Programme time: 13-Oct | 15:30 | 75 min
Room: Arrabida Room (SOTA Benchmark)

Full description and objectives

This workshop examines how AI —multimodal perception, machine learning, and autonomous actuation — is transforming material recovery across the circular economy. Hyperspectral imaging combined with AI enables real-time, non-destructive identification of material composition at conveyor speed, opening radically new possibilities for high-purity sorting of streams that were previously uneconomical to recycle.

The session brings together researchers, technology developers, industrial operators, and policy stakeholders to explore four demonstration sectors:

• E-waste: component-level identification and selective disassembly

• Textile waste: fibre-blend classification for closed-loop recycling

• Battery recycling: chemistry detection for safe and efficient black-mass recovery

• Marine plastics: polymer typing of heterogeneous, degraded debris at sea and on shore

Learning goal: Participants will leave with a concrete understanding of hyperspectral AI pipelines, the readiness level of autonomous sorting across different waste streams, and the data, standardisation, and regulatory conditions needed for industrial deployment.

Predefined discussion questions: What multimodal data strategies maximise cross-sector transferability of models? What industrial, regulatory, and infrastructural conditions must Europe put in place to turn AI-enabled sorting into circular-economy reality?

Why you want to join?

This workshop explores how AI — combining multimodal sensing, hyperspectral imaging, and autonomous robotic sorting — can power the next generation of circular material systems across e-waste, textiles, batteries, and marine plastics.

Running agenda