Laboro.AI Shines at ISWC 2026 with Successful Paper Acceptance

In an extraordinary achievement, Laboro.AI, a leading company specializing in custom AI development, has announced that its primary paper has been accepted for the second consecutive year at the prestigious ISWC 2026 Challenge Track. This international conference, known for showcasing advancements in the intersection of knowledge graphs and AI technology, will be held from October 25 to 29, 2026, in Bari, Italy.

The ISWC 2026 Challenge Track is particularly significant as it serves as a competitive platform where participants showcase their abilities to process knowledge and assess the performance of AI and semantic technologies through shared data and challenges. The specific paper accepted for presentation explores innovative strategies developed through Laboro.AI's engineering team's participation in an international competition titled LLMs4OL.

Research Overview

In this competition, teams leverage large language models (LLMs) to systematically structure knowledge from unstructured text, transforming it into well-organized ontologies. This year, Laboro.AI tackled two significant tasks in the challenge:

1. Flagship Task: This involved constructing an ontology from a single document, incorporating essential terms, their classifications, and relationships comprehensively.
2. Reuse Task: Unlike the previous task, this required expanding an existing ontology rather than reconstructing it from scratch.

Both tasks utilized a lightweight system that enabled the company to secure the top position in the previous year's LLMs4OL. This foundational system demonstrated exceptional performance despite the increased difficulty of this year's challenges.

Throughout the process, Laboro.AI discovered that a unified approach to extraction, without breaking the process into separate stages, led to higher performance levels. Reflecting this discovery, the paper is titled "Laboro.AI at LLMs4OL 2026 Tasks Flagship and Reuse: Less is More for LLM-based Ontology Learning". This embodies the core philosophy of keeping processes simple to achieve optimal results.

Notably, the challenges in this year’s competition were considerably heightened compared to the previous year. In 2025, the evaluation focused on separate stages of ontology construction; however, for the current Flagship task, the completed ontology as a whole was evaluated based on its structural accuracy. Despite competition doubling in size, Laboro.AI proudly achieved third place in the Flagship task, marking another successful year with paper acceptance.

Background on Research

Many organizations possess vast amounts of knowledge scattered across documents, manuals, and reports, yet often lack the means to efficiently harness this data. Most available AI solutions can read documents but struggle to comprehend nuanced knowledge structures, which differ significantly across corporate environments.

An ontology serves as a comprehensive