Harri Valpola's guest appearances

Harri Valpola

Boss Level Podcast

Machine Learning Street Talk (MLST)

Leila Janah

Ben Goertzel

Noam Chomsky

Melissa Perri

Gary Marcus

Sara Hooker

Kenneth Stanley

Max Welling

Lukas Biewald

Linda Liukas

Melanie Mitchell

Pedro Domingos

Rutger Bregman

Thomas Lux

Aaron Dignan

OwlTail

In this episode of Machine Learning Street Talk, Tim Scarfe, Yannic Kilcher and Connor Shorten interviewed Harri Valpola, CEO and Founder of Curious AI. We continued our discussion of System 1 and System 2 thinking in Deep Learning, as well as miscellaneous topics around Model-based Reinforcement Learning. Dr. Valpola describes some of the challenges of modelling industrial control processes such as water sewage filters and paper mills with the use of model-based RL. Dr. Valpola and his collaborators recently published “Regularizing Trajectory Optimization with Denoising Autoencoders” that addresses some of the concerns of planning algorithms that exploit inaccuracies in their world models!  00:00:00 Intro to Harri and Curious AI System1/System 2 00:04:50 Background on model-based RL challenges from Tim 00:06:26 Other interesting research papers on model-based RL from Connor 00:08:36 Intro to Curious AI recent NeurIPS paper on model-based RL and denoising autoencoders from Yannic 00:21:00 Main show kick off, system 1/2 00:31:50 Where does the simulator come from? 00:33:59 Evolutionary priors 00:37:17 Consciousness 00:40:37 How does one build a company like Curious AI? 00:46:42 Deep Q Networks 00:49:04 Planning and Model based RL 00:53:04 Learning good representations 00:55:55 Typical problem Curious AI might solve in industry 01:00:56 Exploration 01:08:00 Their paper - regularizing trajectory optimization with denoising 01:13:47 What is Epistemic uncertainty 01:16:44 How would Curious develop these models 01:18:00 Explainability and simulations 01:22:33 How system 2 works in humans 01:26:11 Planning 01:27:04 Advice for starting an AI company 01:31:31 Real world implementation of planning models 01:33:49 Publishing research and openness  We really hope you enjoy this episode, please subscribe!  Regularizing Trajectory Optimization with Denoising Autoencoders: https://papers.nips.cc/paper/8552-regularizing-trajectory-optimization-with-denoising-autoencoders.pdf Pulp, Paper &amp; Packaging: A Future Transformed through Deep Learning: https://thecuriousaicompany.com/pulp-paper-packaging-a-future-transformed-through-deep-learning/ Curious AI: https://thecuriousaicompany.com/ Harri Valpola Publications: https://scholar.google.com/citations?user=1uT7-84AAAAJ&amp;hl=en&amp;oi=ao Some interesting papers around Model-Based RL: GameGAN: https://cdn.arstechnica.net/wp-content/uploads/2020/05/Nvidia_GameGAN_Research.pdf Plan2Explore: https://ramanans1.github.io/plan2explore/ World Models: https://worldmodels.github.io/ MuZero: https://arxiv.org/pdf/1911.08265.pdf PlaNet: A Deep Planning Network for RL: https://ai.googleblog.com/2019/02/introducing-planet-deep-planning.html Dreamer: Scalable RL using World Models: https://ai.googleblog.com/2020/03/introducing-dreamer-scalable.html Model Based RL for Atari: https://arxiv.org/pdf/1903.00374.pdf

In this episode of Machine Learning Street Talk, Tim Scarfe, Yannic Kilcher and Connor Shorten interviewed Harri Valpola, CEO and Founder of Curious AI. We continued our discussion of System 1 and System 2 thinking in Deep Learning, as well as miscellaneous topics around Model-based Reinforcement Learning. Dr. Valpola describes some of the challenges of modelling industrial control processes such as water sewage filters and paper mills with the use of model-based RL. Dr. Valpola and his collaborators recently published “Regularizing Trajectory Optimization with Denoising Autoencoders” that addresses some of the concerns of planning algorithms that exploit inaccuracies in their world models! 00:00:00 Intro to Harri and Curious AI System1/System 2 00:04:50 Background on model-based RL challenges from Tim 00:06:26 Other interesting research papers on model-based RL from Connor 00:08:36 Intro to Curious AI recent NeurIPS paper on model-based RL and denoising autoencoders from Yannic 00:21:00 Main show kick off, system 1/2 00:31:50 Where does the simulator come from? 00:33:59 Evolutionary priors 00:37:17 Consciousness 00:40:37 How does one build a company like Curious AI? 00:46:42 Deep Q Networks 00:49:04 Planning and Model based RL 00:53:04 Learning good representations 00:55:55 Typical problem Curious AI might solve in industry 01:00:56 Exploration 01:08:00 Their paper - regularizing trajectory optimization with denoising 01:13:47 What is Epistemic uncertainty 01:16:44 How would Curious develop these models 01:18:00 Explainability and simulations 01:22:33 How system 2 works in humans 01:26:11 Planning 01:27:04 Advice for starting an AI company 01:31:31 Real world implementation of planning models 01:33:49 Publishing research and openness We really hope you enjoy this episode, please subscribe! Regularizing Trajectory Optimization with Denoising Autoencoders: <a href="https://papers.nips.cc/paper/8552-regularizing-trajectory-optimization-with-denoising-autoencoders.pdf" rel="nofollow" target="_blank">https://papers.nips.cc/paper/8552-regularizing-trajectory-optimization-with-denoising-autoencoders.pdf</a> Pulp, Paper &amp; Packaging: A Future Transformed through Deep Learning: <a href="https://thecuriousaicompany.com/pulp-paper-packaging-a-future-transformed-through-deep-learning/" rel="nofollow" target="_blank">https://thecuriousaicompany.com/pulp-paper-packaging-a-future-transformed-through-deep-learning/</a> Curious AI: <a href="https://thecuriousaicompany.com/" rel="nofollow" target="_blank">https://thecuriousaicompany.com/</a> Harri Valpola Publications: <a href="https://scholar.google.com/citations?user=1uT7-84AAAAJ&amp;hl=en&amp;oi=ao" rel="nofollow" target="_blank">https://scholar.google.com/citations?user=1uT7-84AAAAJ&amp;hl=en&amp;oi=ao</a> Some interesting papers around Model-Based RL: GameGAN: <a href="https://cdn.arstechnica.net/wp-content/uploads/2020/05/Nvidia_GameGAN_Research.pdf" rel="nofollow" target="_blank">https://cdn.arstechnica.net/wp-content/uploads/2020/05/Nvidia_GameGAN_Research.pdf</a> Plan2Explore: <a href="https://ramanans1.github.io/plan2explore/" rel="nofollow" target="_blank">https://ramanans1.github.io/plan2explore/</a> World Models: <a href="https://worldmodels.github.io/" rel="nofollow" target="_blank">https://worldmodels.github.io/</a> MuZero: <a href="https://arxiv.org/pdf/1911.08265.pdf" rel="nofollow" target="_blank">https://arxiv.org/pdf/1911.08265.pdf</a> PlaNet: A Deep Planning Network for RL: <a href="https://ai.googleblog.com/2019/02/introducing-planet-deep-planning.html" rel="nofollow" target="_blank">https://ai.googleblog.com/2019/02/introducing-planet-deep-planning.html</a> Dreamer: Scalable RL using World Models: <a href="https://ai.googleblog.com/2020/03/introducing-dreamer-scalable.html" rel="nofollow" target="_blank">https://ai.googleblog.com/2020/03/introducing-dreamer-scalable.html</a> Model Based RL for Atari: <a href="https://arxiv.org/pdf/1903.00374.pdf" rel="nofollow" target="_blank">https://arxiv.org/pdf/1903.00374.pdf</a>

Harri Valpola: System 2 AI and Planning in Model-Based Reinforcement Learning

This is my interview with Harri Valpola. Harri is the CEO of Curious AI, a Helsinki-based 20 person startup that’s aiming to build a general artificial intelligence. Now that’s a pretty bold goal when Facebook’s head of AI research just recently said that in terms of general intelligence, we’re not even close to a rat. Harri is a respected researcher in the field of artificial intelligence. He knows his stuff. The previous company he founded, ZenRobotics, builds robots so that’s pretty cool too. We talk about what general artificial intelligence would be like, whether we should try to simulate the human brain or not and how curiosity is a great guide in life. Sponsor link:  Valuemotive: http://valuemotive.com Wunderdog: http://wunder.dog  Links for the episode:  Curious AI Harri on Wired magazine  Links for the podcast:  Mailing list: http://eepurl.com/P0dpf Facebook: https://www.facebook.com/bosslevelpodcast Twitter: https://twitter.com/bosslevelpod iTunes: https://itunes.apple.com/fi/podcast/boss-level/id1041885043 Website: http://www.bosslevelpodcast.com Sami Honkonen on Twitter: http://twitter.com/SamiHonkonen Sami's website: http://www.samihonkonen.com Sami's blog: http://blog.samihonkonen.com

This is my interview with Harri Valpola. Harri is the CEO of Curious AI, a Helsinki-based 20 person startup that’s aiming to build a general artificial intelligence. Now that’s a pretty bold goal when Facebook’s head of AI research just recently said that in terms of general intelligence, we’re not even close to a rat. Harri is a respected researcher in the field of artificial intelligence. He knows his stuff. The previous company he founded, ZenRobotics, builds robots so that’s pretty cool too. We talk about what general artificial intelligence would be like, whether we should try to simulate the human brain or not and how curiosity is a great guide in life. Sponsor link: <ul> <li>Valuemotive: <a href="http://valuemotive.com" rel="nofollow" target="_blank">http://valuemotive.com</a></li> <li>Wunderdog: http://wunder.dog</li> </ul> Links for the episode: <ul> <li><a href="https://thecuriousaicompany.com/" rel="nofollow" target="_blank">Curious AI</a></li> <li><a href="https://www.wired.co.uk/article/harri-valpola-curious-ai-artificial-intelligence-third-wave" rel="nofollow" target="_blank">Harri on Wired magazine</a></li> </ul> Links for the podcast: <ul> <li>Mailing list: <a href="http://eepurl.com/P0dpf" rel="nofollow" target="_blank">http://eepurl.com/P0dpf</a></li> <li>Facebook: <a href="https://www.facebook.com/bosslevelpodcast" rel="nofollow" target="_blank">https://www.facebook.com/bosslevelpodcast</a></li> <li>Twitter: <a href="https://twitter.com/bosslevelpod" rel="nofollow" target="_blank">https://twitter.com/bosslevelpod</a></li> <li>iTunes: <a href="https://itunes.apple.com/fi/podcast/boss-level/id1041885043" rel="nofollow" target="_blank">https://itunes.apple.com/fi/podcast/boss-level/id1041885043</a></li> <li>Website: <a href="http://www.bosslevelpodcast.com" rel="nofollow" target="_blank">http://www.bosslevelpodcast.com</a></li> <li>Sami Honkonen on Twitter: <a href="http://twitter.com/SamiHonkonen" rel="nofollow" target="_blank">http://twitter.com/SamiHonkonen</a></li> <li>Sami's website: <a href="http://www.samihonkonen.com" rel="nofollow" target="_blank">http://www.samihonkonen.com</a></li> <li>Sami's blog: <a href="http://blog.samihonkonen.com" rel="nofollow" target="_blank">http://blog.samihonkonen.com</a></li></ul>

Harri Valpola on third generation artificial intelligence