Read about Ontotext’s KGF22 days dedicated to stories about knowledge graphs in the domains of Industry, Healthcare & Life Sciences and Financial Services
The Knowledge Graph Conference (KGC) has proved to be a must-attend event for all graph enthusiasts. Held from May 8-12 at Cornell Tech in New York, the conference brought together a vibrant community of experts, practitioners and vendors in the graph and semantic tech space. The event attracts individuals interested in graph technology, machine learning and natural language processes in numerous verticals, including publishing, government, financial services, healthcare and life sciences, manufacturing and retail.
The conference positioning focused on knowledge graphs as a mature, enterprise-ready technology for long-term and mission-critical use cases that require security, resilience and scalability. The technology may not seem flashy or exciting, but it delivers specialized capabilities when integrated with third-party solutions. This message resonates with the market positioning of Ontotext as a trusted, stable option for demanding data-centric use cases.
During the conference, the organizers hosted a separate track called the Healthcare and Life Sciences Symposium.
One notable observation was the absence of prominent property graph vendors active in the last couple of KGC editions.The conference focused on knowledge graphs as a mature, enterprise-ready technology for long-term and mission-critical use cases requiring security, resilience and scalability. Click To Tweet
Given the competitive landscape, our Director of Financial Services and Publishing, Peio Popov, presented our partner ecosystem and positioned it as the mature, resilient and secure option for building enterprise knowledge graphs-based projects. Peio called this ecosystem “The Fellowship of the Graph”. This fellowship is an ideal option for enterprises looking for a solution that can be viable for 5+ years to maximize return on investment. Our graph database GraphDB and plugins are designed with great attention to detail, making it perfect for partnering and integration. Its remarkable capabilities shine even brighter when delivered jointly with partners.
The Master class that Peio delivered covered Reconciliation, Text Analytics and Virtualization with Knowledge Graphs, using GraphDB and Text Analysis services tailored explicitly for tasks demanding intricate domain knowledge and linking documents to reference or master data.
There were many conversations centered around the intersection of knowledge graphs and machine learning (ML). An entire conference track was dedicated to topics like graph embeddings, vector representations and matrix operations.
Those techniques are already playing an increasingly important role in the graph space, the capabilities of knowledge graphs and providing new insights previously impossible. As the field continues to evolve, we are sure to see more applications of ML and AI in the graph space, and we expect these techniques to become even more critical for solving complex enterprise problems across various industries.
The general consensus was that knowledge graphs could gain significant advantages by incorporating machine learning capabilities.We are sure to see more applications of ML and AI in the graph space, and we expect these techniques to become even more critical for solving complex enterprise problems across various industries. Click To Tweet
The rise of ChatGPT and other large language models (LLM) has undoubtedly impacted the field of knowledge graphs. These models can generate more coherent and human-like text than ever before, potentially reducing the need for manually curated knowledge graphs.
However, knowledge graphs still have several advantages over LLM. First, knowledge graphs are structured, making it easier to perform complex queries and extract information more precisely and accurately.
They also provide a way to represent relationships between entities, which can be helpful in many applications, such as recommendation systems and fraud detection.
We saw presentations showing the potential for knowledge graphs and LLM to complement each other. For example, knowledge graphs can be used to provide structured data to train LLM, and LLM can be used to extract information from unstructured data sources such as text and images, which can then be incorporated into knowledge graphs. How sweet!
Knowledge graphs will continue to be essential for AI in the era of ChatGPT and LLM. While the technology may continue to evolve and change, the fundamental principles of representing structured data and relationships between entities are important in many applications.
As we saw during the Healthcare and Life Sciences Symposium, knowledge graphs uniquely provide web-scale computation, integration and retrieval over heterogeneous biomedical data and knowledge sources.
In Healthcare and Life Sciences, knowledge graphs have several use cases, such as personalized search, recommendations, clinical decision support, drug discovery and prediction models. In addition, the use of knowledge graphs is expected to drive innovation in Biomedical AI in the 21st century.
However, with the explosion of biomedical knowledge data, there is a need to manage diverse data sources and develop advanced strategies for how it is collected, stored, accessed and structured.
Overall, the Knowledge Graph Conference showed where the technology and community are headed. While focused on connecting members of the graph community, the event highlighted trends around enterprise adoption, competitive positioning and the relationship between knowledge graphs and AI that are worth following as the space continues to evolve. The impact potential for knowledge graphs remains strong, even as the landscape shifts.