Consciousness Club #35: Empiricism versus Nativism in Artificial Intelligence
Speaker: Robert Long (NYU)
Abstract: The age-old dispute between "empiricists" and "nativists" is a foundational issue in philosophy and cognitive science. Empiricists hold that knowledge is (almost) entirely learned from experience, whereas nativists hold that (many) mental capabilities are "built in" and arise independently of experience.
Historically, this has been a debate about human and animal minds. This talk expands the debate to artificial minds. Will truly intelligent AI require innate machinery, or can AI research proceed on empiricist grounds? Building on arguments advanced by Gary Marcus (team AI nativism) and Yann LeCun (team AI empiricism). I'll claim that, even though nativism about human intelligence is largely correct, there's reason to think that empiricism will prove more fruitful in artificial intelligence. However, the key challenge for AI empiricists, as for empiricists in general, is to account for how it is possible to learn abstract concepts from experience.
Entrance is free. Eventbrite registration is required. Click here to register.