Knowledge Graphs and LLMs in Action you own this product

Build AI systems using connected data

Alessandro Negro with Vlastimil Kus, Giuseppe Futia and Fabio Montagna
Forewords by Maxime Labonne, Khalifeh AlJadda

October 2025
ISBN 9781633439894
472 pages

Included with a Manning Online subscription

printed in black & white

available in Polish, Russian, Simplified Chinese

catalog / Data Science / Machine Learning / Knowledge Graphs

resources: Source code Supplememental material Book forum Source code on Github Register your pBook for a free eBook

table of content

Part 1 Foundations of hybrid intelligent systems

1 Knowledge graphs and LLMs: A killer combination

1.1 Knowledge graphs

1.2 Large language models

1.3 KGs and LLMs: Stronger together

1.4 The paradigm shift in data-driven applications

1.4.1 The four pillars of knowledge graphs

1.5 Building data-driven applications using KGs and LLMs

1.5.1 Example use case: Drug discovery and development

1.5.2 Example use case: Conversational AI for customer support

1.5.3 Deciding whether to use a KG

1.6 Knowledge graph technologies

1.6.1 Taxonomies and ontologies

1.7 How do we teach KGs and LLMs?

2 Intelligent systems: A hybrid approach

2.1 What is intelligence?

2.2 Designing an intelligent system

2.2.1 What is an intelligent system?

2.2.2 Categories of intelligent systems

2.2.3 Characteristics of an intelligent system

2.3 Knowledge acquisition and representation

2.4 Reasoning

2.5 Reasoning engines

2.5.1 Limitations of a pure deductive reasoning engine

2.5.2 Using inductive reasoning and ML

2.5.3 The role of LLMs in the reasoning engine

2.6 A KG approach to IASs

Part 2 Building knowledge graphs from structured data sources

3 Create your first knowledge graph from ontologies

3.1 Knowledge graph building: Warmup

3.1.1 Business and domain understanding

3.1.2 Data understanding

3.2 Understanding knowledge graph technologies

3.2.1 RDF or LPG? A goal-driven discussion

3.2.2 Representing edge properties with RDF and LPG

3.3 Building a knowledge graph

3.3.1 Ontology ingestion and processing with neosemantics

3.3.2 Annotation ingestion and processing

3.4 Querying the data

3.5 Reasoning over the KG

4 From simple networks to multisource integration

4.1 Biomedical knowledge graphs and applications

4.2 Multi-omic applications of KGs

4.2.1 Creating a KG from the PPI and protein-disease networks

4.2.2 High-level analysis of the resulting KGs

4.2.3 Domain-specific analysis of the PPI and disease KG

4.3 Pharmaceutical applications of KGs

4.3.1 Deep analysis of the Hetionet knowledge graph

4.3.2 LLM-assisted interpretation of pathway analysis results

4.4 Clinical applications of KGs

4.4.1 LLM-guided clinical decision support analysis

Part 3 Building knowledge graphs from text

5 Extracting domain-specific knowledge from unstructured data

5.1 The archives challenge

5.2 Key concepts of knowledge extraction

5.2.1 Recognizing named entities

5.2.2 Extracting relations

5.3 Building KGs with large language models

5.3.1 Using LLMs

5.3.2 Prompt engineering examples

5.3.3 Prompt engineering guidelines

5.3.4 KG building: Traditional NLP or LLMs?

6 Building knowledge graphs with large language models

6.1 Transforming an archive to a KG

6.1.1 Graph modeling

6.1.2 Creating a metagraph

6.1.3 Normalization and cleansing

6.1.4 Graph-based entity resolution

6.2 Intellectual network analysis: The value of graphs

6.3 Next steps in the Rockefeller Archive Center project

6.4 The value of knowledge graphs in the LLM era

7 Named entity disambiguation

7.1 From recognition to disambiguation

7.2 Understanding named entity disambiguation

7.3 Domain-based NED and LLMs

7.4 Business and domain understanding

7.4.1 Context

7.4.2 Use case definition

7.5 Understanding the data

7.5.1 Unstructured data

7.5.2 Domain ontologies

7.6 Building a SoHO knowledge graph

7.6.1 Defining the schema

7.6.2 Processing and ingesting documents

7.6.3 Disambiguating and ingesting medical entities

7.6.4 Processing, loading, and mapping ontologies

7.6.5 Generating entity co-occurrences

7.7 KG-based use cases

7.7.1 Conceptual search

7.7.2 Structured knowledge-based search

7.7.3 KG-based interpretability and discovery

7.7.4 Uncovering new knowledge

8 NED with open LLMs and domain ontologies

8.1 Understanding limitations of traditional NED systems

8.2 Ingesting the domain ontology

8.3 Setting up the model with Ollama and Llama 3.1 8B

8.4 End-to-end NED process

8.4.1 Named entity recognition

8.4.2 Candidate selection

8.4.3 Candidate disambiguation

8.5 Conclusions

Part 4 Machine learning on knowledge graphs

9 Machine learning on knowledge graphs: A primer approach

9.1 Machine learning on graphs: Why?

9.2 Machine learning on graphs: What?

9.2.1 Node classification

9.2.2 Link prediction (a.k.a. relationship prediction)

9.2.3 Clustering and community detection

9.2.4 Graph classification

9.3 Machine learning on graphs: How?

9.3.1 Node classification and link prediction

9.3.2 Graph classification

9.3.3 Graph clustering

10 Graph feature engineering: Manual and semiautomated approaches

10.1 Manual node features

10.1.1 Degree

10.1.2 Triangles

10.1.3 Density

10.1.4 Geodesic (or shortest) path

10.1.5 Closeness

10.1.6 Betweenness

10.1.7 PageRank

10.1.8 Prediction

10.2 Manual relationship features

10.2.1 Node-based representation

10.2.2 Path-based features

10.3 Semiautomated feature extraction

10.3.1 Performing ReFeX manually

10.3.2 Performing ReFeX automatically with code

11 Graph representation learning and graph neural networks

11.1 Embeddings in graph representation learning

11.1.1 Understanding graph embeddings: From discrete to continuous

11.1.2 Real-world applications and examples

11.2 The encoder–decoder model

11.2.1 The encoder: Converting graph structure to vectors

11.2.2 The decoder: Reconstructing graph properties

11.2.3 The power of the framework

11.2.4 Node2Vec: An example of an encoder–decoder framework

11.3 Shallow embeddings: A first approach to graph representation

11.3.1 Understanding shallow embeddings

11.3.2 Limitations of shallow embeddings

11.4 Embeddings in knowledge graphs

11.4.1 Loss function

11.4.2 Multirelationship decoder

11.5 Message passing and graph neural networks

11.5.1 The message-passing framework: A neural conversation

11.5.2 Motivation and intuition: Why message passing works

11.5.3 The basic GNN model

11.5.4 Message passing with self-loops

11.6 Generalized aggregation and update methods

11.6.1 Neighborhood normalization

11.6.2 Neighborhood attention

11.6.3 Multihead attention and transformer connections

11.6.4 Generalized update methods

11.7 The synergy of GNNs and LLMs

12 Node classification and link prediction with GNNs

12.1 Node classification for anti-money laundering applications

12.1.1 Input data

12.1.2 Graph processor: Data preparation

12.1.3 Graph processor: Homogeneous PyG graph

12.1.4 Encoder–decoder architecture

12.1.5 Evaluation and analysis

12.2 Link prediction for movie recommendations

12.2.1 Input data

12.2.2 Graph processor: Data preparation

12.2.3 Graph processor: Heterogeneous PyG graph

12.2.4 Encoder–decoder architecture

12.2.5 Evaluation and analysis

Part 5 Information retrieval with knowledge graphs and LLMs

13 Knowledge graph–powered retrieval-augmented generation

13.1 AI agents

13.2 Chatting with the LLM

13.3 Challenges in the production environment

13.4 Chatting with the AI about private data

13.4.1 Retrieval-augmented generation

13.4.2 Vector-based RAG limitations

13.4.3 Graph RAG

13.4.4 Reasoning agents

13.4.5 Let’s chat with our KG

14 Asking a KG questions with natural language

14.1 Querying a knowledge graph in the policing domain

14.1.1 Enabling domain experts with knowledge graphs

14.2 RAG for KG querying: Capabilities and challenges

14.2.1 RAG effectiveness with complete context

14.2.2 RAG fragility with incomplete retrieval

14.3 Schema-based approach for querying KGs

14.3.1 Understanding and using graph schemas

14.4 Think like an expert: Using metadata for enhanced querying

14.5 Intent detection: Understanding user expectations

14.5.1 Classifying by visualization type

14.5.2 Is it data, documentation, or just complaining?

14.6 From schema to LLM-ready context

14.6.1 Schema extraction and representation

14.6.2 Enriching schemas with descriptive annotations

14.6.3 A practical approach to schema representation

14.7 It’s time to think: Understanding LLM reasoning

14.7.1 The order matters: Answer first vs. reasoning first

14.7.2 Thinking in queries: From text to Cypher

14.7.3 Structuring output for reliable query generation

14.8 Response summarization: From results to insights

15 Building a QA agent with LangGraph

15.1 Building the LangGraph pipeline

15.1.1 System architecture overview

15.1.2 Configuring pipeline components

15.1.3 Schema translation service

15.1.4 State management design

15.1.5 Pipeline agent implementation

15.1.6 Pipeline integration layer

15.2 Streamlit application

15.2.1 Application overview

15.2.2 LangGraph integration

15.3 Expert-emulating investigation

15.3.1 Identifying the initial case

15.3.2 Spatial analysis of surveillance coverage

15.3.3 Vehicle pattern detection

15.3.4 Context-aware request refinement

15.3.5 Historical record analysis

15.4 Future directions and enhancements

15.4.1 Learning from use

15.4.2 Enhancing core capabilities

15.4.3 Advanced evolution paths

Appendixes

Appendix A: Introduction to graphs

A.1 What is a graph?

A.2 Graphs as models of networks

A.3 Representing graphs

Appendix B: Neo4j

B.1 Introduction to Neo4j

B.2 Installing Neo4j

B.2.1 Installing a Neo4j server

B.2.2 Neo4j Desktop installation

B.3 Cypher

B.4 Installing plugins

B.4.1 Installing APOC Core

B.4.2 GDS installation

B.5 Cleaning

Appendix C: Building knowledge graphs from structured sources

C.1 MicroRNA–disease association: Warmup

C.1.1 Key concepts

C.1.2 Business understanding

C.1.3 Data understanding

C.2 Building the miRNA knowledge graph

C.2.1 Importing known miRNA–disease connections

C.2.2 Importing the disease ontology

C.2.3 Importing miRNA information

C.3 Exploring and analyzing the miRNA KG

Appendix D: references

Overview

9 Machine learning on knowledge graphs: a primer approach

Intelligent Advisory Systems need to go beyond querying and manual exploration of knowledge graphs to uncover the rich, implicit structure they encode. This chapter motivates machine learning on knowledge graphs as the natural way to surface insights at scale across domains such as drug discovery, personalized medicine, and law enforcement. It argues that graphs provide a universal data representation, allow many problems to be cast as a small set of graph tasks, and natively capture dependencies that violate the independent-and-identically-distributed assumption of classical ML. It also highlights the growing maturity of graph ML tooling, the complementary strengths of Graph Neural Networks and Large Language Models, and the importance of keeping the discipline problem-driven.

The chapter organizes graph ML work into node-focused and graph-focused tasks. Node classification predicts labels for nodes within a single graph and is often treated as semi-supervised because models exploit the full network structure during training; its effectiveness depends on modeling homophily, heterophily, and structural equivalence rather than assuming i.i.d. data. Link (relationship) prediction addresses missing or future edges in evolving, incomplete graphs and powers use cases from recommendations to knowledge graph completion. Community detection/graph clustering is typically unsupervised, discovers latent group structure from topology alone, and supports summarization and segmentation at scale. Graph classification treats each graph as a data point (e.g., molecules) and learns mappings from whole-graph structure to labels, contrasting with node-level prediction.

Methodologically, there are two routes: use graph-native algorithms that operate directly on structure, or transform graphs into feature vectors and apply standard ML. The chapter emphasizes feature engineering as the linchpin—featurizing nodes, edges, or whole graphs, keeping train/predict pipelines consistent, and applying scaling and robust classifiers. A worked example on the Karate Club shows that naive features (degree) and even generic embeddings (Node2Vec) can be unstable, while simple, domain-aware features that reflect neighborhood composition deliver large, reliable gains. For graph classification, features summarize entire graphs; for clustering, algorithms like Louvain or Label Propagation need no featurization or training. The chapter sets up the remainder of the book to progress from manual feature design to automated representation learning with GNNs, underscoring that final performance hinges on representation quality and careful tuning.

Node-focused and graph-focused tasks are represented in terms of expected input and output during training and prediction.

Node classification typical flow. As with many supervised machine learning tasks, it has two phases. The first is the training and the second is the prediction that, in this case, translates into classifying unclassified nodes.

Link prediction typical flow. Link prediction has two phases. During the first one, the training, a graph is taken and the model is trained to recognize whether a link exists or not. During the prediction phase, the model provides an output for each pair of nodes representing the target relationship to predict. For each, it will provide an existence probability.

Community Detection flow. The input is a graph, and the output is a mapping between nodes and groups. In the image, the groups are represented as colors and different circles.

Graph classification flow. Like many supervised tasks, it has two phases. During the training, it takes different graphs with the related classes and trains a model to recognize the class of each. During the prediction, the model takes a graph and predicts its class.

A graph-based classification approach (collective) versus traditional algorithms for classification.

Training flow for node classification and link prediction. A critical step is the “featurization” of nodes and relationships. Once these steps have been done, the vectors can be passed to a classic algorithm. Both node classification and link prediction can be seen as classification tasks.

Prediction flow for node classification and link prediction. The featurization process should be the same as the training. The classifier model built previously is used in this phase for making the predictions.

Karate Club graph as produced by the code in listing 9.1, the colors of the nodes represent the club the member ultimately joined.

High-level training process for the graph classification task. The output is a classifier model that has been trained on the known classes of different graphs.

High-level prediction process for the graph classification task. The output in this case is the classes assigned to the previous unlabeled graphs.

High-level graph clustering task process. This is an example of a pure graph algorithm approach that doesn’t require an intermediate transformation in a vector representation or other formats.

The result of the community detection algorithm applied to the Karate Club network. The algorithm has identified 4 different communities. All but one contain members of the same groups. Only in one case, there is a spurious element, but it is very well connected to other people belonging to a different group.

Summary

Machine Learning on graphs emerges as a crucial approach because it naturally handles interconnected data, provides universal data representation, and enables the modeling of complex problems through a limited set of computational tasks.
Unlike traditional machine learning that assumes independent and identically distributed (i.i.d.) data points, graph-based approaches explicitly leverage connections and dependencies between nodes, better reflecting real-world relationships and patterns.
Machine learning tasks on graphs fall into two main categories: node-focused (like node classification and link prediction) where tasks are performed on a single graph, and graph-focused (like graph classification) where each graph is a separate data point.
Node classification and link prediction are typically semi-supervised tasks, utilizing both labeled and unlabeled data, along with the graph's structural information. Community detection, in contrast, is usually unsupervised and operates directly on graph structure.
Feature engineering plays a critical role these tasks. Simple approaches (like using node degree) often underperform compared to more sophisticated features that capture domain knowledge (like homophily-based metrics) or structural patterns.
Node embeddings can be generated through various approaches, from manual feature engineering to automated techniques like Node2Vec. However, autonomous embeddings require careful tuning to avoid generating overly homogeneous representations.
The success of graph-based machine learning often depends on finding the right balance between automated feature learning, domain knowledge incorporation, and appropriate algorithm selection for the specific task at hand.

A. Negro, Graph-Powered Machine Learning. Shelter Island, NY, USA: Manning Publications, 2021
A. Mukkamala, P. Szolovits, and J. Lin, "Representation Learning: From Diagnoses to Healthcare Analytics," Data Science, vol. 4, no. 2, pp. 125-146, Sep. 2021. Accessed: May 15, 2024 [Online]. Available: https://content.iospress.com/articles/data-science/ds007
Y. Ma and J. Tang, Deep Learning on Graphs. Cambridge, UK: Cambridge University Press, 2021
W. L. Hamilton, Graph Representation Learning. San Rafael, CA, USA: Morgan & Claypool Publishers, 2020
T. N. Kipf and M. Welling, "Semi-Supervised Classification with Graph Convolutional Networks," in International Conference on Learning Representations (ICLR 2016), San Juan, Puerto Rico, 2016. Accessed: May 15, 2024 [Online]. Available: https://arxiv.org/abs/1609.02907
M. McPherson, L. Smith-Lovin, and J. Cook, "Birds of a Feather: Homophily in Social Networks," Annual Review of Sociology, vol. 27, no. 1, pp. 415-444, 2001
D. Zhou, O. Bousquet, T. Lal, J. Weston, and B. Scholkopf, "Learning with Local and Global Consistency," in Proceedings of the 17th International Conference on Neural Information Processing Systems (NeurIPS 2004), Vancouver, BC, Canada, 2004. Accessed: May 15, 2024 [Online]. Available: https://proceedings.neurips.cc/paper/2003/hash/87682805257e619d49b8e0dfdc14affa-Abstract.html
C. Donnat, M. Zitnik, D. Hallac, and J. Leskovec, "Graph Wavelets for Structural Role Similarity in Complex Networks," in Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2018), London, UK, 2018. Accessed: May 15, 2024 [Online]. Available: https://dl.acm.org/doi/10.1145/3219819.3220025
Z. Yang, W. Cohen, and R. Salakhutdinov, "Revisiting Semi-Supervised Learning with Graph Embeddings," in Proceedings of the 33rd International Conference on Machine Learning (ICML 2016), New York, NY, USA, 2016. Accessed: May 15, 2024 [Online]. Available: https://proceedings.mlr.press/v48/yanga16.html
L. Tang and H. Liu, "Relational Learning via Latent Social Dimensions," in Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2009), Paris, France, pp. 817-826, 2009. Accessed: May 15, 2024 [Online]. Available: https://dl.acm.org/doi/10.1145/1557019.1557109
L. A. Adamic and E. Adar, "Friends and Neighbors on the Web," Social Networks, vol. 25, no. 3, pp. 211-230, 2003
R. Ying, R. He, K. Chen, P. Eksombatchai, W. L. Hamilton, and J. Leskovec, "Graph Convolutional Neural Networks for Web-Scale Recommender Systems," in Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2018), London, UK, pp. 974-983, 2018. Accessed: May 15, 2024 [Online]. Available: https://dl.acm.org/doi/10.1145/3219819.3219890
M. Nickel, K. Murphy, V. Tresp, and E. Gabrilovich, "A Review of Relational Machine Learning for Knowledge Graphs," Proceedings of the IEEE, vol. 104, no. 1, pp. 11-33, 2015. Accessed: May 15, 2024 [Online]. Available: https://ieeexplore.ieee.org/document/7358050
M. Zitnik, M. Agrawal, and J. Leskovec, "Modeling polypharmacy side effects with graph convolutional networks," Bioinformatics, vol. 34, no. 13, pp. i457–i466, 2018.
A. Bordes, N. Usunier, A. Garcia-Duran, J. Weston, and O. Yakhnenko, "Translating Embeddings for Modeling Multi-relational Data," in Proceedings of the 26th International Conference on Neural Information Processing Systems (NeurIPS 2013), Lake Tahoe, NV, USA, 2013. Accessed: May 15, 2024 [Online]. Available: https://papers.nips.cc/paper_files/paper/2013/hash/1cecc7a77928ca8133fa24680a88d2f9-Abstract.html
Wang, H. Huang, C. Ding, and F. Nie, "Predicting Protein–Protein Interactions from Multimodal Biological Data Sources via Nonnegative Matrix Tri-Factorization," Journal of Computational Biology, vol. 20, no. 4, pp. 344-358, 2013. Accessed: May 15, 2024 [Online]. Available: https://doi.org/10.1089/cmb.2012.0273
G. Berlusconi, F. Calderoni, N. Parolini, M. Verani, and C. Piccardi, "Link Prediction in Criminal Networks: A Tool for Criminal Intelligence Analysis," PLOS ONE, vol. 11, no. 4, e0154244, 2016. Accessed: May 15, 2024 [Online]. Available: https://doi.org/10.1371/journal.pone.0154244
M. Agrawal, M. Zitnik, J. Leskovec, "Large-scale analysis of disease pathways in the human interactome," in Proc. Pacific Symposium on Biocomputing (PSB), 2018, pp. 111-122. Accessed: Nov. 18, 2024 [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC5731453/
S. Pandit, D. Chau, S. Wang, and C. Faloutsos, "NetProbe: A fast and scalable system for fraud detection in online auction networks," in Proc. 16th International World Wide Web Conference (WWW), 2007, pp. 201-210. Accessed: Nov. 18, 2024 [Online]. Available: https://www.cs.cmu.edu/~christos/PUBLICATIONS/netprobe-www07.pdf
U. N. Raghavan, R. Albert, and S. Kumara, "Near linear time algorithm to detect community structures in large-scale networks," Physical Review E, vol. 76, no. 3, 036106, 2007. Accessed: Nov. 18, 2024 [Online]. Available: https://arxiv.org/pdf/0709.2938
J. Gilmer, S. S. Schoenholz, P. F. Riley, O. Vinyals, and G. E. Dahl, "Neural message passing for quantum chemistry," in Proc. 34th International Conference on Machine Learning (ICML), 2017, pp. 1263-1272. Accessed: Nov. 18, 2024 [Online]. Available: https://proceedings.mlr.press/v70/gilmer17a.html
Y. Li, C. Gu, T. Dullien, O. Vinyals, and P. Kohli, "Graph matching networks for learning the similarity of graph structured objects," in Proc. 36th International Conference on Machine Learning (ICML), 2019, pp. 3835-3845. Accessed: Nov. 18, 2024 [Online]. Available: https://proceedings.mlr.press/v97/li19d.html
P. Sen, G. Namata, M. Bilgic, L. Getoor, B. Galligher, and T. Eliassi-Rad, "Collective classification in network data," AI Magazine, vol. 29, no. 3, pp. 93-106, 2008. Accessed: Nov. 18, 2024 [Online]. Available: https://doi.org/10.1609/aimag.v29i3.2157
24. W. W. Zachary, "An information flow model for conflict and fission in small groups," Journal of Anthropological Research, vol. 33, no. 4, pp. 452-473, 1977. Accessed: Nov. 18, 2024 [Online]. Available: https://www.jstor.org/stable/3629752
A. Grover and J. Leskovec, "node2vec: Scalable Feature Learning for Networks," in Proc. 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2016, pp. 855-864. Accessed: Nov. 18, 2024 [Online]. Available: https://arxiv.org/abs/1607.00653

FAQ

What is “machine learning on knowledge graphs,” and why is it critical?

It is the application of ML methods that directly leverage the structure and semantics of knowledge graphs (nodes, relationships, attributes) to make predictions, discover patterns, and generate insights. It’s critical because many real-world data items are interdependent, not i.i.d., and much of the useful signal lies in relationships. This enables Intelligent Advisory Systems to surface non-obvious insights in domains like drug discovery, personalized medicine, and law enforcement where human-only exploration is impractical.

What are the main categories of graph ML tasks?

They are broadly divided into node-focused tasks and graph-focused tasks. Node-focused tasks (e.g., node classification, link/relationship prediction, community detection) operate on one graph and predict properties of nodes or edges. Graph-focused tasks (e.g., graph classification) treat each entire graph as a single data point and predict a label for the whole graph.

Why does the i.i.d. assumption break in graph ML, and what patterns replace it?

In graphs, data points (nodes) are connected and influence one another, violating the independence assumption. Instead, patterns like homophily (similar nodes connect), structural equivalence (nodes with similar neighborhoods share roles), and heterophily (dissimilar nodes connect) guide learning. This is why node classification and link prediction usually consider both node features and network structure.

Is node classification supervised, unsupervised, or something else?

It is typically considered semi-supervised: during training, models often see the entire graph (including unlabeled nodes) and exploit neighborhood information to learn, even though target labels are unknown for some nodes. Note that “semi-supervised” in classical ML assumes i.i.d. data, which graph data violates—another reason graph ML doesn’t cleanly fit classical categories.

What is link prediction (relationship prediction), and why are edges often missing?

Link prediction estimates whether a connection exists or will form between two nodes; relationship prediction extends this to predict the specific relationship type. Edges are often missing because graphs are incomplete (unobserved or intentionally hidden connections) and dynamic (new interactions continually arise). Applications include recommendations, knowledge graph completion, fraud/crime analysis, and biomedical discovery.

How do community detection and graph clustering differ from traditional clustering like K-means?

Traditional clustering operates on independent vectors, while graph clustering/community detection leverages the network’s topology (edges, directions, and weights). Algorithms like Louvain and Label Propagation use connectivity patterns (e.g., modularity optimization or label propagation) to find densely connected groups without requiring feature extraction or a training phase.

How does graph classification differ from node classification?

Node classification predicts labels for nodes within one graph, using the network’s connectivity. Graph classification treats each entire graph as one sample (often i.i.d.), learning a mapping from graphs to labels (e.g., molecular toxicity). The challenge is crafting representations that capture both node attributes and structural patterns at the whole-graph level.

What are the two main implementation approaches: graph-native vs. feature-engineered?

Graph-native methods (e.g., collective classification, community detection) operate directly on the graph structure. Feature-engineered approaches first convert nodes/edges/graphs into vectors and then apply conventional ML/DL methods. Modern GNNs automate much of this representation learning, producing embeddings that encode structure and attributes for downstream tasks.

Why is feature engineering (featurization) on graphs so important?

Model performance hinges on informative representations of nodes, relationships, and graphs. Domain-aware features (e.g., degrees by neighbor group to capture homophily) can dramatically outperform naive features; even automatic methods like Node2Vec or GNNs require thoughtful parameterization. Critically, the featurization used in prediction must match the one used in training, and feature scaling can be essential for stable learning.

How do GNNs and LLMs complement each other in Intelligent Advisory Systems?

GNNs excel at learning from graph structure to detect complex patterns (e.g., molecular interactions, patient similarity networks, suspicious transaction networks). LLMs interpret and synthesize unstructured text (papers, guidelines, reports) and explain model findings in clear language. Together, they produce pattern discovery plus context-appropriate, explainable recommendations, amplifying the impact of knowledge graphs in real-world decision-making.

pro $24.99 per month

access to all Manning books, MEAPs, liveVideos, liveProjects, and audiobooks!
choose one free eBook per month to keep
exclusive 50% discount on all purchases
renews monthly, pause or cancel renewal anytime

lite $19.99 per month

access to all Manning books, including MEAPs!

team

5, 10 or 20 seats+ for your team - learn more

eBook

pdf, ePub, online

$64.99 $32.49

you save $32.50 (50%)

include audio $24.99 $12.49

pro $24.99 per month

access to all Manning books, MEAPs, liveVideos, liveProjects, and audiobooks!
choose one free eBook per month to keep
exclusive 50% discount on all purchases
renews monthly, pause or cancel renewal anytime

lite $19.99 per month

access to all Manning books, including MEAPs!

team

5, 10 or 20 seats+ for your team - learn more

eBook

$64.99 $32.49

you save $32.50 (50%)

include audio $24.99 $12.49

eBook

pdf, ePub, online

$64.99 $32.49

you save $32.50 (50%)

include audio $24.99 $12.49

pro $24.99 per month

access to all Manning books, MEAPs, liveVideos, liveProjects, and audiobooks!
choose one free eBook per month to keep
exclusive 50% discount on all purchases
renews monthly, pause or cancel renewal anytime

lite $19.99 per month

access to all Manning books, including MEAPs!

team

5, 10 or 20 seats+ for your team - learn more