
Sign up to save your podcasts
Or


NLP, transformer architectures, and machines beating humans at their own game: Jon Krohn talks to Alexander H. Miller about his work in building a machine that can outsmart humans in the game of Diplomacy by engineering powers of persuasion and collusion to its own advantage.
This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick (linkedin.com/learning/instructors/keith-mccormick). Interested in sponsoring a SuperDataScience Podcast episode? Visit JonKrohn.com/podcast for sponsorship information.
In this episode you will learn:
• Training a natural language model to interact with Diplomacy players [05:07]
• Processing speeds for a Diplomacy bot [29:32]
• Using transformer architectures [37:25]
• How Diplomacy AI actually works [43:25]
• CICERO's potential real-world applications [55:28]
• How to R&D an AI project [59:27]
• How to become an AI Research Manager [1:06:12]
Additional materials: www.superdatascience.com/663
By Jon Krohn4.6
294294 ratings
NLP, transformer architectures, and machines beating humans at their own game: Jon Krohn talks to Alexander H. Miller about his work in building a machine that can outsmart humans in the game of Diplomacy by engineering powers of persuasion and collusion to its own advantage.
This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick (linkedin.com/learning/instructors/keith-mccormick). Interested in sponsoring a SuperDataScience Podcast episode? Visit JonKrohn.com/podcast for sponsorship information.
In this episode you will learn:
• Training a natural language model to interact with Diplomacy players [05:07]
• Processing speeds for a Diplomacy bot [29:32]
• Using transformer architectures [37:25]
• How Diplomacy AI actually works [43:25]
• CICERO's potential real-world applications [55:28]
• How to R&D an AI project [59:27]
• How to become an AI Research Manager [1:06:12]
Additional materials: www.superdatascience.com/663

475 Listeners

1,086 Listeners

339 Listeners

769 Listeners

156 Listeners

268 Listeners

209 Listeners

141 Listeners

90 Listeners

131 Listeners

150 Listeners

208 Listeners

559 Listeners

266 Listeners

70 Listeners