We suggest the following topics for the roundtable discussion. We will compile the answers and make them available here as well.
- discipline relations: split work, or everyone a generalist?
- validation: only validate existing theories or propose new ones?
- reproducibility: what if data is private/proprietary?
- ethics: what questions should we ask (over/underexposure)?
- bias prevention: data collection, experimental setup, models?
- bias influence: effect of inherent data bias (gender/language)?
Please also remember the questionnaire