• Blog
  • Informational

Do Large Language Models Dream of Knowledge Graphs – Impressions from Day 2 At SEMANTiCS 2023

What if ChatGPT is the killer app for the Semantic Web? Why not train LLMs on lexical resources? And how do we tame the genie of LLMs for the healthcare industry? Read our report from Day 2 of SEMANTiCS 2023 to find out.

October 13, 2023 6 mins. read Teodora Petkova

Did you know that, if you add “take a deep breath” to a prompt, chances are you will get more accurate results from large language models (LLMs)? I didn’t either. I learned that fact from a comment in the audience on the second day of  SEMANTICS 2023 – the European conference series focused on semantic technologies ever since 2005.

What If ChatGPT Is the Killer App for the Semantic Web?

The last day of SEMANTiCS started as excitingly as all the rest. This time with a SPARQL query. 

Aidan Hogan, associate Professor at the Department of Computer Science, University of Chile, and an Associate Researcher at the Millennium Institute for Foundational Research on Data (IMFD), asked ChatGPT:

“Can you give me a SPARQL query for Wikidata to find Turin Award winners who were born in Latin America?”

Further, Hogan elaborated on the ways we can use the LLM to access knowledge from knowledge graphs, sparkling (pun intended) an interesting inner dialogue: What if we used ChatGPT to write SPARQL queries instead of natural language answers?

Do Knowledge Graphs Dream of Large Language Models?

Showcasing in real-time how ChatGPT can be used to write SPARQL queries, Aidan brought up the point that LLMs can be a potential entry point to the Semantic Web home. He shared the need for more research at the intersection of LLMs and knowledge graphs.

A slide from the publicly available Keynote of Aidan Hogan at SEMANTICS 2023

Hogan also spoke about the current capabilities and limitations of LLMs and compared them to those of knowledge graphs in terms of usability, explainability, provenance, resoluteness, mutability, timeliness, comprehensiveness, utility, expressiveness, cost, and truth.

“[LLMs] call into question a fundamental tenet of Data Management: that in order to address non-trivial information needs, the first step is to explicitly structure data in order to lift them from the ambiguous swamp of our human language. Aidan Hogan”

Throughout his presentation [PDF], he made a plethora of academic references on all the open questions deriving from use cases where the interplay between knowledge graphs and LLMs is involved.

Aidan Hogan at SEMANTiCS 2023. Source: SWC

Aidan concluded by saying that LLMs can make knowledge graphs more usable and widespread as a technology. He invited everyone to contribute to Transactions on Graph Data and Knowledge (TGDK) – an open-access journal publishing research about graph data and knowledge. He also reminded us all about his wonderful book, available online with open access.

Feeding The Graph With Lexical Resources

Somewhat complementary to Hogan’s presentation was the talk Data Spaces and Lexical Resources. I say complimentary because other than using Linked Data to better work with LLMs and knowledge graphs, we can also use lexical resources to enrich LLM’s “learning” path.

Ilan Kernermann, CEO of Lexicala, and Martin Kaltenböck, CFO of Semantic Web Company (SWC), presented an approach to making use of Language Models and Knowledge Graphs in Data Spaces and Data Markets to foster data exchange and semantic interoperability. Both speakers talked about common metadata standards and adequate language resources as key enablers of efficient interoperable, multilingual projects.

Ilan also spoke about the flaws of LLMs in terms of language-specific flaws, to kind of complement the flaws Hogan showed.

Ilan Kernermann presenting the flaws of LLM technology at SEMANTiCS 2023

From Ilan’s perspective, one rooted in years of work and research to provide multilingual lexical data solutions for the Language Technology industry and academia, these flaws were to be addressed by lexicology. Ilan argued that we could map the DNA of language by organizing lexical resources data-wise. 

Among the language resources Ian listed were systemic lexical mapping, definitions, sense indicators, typical language patterns, phraseology, domain classification, grammatical details, morphology, and cross-lingual correspondence.

Regulations For The Genie Out Of The Bottle: LLM-Based Chatbots In Medicine

The last talk I attended was titled: Laws For A Genie: Governance And Evidence Frameworks For Large Language Model-Based Chatbots In Medicine given by Stephen Gilbert, Professor of Medical Device Regulatory Science at the Else Kröner Fresenius Center for Digital Health, Technische Universität Dresden. It was an entertaining, highly informative, and thoughtful walk through the ethical and technological aspects of the use of LLMs in medicine. 

Stephen Gilbert stressed how important the regulation and ethical frameworks were for using LLMs in the healthcare domain. For, despite the great medical potential for LLM-based generative chat tools (such as ChatGPT or Google’s MedPaLM), LLMs with explainability, low bias, predictability, correctness, and verifiable outputs do not currently exist. This calls for developing frameworks and rules that will eventually tame the genie that is already out of the bottle. 

Data-Centric Endeavors Are Good But Knowledge-Centric Ones Are Better

The conference took place in conjunction with LANGUAGE INTELLIGENCE 2023 – an event showcasing the latest developments in Multilingual Artificial Intelligence. It offered 3 days full of academic and industry tracks, business talks, tutorials, and workshops. If I were to talk about some of the insights I got from the parallel event, I would probably end up with 3 more posts. Thankfully, lt-innovate.org already did a concise wrap-up. 

But, if I am to distill the gist, it would be: language technology.

Just like the typewriter in the hall hosting the Poster’s park, LLMs are yet another tool poised to change the way we work with language. And yes, although this tool might look sophisticated and exciting, it comes with many ethical concerns and a lot of work for us to do before we are able to use it responsibly and sustainably. As with any technology, the thing that matters most is the knowledge we manage to transfer via it. And, as we all know, when it comes to machine-mediated knowledge, a little semantics goes a long way.

A typewriter in the hall hosting the Poster’s park at SEMANTiCS 2023

SEMANTiCS 2023 proved that semantic technologies are invaluable when it comes to encoding knowledge and working towards better spaces of collaboration, innovation, and most importantly shared understanding. 

LLMS are only as knowledgeable as the data we feed them so let’s see how knowledgeable our LLMs will be and how knowledge-centric our projects are a year from now.

Looking forward to seeing you in Amsterdam for the 20th edition of SEMANTiCS in 2024!

Article's content

Marketing Expert at Ontotext

Teodora is a philologist fascinated by the metamorphoses of text on the Web. Curious about our networked lives, she explores how the Semantic Web vision unfolds, transforming the possibilities of the written word. From 2022 on, Teodora helps with the creation and curation of the Ontotext knowledge graph to foster information ecology out of marketing content that will enable relevant user experiences across Ontotext's universe.

GraphDB in Action: Smells Like Semantics Spirit

Read about a project called Odeuropa and a number of exciting applications delivered by it with our RDF graph database humming behind them

KGF 2023: Bikes To The Moon, Datastrophies, Abstract Art And A Knowledge Graph Forum To Embrace Them All

Read about this year’s Knowledge Graph Forum – a space where Ontotext and partners presented smart and linked ways to tame and thrive on complexity, rather than be drowned by it

Do Large Language Models Dream of Knowledge Graphs – Impressions from Day 2 At SEMANTiCS 2023

Read our report from Day 2 of SEMANTiCS 2023 to find out if ChatGPT is the killer app for the Semantic Web, how do we tame the genie of LLMs for Healthcare and more

Can LLMs Become Knowledgeable – Impressions from Day 1 At SEMANTiCS 2023

Read about the interplay between LLMs & KGs and how business and academia tackle them in our report from Day 1 at SEMANTiCS 2023

It’s Time We Give Each Other More Data Spaces: Impressions from the Pre-conference Day at SEMANTiCS 2023

Read about SEMANTiCS pre-conference day, which covered the topics of interoperability, ESG data, knowledge engineering, scholarly communication, and academia & industry collaboration.

GraphDB in Action: Navigating Knowledge About Living Spaces, Cyber-physical Environments and Skies 

Read about three inspiring GraphDB-powered use cases of connecting data in a meaningful way to enable smart buildings, interoperable design engineering and ontology-based air-traffic control

Your Knowledge Graph Journey In Three Simple Steps

A bird’s eye view on where to start in building a knowledge graph solution to help your business excel in a data-driven market

GraphDB in Action: Putting the Most Reliable RDF Database to Work for Better Human-machine Interaction

Read about the world of academia research projects that use GraphDB to meet the challenges of heterogeneous data across various domains

Knowledge Graphs for Retail – Connecting People, Products and Platforms

Read about how knowledge graphs can serve the retail industry’s growing need to connect, manage and utilize data efficiently, aligning it in a collaborative data ecosystem

Data Wants To Be Truly Sovereign: Designing Data Spaces with Linked Data Principles In Mind

Read about what data spaces are and how semantic technologies and Linked Data can make them a stronger and safer mechanism for commercial data exchange

GraphDB in Action: Powering State-of-the-Art Research

Read about how academia research projects use GraphDB to power innovative solutions to challenges in the fields of Accounting, Healthcare and Cultural Heritage

KGF22: Knowledge Graphs and The Not So Quiet Cognitive Revolution

Read about Ontotext’s KGF22 days dedicated to stories about knowledge graphs in the domains of Industry, Healthcare & Life Sciences and Financial Services

KGF22: Wittgenstein, Developers Empathy and Other Semantic Data Considerations

Read about our event report from Ontotext’s Knowledge Graph Forum 2022, highlighting expert insight on building knowledge graphs and designing enterprise-grade solutions with semantic technologies.

A Little SEMANTiCS Goes A Long Way

Take a sneak peek at some of the keynote speeches and tutorials throughout SEMANTiCS 2022

It Takes A Village To Raise An Enterprise Knowledge Graph

Read about the design processes behind crafting knowledge-graph enabled solutions and explore some of the stories of our partners.

Smart Buildings Are Built of Smart Data: Knowledge Graphs for Building Automation Systems

Read about how knowledge graphs offer a sustainable solution for harnessing and making sense of heterogeneous data in the building automation industry.

Metadata Moves: Knowledge Graph Technology for Logistics

Read about how the world of metadata humming behind the logistics and other supply chain processes can benefit from using knowledge graph technology.

Electrical Standards, Smart Grids and Your Air Conditioner

Read about how applying Linked Data principles and semantic technology to electricity data can make for a more efficient, reliable and sustainable electricity market.

The Semantic Web: 20 Years And a Handful of Enterprise Knowledge Graphs Later

Read about how the Semantic Web vision reincarnated in thousands of Linked Open Data datasets and millions of Schema.org tagged webpages. And how it enables knowledge graphs to smarten up enterprises data.

Metadata is Like Packaging: Seeing Beyond the Library Card Metaphor

Read about what metadata is, why it is important and how it can enhance the ways information flows across the enterprise.

From Fragmented Data to a Comprehensive Knowledge Graph: The Case for Building an R&D Repository

Read about how enterprise knowledge graphs can unlock meaning and thus create a smart future-proof living repository of scientific data and its relationships.

Texts Without Pages: Advancing Text Analytics with Content Enrichment

Read about how text analytics can be brought forward with content enrichment processes for better text authoring, delivery and navigation.

A Shield Built of Connected Data: Knowledge Graphs Meet Cybersecurity

Read about how a knowledge graph can help organizations stay vigilant of the increasing number of cyber threats, keeping malicious attacks at bay with the help of semantics.

Digital Twins: If It Sounds Like Cyberpunk, It’s Because It Is

Read about what digital twins are, what makes them attractive to companies and how digital twins relate to semantic technology and enable organizations to design, simulate and validate various scenarios virtually.

Eating the Knowledge Soup, Literally

Read about the fluid essence of knowledge and the capability of knowledge graphs to power an information-rich platform of diverse facts about anything, a broccoli soup included.

If Curiosity Cabinets Were Knowledge Graphs

Read about why and how knowledge graph technology can help build networks of interwoven digital objects in the world of cultural heritage.

On the Hunt for Patterns: from Hippocrates to Supercomputers

Read about the ExaMode project that will help medical professional use the power of supercomputers and knowledge graphs for more efficient patient care through data-driven diagnoses.

Crafting a Knowledge Graph: The Semantic Data Modeling Way

Read about how to build a knowledge graph the semantic data modeling way in 10 steps, provided by our knowledge graph technology experts.

A Graphful of Investment Opportunities

Read about the story of an algorithm that mines data to narrow down opportunities for investing.

Okay, You Got a Knowledge Graph Built with Semantic Technology… And Now What?

Read about how knowledge management can be made smarter using a knowledge graph built with semantic technology.

If Johnny Mnemonic Smuggled Linked Data

Read about how semantic technology and Linked Data can help enterprises benefit from smart data management and retrieval practices.

Data, Databases and Deeds: A SPARQL Query to the Rescue

Read about why and how SPARQL queries make for a better search in diverse datasets across an organization in an integrated way.

Semantic Technology and the Way We See the World

Read about how semantic technology can help you set the wheels for many processes related to еfficient data management and governance.

Telling Stories with an RDF Database

Read about the opportunities for authoring and publishing workflows opened by an RDF triplestore.

The Power of URI or Why Odysseus Called Himself Nobody

Read about URI and its power to enable the sharing and reuse of machine-readable data with minimum integration costs.

From Cultivating Nature to Cultivating Data: Semantic Technology and Viticulture

Learn how the potential that Big Data streams bring to grape and wine production can be harnessed with the right kind of technology.

The Knowledge Graph and the Enterprise

Read about the knowledge graph and about how many enterprises are already embracing the idea of benefiting from it.

It Don’t Mean a Thing If It Ain’t Got Semantics

Learn how you can turn data pieces into actionable knowledge and data-driven decisions with an RDF database.

The Bounties of Semantic Data Integration for the Enterprise

Learn about the potential semantic data integration carries for piecing massive amounts of data together.

Here’s a Graph, Go Figure! Coupling Text Analytics with a Knowledge Graph

Learn why and how a Knowledge Graph boosts significantly Text Analytics processes and practices and makes text work for us in a more meaningful way.

Cognitive Computing: Let’s Play an Awareness Game

Read about the new breed of computing is emerging before our eyes – cognitive computing and join us in our Awareness Game.

Machine Learning and Our (Insatiable) Penchant for Making Things Smarter

Read about how machines can be of great help with many tasks where fast and error-free computation over big amounts of data are required.

Staying In the Vanguard of Digital Transformation with Open Data

Learn about Open Data and its potential to be part of smart solutions to data problems and innovative products and services.

Whose Meaning? Which Ontology?

Read about how ontologies open up opportunities for a new class of tools to power information consumption and knowledge management.

Shiny Happy Data: A Praise for RDF

Learn how to choose the right solution for working with your data the conceptual framework of “happy connected people”.

Enterprise Metadata Matters: From Having Data to Acting Upon Them

Learn more about the importance of being metadata-driven today in our latest SlideShare presentation.

Data Daiquiri: The Power of Mixing Data

Learn how companies can tap into the power of the data coming their way by integrating the huge data flows with their proprietary data.

Migrating to GraphDB: Your Why and How in 20 slides

Learn about the steps you need to migrate your data to GraphDB to use it as a smart brain on top of your legacy systems.

Got meaning? Or Why an RDF Graph Database Is Good for Making Sense of Your Data

Read about how you can create systems capable of discovering relationships and detecting patterns within all kinds of data.

Brains Armored with Smart Data

Read our thoughts rising from questions such as “Will Giant Brains Rule the World?” and “Can a mechanical brain replace you?”

One Step Closer to Intertwingularity: Semantic Metadata

Learn about how semantic metadata allows us to add granularity to an object, interlink it to other objects and make it easy to search.

Exceptional User Experiences with Meaningful Content NOW

Content enrichment and semantic web technologies are key to efficient content management. Learn why and see these technologies in action.

Semantic Information Extraction: From Data Bits to Knowledge Bytes

Learn about semantic information extraction and how it pulls out meaningful data from textual sources, ready to be leveraged for insights, decisions and actions.

Weaving Data Into Texts: The Value of Semantic Annotation

Read about how semantic annotation links certain words to context and references that can be processed by an algorithm.

Exploring Linked Open Data with FactForge

Learn about FactForge and how you can turn the opportunities that data flows on the web can pour into our business into a real experience.

What is GraphDB and how can it help you run a smart data-driven business?

Learn about GraphDB in a simple and easy to understand way and see what Ontotext’s semantic graph database has to do with pasta making.

Linked Data for Libraries: Our New Librarians

Learn how semantic technologies can bring audiences back to libraries and make library archives and collections visible and accessible.

Working with Data Just Got Easier: Converting Tabular Data into RDF Within GraphDB

Read about OntoRefine – a new tool that allows you to do many ETL (extract, transform and load) tasks over tabular data.

GraphDB: Answers for Kids of All Ages

Read about how GraphDB can help you clean up messes of data (just like your room) – whether you are a kid or not.

The Knowledge Discovery Quest

Learn how by joining the dots, semantic search enhances the way we look for clues and compare correlations on our knowledge discovery quest.

Connectivity, Open Data and A Bag of Chips

Learn how LOD’s connectivity allows data to be shared seamlessly, used and reused freely. As simple as a bag of chips.

Data Integration: Joining the Data Pieces of Your Business Puzzle

Learn how to use information interconnectedness to integrate, interpret and ultimately make sense of data.

Cooking Up the Semantic Web

Read about the Semantic Web and what it takes to reach its potential and evolve from a Web of Documents to a Web of Data.

Semantic Search: The Paradigm Shift from Results to Relationships

Read about semantic search and how it takes information retrieval to the next level and puts information at our fingertips.

A Web of People and Machines: W3C Semantic Web Standards

Learn how and why Semantic Web Standards are to serve the Web of Data for better collaboration between people through computers.

Thinking Outside the Table

Learn how to manage highly connected data, working with complex queries and having readily available relationships, without the need to express them explicitly.

Our Networked Lives, Publishing and Semantic Technologies

Read about how semantic technology helps publishing handle data in an interconnected way, attaching machine-processable and readable meaning to them.

Why Graph Databases Make a Better Home for Interconnected Data Than the Relational Databases?

Read about how you can turn data into a resource, easily accessed and effectively used across the organization with a graph database.

Text, Data and the Roman Roads: Semantic Enrichment

Read about semantic enrichment and the unique opportunity it offers for interconnecting objects to facilitate knowledge discovery.

4 Things NOW Lets You Do With Content

Go beyond conventional publishing with Ontotext’s News On the Web and get the feel of how you can discover and consume content with semantic technology.