OntoWEEE

OntoWEEE uses language-model agents to extract information from research papers about waste electrical and electronic equipment (WEEE), then links it with product catalog data. The resulting knowledge graph connects products, components, recycling processes and materials, while preserving source evidence and experimental conditions.

Open the interactive demo

What is in the demo?

The current demo brings together 20 research papers, with a focus on batteries, and 12 mobile phone catalog records connected to paper entities. Nine catalog connections are model alignments; three are weaker variant links. The demo includes a fused graph, extraction traces for all 20 papers, and guided walkthroughs for six examples.

How to explore it

• Knowledge graph: Search for an entity, such as lithium-ion battery or Nokia 5110, or filter by paper, entity type or relation. Expand groups to see individual samples. Click a relation to inspect its measurements, conditions and evidence. Drag a node to move it, drag empty space to pan, and scroll to zoom.
• Extraction trace: Select a paper and choose Browse freely to follow the archived table and text extraction stages. For the six guided examples, Start walkthrough explains the steps. Evidence links connect graph results to the relevant replay records and available source passages or table cells.
• Paper: Find links to the original publications and the GSMArena catalog source.

How to read the graph

A group of nodes with the same name is a browsing aid. Samples from different papers remain separate, with their own measurements and conditions. A phone model link aligns a paper's model designation with a catalog record; it does not transfer measurements between them.

Read recovery values together with their experimental conditions. The traces replay archived extraction outputs and reviewed corrections, so you can inspect how the displayed results were obtained.