Question to LAION team. Scaling open datasets is powerful, but does it resolve ambiguity in data? If inputs are inconsistent, larger datasets just scale that inconsistency. What if the real step is eliminating ambiguity at the data level before AI?
Question to LAION team.
Scaling open datasets is powerful, but does it resolve ambiguity in data?
If inputs are inconsistent, larger datasets just scale that inconsistency.
What if the real step is eliminating ambiguity at the data level before AI?