Skip to content
49. Nvidia A100, Training Game AI, Neural Networks | AI Developer Interview episode artwork

May 20, 2020 1 hr 15 min

49. Nvidia A100, Training Game AI, Neural Networks | AI Developer Interview

from Broken Silicon Hosted by Moore's Law Is Dead

Hot off the heels of the announcement of Nvidia A100, we have an AI Developer on to talk about what these massive GPU’s are actually used for – AI & ML applications. SPONSOR: https://www.cdkoffers.com 25% Windows Code discount: brokensilicon 3% off all software: dieshrink Win10 pro oem key 13$: https://bit.ly/2Wdfghh Win10 Home oem key 10$: https://bit.ly/3dsbSFi Office 2016 key 27$: https://bit.ly/2Wb3sMx Office 2019 key 44$: https://bit.ly/2WdfBAz Win10 pro oem+Office 2016 35$: https://bit.ly/3bkJeVb 1:21 AI Buzzwords & Introductions (some sound issues) 6:03 What is a Neural Network? 12:53 Challenges in Training Neural Networks 16:21 Sparsity and Pruning 24:03 How will AI Improve our Lives? 35:42 Bottlenecks to AI Research 42:03 AI & ML in Gaming 47:38 CUDA Intrenchment, AMD’s ability to enter the market 50:58 Cerebras & Graphcore 55:53 AMD vs Nvidia Graphics 1:03:13 Comparing Synthetic AI to Biological Brains https://www.thispersondoesnotexist.com/ https://devblogs.nvidia.com/nvidia-ampere-architecture-in-depth/ https://www.cerebras.net/cerebras-wafer-scale-engine-why-we-need-big-chips-for-deep-learning/ https://cirrascale.com/graphcore-cloud.php?utm_term=graphcore&utm_source=adwords&utm_medium=ppc&utm_campaign=Graphcore&hsa_net=adwords&hsa_grp=86815837387&hsa_mt=p&hsa_tgt=kwd-606839681642&hsa_kw=graphcore&hsa_src=g&hsa_acc=2793442874&hsa_cam=8112712212&hsa_ver=3&hsa_ad=397149594120&gclid=CjwKCAjwwYP2BRBGEiwAkoBpAuc9L_R5wt9rbYuJ7jeopg5W5YdYxaFBkNVpQz336TjcNJYg9jyZ2BoCX0MQAvD_BwE https://arxiv.org/pdf/1803.03635.pdf https://arxiv.org/pdf/1807.01281.pdf https://podcasts.google.com/feed/aHR0cHM6Ly9jaGFuZ2Vsb2cuY29tL3ByYWN0aWNhbGFpL2ZlZWQ/episode/Y2hhbmdlbG9nLmNvbS83LzcwNA?hl=en&ved=2ahUKEwiuzYfG_7jpAhUIHs0KHSavAhAQjrkEegQIBxAE&ep=6 https://twimlai.com/upside-down-reinforcement-learning/?fbclid=IwAR18e99xWPGUN0_SVG84IKUaT5KYSKVhzwMdq37goiOdDwRuUf8s9Z_8GSI https://en.wikipedia.org/wiki/Reinforcement_learning https://en.wikipedia.org/wiki/Concept_learning https://en.wikipedia.org/wiki/Unsupervised_learning#cite_note-2 https://arxiv.org/abs/1912.02875 https://arxiv.org/pdf/1912.02877.pdf https://www.alexirpan.com/2018/02/14/rl-hard.html https://www.microsoft.com/en-us/research/blog/turing-nlg-a-17-billion-parameter-language-model-by-microsoft/ https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot/ https://openai.com/blog/jukebox/ https://www.pnas.org/content/pnas/116/43/21854.full.pdf

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

49. Nvidia A100, Training Game AI, Neural Networks | AI Developer Interview

0:00 1:15:16

No transcript yet.

Requests help us decide what to transcribe next.

Frequently asked questions

How long is this episode of Broken Silicon?

This episode is 1 hour and 15 minutes long.

When was this Broken Silicon episode published?

This episode was published on May 20, 2020.

Can I download this Broken Silicon episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.