2026
BA-FedSHAP: A reproducible toolkit for auditing background-induced attribution drift
SSRN
Preprint on how the choice of background distribution alters federated attribution results, and what that does to an audit.
PublicationsAuthor
Forskningsområde
Integritetsbevarande samverkande AI
Institutioner som inte kan samla sina känsliga data kan ändå lära tillsammans, om styrningen byggs jämsides med metoden.
- Federated learning
- Privacy
- AI assurance
Vad måste gälla, tekniskt och institutionellt, för att federerad analys av känsliga dataset ska vara förtroendevärd och inte bara decentraliserad?
Federated learning is often presented as a privacy guarantee. It is better understood as a change in where the risk sits. Moving computation to the data removes one exposure and introduces others: attribution that shifts with modelling choices, updates that leak, and a governance question about who is accountable when no single party holds the dataset.
I work on both halves of that, the technical behaviour of federated methods under audit, and the institutional arrangements that make a federation legible to a regulator or a ministry.
Resultat
Relaterade arbeten
2026
BA-FedSHAP, auditing background-induced attribution drift
Measures how much a federated explanation depends on an arbitrary modelling choice.
Research softwareAuthor
2025
Federated learning for children's data: encoding trust in education systems
UNICEF Innocenti, Office of Strategy and Evidence
Hävdar att stater kan använda känsliga utbildningsdata utan att centralisera dem, och vilken styrning det kräver.
PolicyAuthor