Assertional Knowledge Engineering
Information Extraction
- From NLP community: relation extraction
- closed IE vs open IE
- related survey: https://arxiv.org/abs/2312.17617
- related approach: https://aclanthology.org/2024.findings-acl.839/
- extract structural knowledge from natural language texts.
- typical: Named Entity Recognition (NER), Relation Extraction (RE) and Event Extraction (EE)
- References: Derong Xu, Wei Chen, Wenjun Peng, Chao Zhang, Tong Xu, Xiangyu Zhao, Xian Wu, Yefeng Zheng, and Enhong Chen. Large language models for generative information extraction: A survey. CoRR, abs/2312.17617, 2023
Contributors:
- Diego Collarana (FIT)
- Sven Hertling (FIZ), Harld Sack (FIZ), Heike Fliegl (FIZ)
- Desiree Heim (DFKI)
- Markus Schröder (DFKI)
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
KG Completion (A-Box)
Link Prediction
Relation Prediction
Fact Checking / Triple Testing
Literal Completion (labels/comments/descriptions)
Entity Linking (between KGs)
- Given two entity representations as text (verbalized or in RDF serialization), ask LLM if the entities are referring to the same real-world entity.
- candidate generation can be done via Sentence BERT models
- typical: mapping mentions to their corresponding entities given a context
- between KGs: ... ?
- also embedding based entity linking, more general: "Neural entity linking"
- LLM: good for interpreting uncommon mention
- References: Ozge Sevgili, Artem Shelmanov, Mikhail Y. Arkhipov, Alexander Panchenko, and Chris
Biemann. Neural entity linking: A survey of models based on deep learning. Semantic Web, 13(3):527–570, 2022
Contributors:
- Sven Hertling (FIZ), Harald Sack (FIZ), Heike Fliegl (FIZ)
- Desiree Heim (DFKI)
- Markus Schröder (DFKI)
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Entity Disambiguation
- Disambiguation: given mentions and contexts, need a function which maps them to entities in KG
- Oftentimes also part of IE pipelines (e.g. KG Completion)
- has typical steps: candidate generation, entity ranking
- Direct LLM assistance: Verbalized description of link candidates compared against a text snippet that contains the entity that should be linked/disambiguated + formulated as a classification task
- Indirect LLM assistance: E.g., using LLM-assisted KG IE from texts and compare the found entities and relationships against the neighborhood of the link candidates
- References: Golnaz Shapurian. Large language models and knowledge graphs for astronomical entity disambiguation. arXiv preprint arXiv:2406.11400, 2024; Shaojun Liu and Yanfeng Fang. Use large language models for named entity disam-
biguation in academic knowledge graphs. In 2023 3rd International Conference on
Education, Information Management and Service Science (EIMSS 2023), pages 681–
691. Atlantis Press, 2023
Contributors:
- Desiree Heim (DFKI)
- Markus Schröder (DFKI)
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Terminological Knowledge Engineering
Ontology Design
Contributors:
- Sven Hertling (FIZ), Harald Sack (FIZ), Heike Fliegl (FIZ)
- Michael Wetzel (Coreon)
- Sabine Mahr (word b sign)
- ...
Competency Question (CQ) Generation
User Stories / Personas Generation
Ontology Learning (Automated ontology design from text)
Ontology Evaluation
Contributors:
- Sven Hertling (FIZ), Harald Sack (FIZ), Heike Fliegl (FIZ)
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Competency Question (CQ) generation (from given ontologies)
CQ to SPARQL
Ontology Mapping
Contributors:
- Sven Hertling (FIZ), Harald Sack (FIZ), Heike Fliegl (FIZ)
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Ontology Documentation
Contributors:
- Daniel Baldassare (doctima)
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Draft from Daniel Baldassare :
- Data model description: nodes and relationships classes
- Metadata model description: nodes and relationships' s metadata (identifiers, optional and required metadata)
- Use with LLM:
- How to use it for standard RAG ( which embedding model)
- How to use it for GraphRAG/semantic layer (which embedding model, which additional metadata)
Class and Relation Descriptions/Labels
Reasoning
Approx/Probabilistic Reasoning via LLMs (LLM supported)
Contributors:
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Constraint Checking
Contributors:
- Robert David
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Data Repairs (→ maybe move to completion?)
Contributors:
- Robert David
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Downstream Tasks
KG/Ontology Embeddings
Contributors:
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- Please add yourself if you want to contribute ...
- ...
Please add other Downstream Tasks ...
Please add other Downstream Tasks ...
Please add other Downstream Tasks ...
User Interface / Access
Natural Language Interface to KG
- Natural Language to SPARQL
- Text2QueryLanguage (e.g. Text2SPARQL)
- Direct translation of user queries into equivalent knowledge graph query languages by prompting an LLM
- The knowledge graph structure can be given either by including a (sub)schema or a subgraph in the query translation prompt
- KG-RAG
- Can be done by extracting knowledge graph entities and potentially relevant relationships from the user query and retrieving relevant triples from the knowledge graph which are given as a context to the prompt containing the user query
- References: Jacopo D’Abramo, Andrea Zugarini, and Paolo Torroni. Dynamic few-shot learning for knowledge graph question answering. arXiv preprint arXiv:2407.01409, 2024; ] Dmitrii Pliukhin, Daniil Radyush, Liubov Kovriguina, and Dmitry Mouromtsev. Improving subgraph extraction algorihtms for one-shot sparql query generation with large language models. In QALD/SemREC@ ISWC, 2023.
Contributors:
- Diego Collarana (FIT)
- Sven Hertling (FIZ), Harld Sack (FIZ), Heike Fliegl (FIZ)
- Desiree Heim (DFKI)
- Markus Schröder (DFKI)
- Sabine Mahr (word b sign)
- Please add yourself if you want to contribute ...
- ...
KG to Natural Language (verbalization)
- OWL constructs to natural language (usually deterministic and no LLM involved?)
- Typically triple-based verbalization
- Different granularity of verbalizations (very exact subject/predicate/object sentences vs summary of subgraphs)
- ABOX statement verbalization: Often employed in RAG-based KGQA processes
- TBOX statement verbalization: Relevant for Text2QueryLanguage, Ontology-based KG Completion tasks
- References: Phillip Schneider, Manuel Klettner, Elena Simperl, and Florian Matthes. A comparative analysis of conversational large language models in knowledge-based text generation. arXiv preprint arXiv:2402.01495, 2024; Sven Hertling and Heiko Paulheim. Olala: Ontology matching with large languagemodels. In Proceedings of the 12th Knowledge Capture Conference 2023, pages 131–139, 2023
Contributors:
- Daniel Baldassare (doctima)
- Sven Hertling (FIZ), Harald Sack (FIZ), Heike Fliegl (FIZ)
- Desiree Heim (DFKI)
- Markus Schröder (DFKI)
- Sabine Mahr (word b sign)
- ...
Multilingual Translation of Literals
Contributors:
- Michael Wetzel (Coreon)
- Sabine Mahr (word b sign)
- Please add yourself if you want to contribute ...
- ...