Search

Home > Data Skeptic > Multi-Agent Diverse Generative Adversarial Networks
Podcast: Data Skeptic
Episode:

Multi-Agent Diverse Generative Adversarial Networks
Category: Religion & Spirituality
Duration: 00:29:19
Publish Date: 2017-05-12 10:00:00
Description:

Despite the success of GANs in imaging, one of its major drawbacks is the problem of 'mode collapse,' where the generator learns to produce samples with extremely low variety.

To address this issue, today's guests Arnab Ghosh and Viveka Kulharia proposed two different extensions. The first involves tweaking the generator's objective function with a diversity enforcing term that would assess similarities between the different samples generated by different generators. The second comprises modifying the discriminator objective function, pushing generations corresponding to different generators towards different identifiable modes.
Total Play: 0

Users also like

300+ Episodes
Good Law | B .. 80+     10+
400+ Episodes
The Knowledg .. 200+     10+
1K+ Episodes
Entrepreneur .. 600+     50+