a knowledge graph integrating over 50 years of biomedical data • Leveraged to predict new uses for drugs by using the graph topology to create features to predict new links 15

16.Knowledge Graphs - het.io • HetioNet is a knowledge graph integrating over 50 years of biomedical data • Leveraged to predict new uses for drugs by using the graph topology to create features to predict new links 16

17. Knowledge Graphs: getting started SparkCypher & Neo4j Neo4j Graph SparkGraph Morpheus Platform Graph Graph Transactions Analytics • Merge distributed data • Build a graph data • Move to Neo4J to build into dataframes pipeline to bring into expert queries and • Reshape your tables native graph persist your graph into graphs • Bring graph features • Explore cypher queries back to ML pipeline 17

18.The Steps of Graph Data Science Knowledge Graph Graph Graph Feature Native Engineering Learning Complexity Graph Neural Networks Graph Embeddings Graph Algorithm Query Based Feature Feature Engineering Query Based Engineering Knowledge Graph Delivery Timeline Neo4J for Graph Persistence

19.Graph Feature Engineering Make use of your existing machine learning pipeline: • Tabular data from Spark • Enriched with graph based features from Neo4j • Combined into a single model building pipeline MACHINE LEARNING LIBRARY 19

20.Categories of Graph Features Community Centrality / Pathfinding Detection Importance & Search Detects group Determines the Finds optimal paths clustering or partition importance of distinct or evaluates route options nodes in the network availability and quality Heuristic Link Prediction Embeddings Similarity Estimates the likelihood of Vectors that capture Evaluates how nodes forming a relationship connectivity or topology alike nodes are 20

21.Financial Crime: Detecting Fraud Many large financial institutions have existing pipelines to identify fraud Graph based features improve accuracy: • Connected components to identify disjoint graphs • PageRank to measure influence • Louvain to identify communities • Jaccard to measure account similarity 21

22.Financial Crime: Detecting Fraud Many large financial institutions have existing pipelines to identify fraud Graph based features improve accuracy: • Connected components to identify disjoint graphs • PageRank to measure influence • Louvain to identify communities • Jaccard to measure account similarity 22

23. Graph Feature Engineering: getting started Graph Graph Transactions Analytics • Merge distributed data • Build a graph data • Move to Neo4J to build into dataframes pipeline to bring into run native graph • Reshape your tables native graph algorithms into graphs • Bring graph features • Write algorithm derived • Explore graph algorithms back to ML pipeline features to persistent graph 23

24. Graph Features in Neo4J Pathfinding Centrality / Community & Search Importance Detection • Parallel Breadth First Search • Degree Centrality • Triangle Count • Parallel Depth First Search • Closeness Centrality • Clustering Coefficients • Shortest Path • CC Variations: Harmonic, Dangalchev, • Connected Components (Union Find) • Single-Source Shortest Path Wasserman & Faust • Strongly Connected Components • All Pairs Shortest Path • Betweenness Centrality • Label Propagation • Minimum Spanning Tree • Approximate Betweenness Centrality • Louvain Modularity – 1 Step & Multi-Step • A* Shortest Path • PageRank • Balanced Triad (identification) • Yen’s K Shortest Path • Personalized PageRank • K-Spanning Tree (MST) • ArticleRank • Random Walk • Eigenvector Centrality Link Similarity Prediction • Euclidean Distance • Adamic Adar • Cosine Similarity • Common Neighbors • Jaccard Similarity • Preferential Attachment neo4j.com/docs/ • Overlap Similarity • Resource Allocations • Same Community graph-algorithms/current/ • Pearson Similarity • Total Neighbors 24

25.The Steps of Graph Data Science Knowledge Graph Graph Graph Feature Native Engineering Learning Complexity Graph Neural Networks Graph Embeddings Graph Algorithm Query Based Feature Feature Engineering Query Based Engineering Knowledge Graph Delivery Timeline Neo4J for Graph Persistence

26.Graph Embeddings Embeddings transform graphs into a vector, or set of vectors, describing topology, connectivity, or attributes of nodes and edges in the graph • Vertex embeddings: describe connectivity of each node • Path embeddings: traversals across the graph • Graph embeddings: encode an entire graph into a single vector 26

27.Graph Embeddings - Recommendations Explainable Reasoning over Knowledge Graphs for Recommendation 27

28.Graph Embeddings - Recommendations Explainable Reasoning over Knowledge Graphs for Recommendation 28

29. Graph Embeddings: Getting Started Graph Graph Transactions Analytics • Merge distributed data • Build a graph data • Move to Neo4J to build into dataframes pipeline to bring into expert queries and • Reshape your tables native graph persist into graphs • Bring graph features • Stay tuned for DeepWalk • Explore graph algorithms back to ML pipeline and DeepGL 29