science1 distinct publisher
Mitchell's complaint: we grade models with human exams when the control group is animals and infants
On Quanta's podcast, the Santa Fe Institute's Melanie Mitchell says we lack adequate methods for measuring machine cognition, and points to how psychologists test babies and animals.
Publishers:quantamagazine.org
Reality
- Evidence32
- Adoption
- Insufficient
- Hype gap+10
- Incentives30
- Confidence45