Most learning is superficial and fades quickly. This podcast will equip you to move to learning that is durable because it is deep. Deep learning lasts because it respects the way the brain works. Inquiring minds want to know "how" and "why"—not just what!
Long-form technical deep dives with Andrej Karpathy and Justin Johnson
Welcome to The Deep Learning Crowd Podcast. We talk about individual journey's in the world of AI and discuss topics around Deep Learning. We discover first hand from our guests about some of the most interesting applications their companies are using right now.
Find me on Github/Twitter/Kaggle @SamDeepLearning. Find me on LinkedIn @SamPutnam. This Podcast is supported by Enterprise Deep Learning | Cambridge/Boston | New York City | Hanover, NH | http://www.EnterpriseDeepLearning.com. Contact: Sam@EDeepLearning.com, 802-299-1240, P.O. Box 863, Hanover, NH, USA, 03755. We move deep learning to production. I teach the worldwide Deploying Deep Learning Masterclass at http://www.DeepLearningConf.com in NYC regularly and am a Deep Learning Consultant ser ...
Most AI research today is done by elite universities and corporate labs. The pursuit of science and creativity should be a viable project for any person, from any walk of life, who is excited by that feeling of mystery and building something that grows. chloe is an end to end neural network chatbot written in PyTorch based on the transformer. Accomplishing goals through conversation is a task we can relate to, chatbots are an ideal agent through which to connect new research to our current u ...
Real thinking involves chewing on a compelling question. Powerful answers invoke cause and effect. Those answers have immediate implications as well as long-term consequences and both of those lead to actions.저자 Michael Gray
Thinking is driven by questions. Questions are answered through the interaction of necessary assumptions, a fact base to which thinking is accountable, and—most of all—a conceptual framework. Conceptual frameworks harness the power of patterns to look for parallels that leverage past learning to solve present problems.…
Every area of human endeavor is an outworking of a way of thinking. We all default to a particular way of thinking, usually without recognizing it. This episode is designed to help you be intentional about where your conclusions are coming from. The core of your thinking is the combination of point-of-view, motivation (what you are trying to accomp…
Great conversations are driven by empathetic listening that results in good questions. Good questions encourage the other person to open up and share. Good questions give the questioner an opportunity to learn from another person’s life experience. In this episode my guest, Laura from Asia, characterizes being a good questioner as showing hospitali…
Principles are the power tools of thinking. Learn how to construct principles that satisfy your need for things to make sense. When something makes sense, you won’t have to struggle to remember it or to use it in problem-solving.저자 Michael Gray
Patterns don’t simply emerge on their own. They are the fruit of “creative scrabbling” through Subsidiary-Focal Integration (SFI). This episode will operationalize SFI.저자 Michael Gray
"Creativity is just connecting things" was Steve Jobs summary. Learn how to create transformative patterns through connecting concepts.저자 Michael Gray
Finding a pattern in a collection of specifics through induction is the essence of the transformative insight that we call the "aha" moment. Learn how to increase the frequency and wattage of your lightbulb moments.저자 Michael Gray
Pattern-recognition is the most extraordinary capability of the human brain. We use induction to formulate these regularities as our concept categories. This episode will demystify inductive reasoning to enable better thinking.저자 MICHAEL GRAY
The inability to remember much of anything before our third year of life shows us what must happen to enable lasting learning as adults.저자 Michael Gray
The purpose of exploration is to shake things up—to encounter new ideas which may initially be puzzling but which may hold the key to answering the compelling questions you care about.저자 Michael Gray
Focused exploration is the third disposition of Learner's Mind. Exploration need not be time inefficient. We explore the story of Watson and Crick in their pursuit of the structure of DNA. This is an example of the power of exploration to quickly solve a problem that appeared insoluble.저자 Michael Gray
D
Deep and Durable Learning

1
Developing Learner's Mind: Purposeful Persistent Perception
23:14
23:14
나중에 재생
나중에 재생
리스트
좋아요
좋아요
23:14
Developing Learner's Mind requires an openness to and a curiosity about the world around you. Curiosity is manifested by a willingness to pay attention to what exploration uncovers followed by the cultivation of perception through which you really start to listen.저자 Michael Gray
Learner's mind begins with curiosity and curiosity begins with paying attention to the world around you. Paying attention broadens options and leads to better conclusions. Learn how to cultivate attention.저자 Michael Gray
Caring about students means prioritizing their needs as learners. This includes centering courses on a manageable number of core concepts as well as considering prior knowledge and what role the course will play in their future learning. Students need to participate actively without fear that their contributions will cause them to be judged.…
Adam is the CEO of KONUX, a leading German AI scale-up that is transforming railway operations for a sustainable future. Adam joined KONUX as CEO to help continue their success and help elevate them to a new level. As a company, they focus heavily on AI and IoT and have raised hundreds of millions. In our chat today we talk about Adams's background…
D
Deep and Durable Learning

1
Wrestling with Rusting: Rooting Courses in a Compelling Question
33:42
33:42
나중에 재생
나중에 재생
리스트
좋아요
좋아요
33:42
Great courses are designed to engage the curiosity of students. In this episode Dr. David Gardenghi explains how he roots his chemistry for engineers course in the central engineering problem of corrosion.저자 Michael Gray, David Gardenghi
Dr. Amy Tuck teaches the functioning of the human immune system through a logically connected narrative. Immunity is due to an intricately interconnected multilevel system designed to assure our survival in a hostile world. Listen as Dr. Tuck shares her strategy for creating deep and durable learning that answers real life application questions.…
The Summer Institute in Teaching Science (SITS) is a four summer program of ten weeks per summer. The aim of SITS is to transform discipline-specific knowledge experts into effective teachers who are able to take their knowledge apart to create an optimal path for learning. In this episode recent SITS graduate, Dr. David McKinney, reflects on his p…
The 7C's of Cognition: Curiosity, Conceptualization, Connectivity, Constrained Capacity, Chunking, Consolidation, and Creativity. Creating Coherence.저자 Michael Gray
Florian is an 18-year-old entrepreneur and founder of Fast AI Movies. Despite Florian's age, he is advancing very fast in AI and business. At 13 years old he taught himself programming and then AI. In our chat today we talk about Florian's unique journey into his passion for AI. We were able to really discover the mindset of a young entrepreneur an…
Your brain is never off duty. While your body sleeps your brain actively and purposefully retrieves prior knowledge to sharpen it and harmonize it with new learning through the process of consolidation.저자 Michael Gray
Learning is sabotaged when the cognitive load of the task is too high as well as when it is too low. Learn how to optimize cognitive load.저자 Michael Gray
Luisa is the Co-Founder & Space Systems Engineer for the start up Vyoma in Munich. Luisa has plans to build and operate an innovative, satellite-based observation system for space debris with the power of AI. In our chat today we discover their journey so far and the unique drive of an employee first business model. We talk computer vision and mach…
Chunking removes the obstacle of our extremely limited working memory by leveraging the hard-wired pattern making of the brain to create networks of powerful, logically-linked concepts.저자 Michael Gray
The brain ruthlessly discards information, but relentless forgetting allows us to remain flexible as learners.저자 Michael Gray
New concepts don’t float around but are logically negotiated into an appropriate place in our mental framework of concepts.저자 Michael Gray
The process of concept (idea) formation in the brain is the universal mechanism for learning with understanding.저자 Michael Gray
Curiosity is essential and is the foundation of learning.
T
The Deep Learning Crowd

1
Firas Lethaus - Co-Founder of Volkswagen AI Team & Head of Deep Learning
29:55
29:55
나중에 재생
나중에 재생
리스트
좋아요
좋아요
29:55
Firas is the Co-Founder of Volkswagens AG AI Team and Head of Deep Learning at the Expert Centre in Munich. A gentlemen who has dedicated his life to Data and AI. Firas has worked for multiple big players in the industry such as Audi, DLR & now Volkswagen. Today we get to learn from Firas's perspective on the current state of AI, effectively buildi…
Knowledge of the human brain enables us to optimize learning. Exploring the 7 C's of Cognition.
T
The Deep Learning Crowd

1
Alexey Grigorev - Principle Data Scientist & Founder of DataTalks.Club
44:47
44:47
나중에 재생
나중에 재생
리스트
좋아요
좋아요
44:47
Alexey Grigorev is a Principle Data Scientist at OLX Group and founder of DataTalks.Club with the vision to be the go to community for everything data. Alexey is a true expert and today he channels his knowledge into topics around the booming area of MLOps. Alexey shares advice for young and aspiring Data Scientists and we will also find out what h…
There is currently a mood of dismissal and even denigration of experts. "Strongly held opinions should be granted the status of facts," say some laypeople. This episode explains why learning requires transformative expertise outside oneself.저자 Mike Gray
T
The Deep Learning Crowd

1
The Deep Learning Crowd - Ritesh Kanjee - CEO & Founder - Augmented Startups
52:30
52:30
나중에 재생
나중에 재생
리스트
좋아요
좋아요
52:30
Ritesh Kanjee is a computer vision expert who has channelled his skillset into training and mentoring which led him to create Augmented Startups. Ritesh prides himself in being a Computer Vision and AI content creator and has amassed nearly 100,000 subscribers on YouTube. His company, Augmented Startups are at the forefront of Artificial Intelligen…
Educational curricula are obsessed with information. Rigor means more information is covered. Learn how to overcome the systemic prioritization of information over real knowing.저자 Mike Gray
Knowledge requires a human knower. A person can be said to know something when they believe it to be true and they can justify that belief using reason and evidence.저자 Mike Gray
Information and Knowledge are not the same thing. Learning is much more than collecting information! Join me as we explore how data is used to create information and information is used by human knowers to create knowledge.저자 Mike Gray
T
The Deep Learning Crowd

1
The Deep Learning Crowd - Nikolaas Steenbergen - Co-founder - 030Solutions.
34:47
34:47
나중에 재생
나중에 재생
리스트
좋아요
좋아요
34:47
In todays episode I will be speaking with Nikolaas Steenbergen, Co-Founder of 030Solutions. A brand new start-up based in Berlin which uses laser and image data from mobile mapping vehicles or drones, using a smart AI software to survey, record and analyze data for mapping distributed infrastructure such as electricity distribution networks. Nikola…
Information is readily available, but so is misinformation. Don't believe everything you see on your screen. Actively resist your native confirmation bias that shops for prepackaged opinions that square with your preconceptions/tribe. Learning requires intellectual humility that assumes a willingness to be taught.…
T
The Deep Learning Crowd

1
The Deep Learning Crowd - Leon Szeli - Co-Founder & CEO - Presize
40:58
40:58
나중에 재생
나중에 재생
리스트
좋아요
좋아요
40:58
In this episode I will be speaking with Leon Szeli, CEO and Co-Founder of Presize, an exciting Start-up in Munich which has already created huge waves in the eCommerce and AI world. I will be getting to know Leon and how his progressed so quickly but effectively in the last 2 years. We will be covering his unique product in the field of computer vi…
T
The Deep Learning Crowd

1
The Deep Learning Crowd – Dr. Holger Bartel – CEO & Founder
39:38
39:38
나중에 재생
나중에 재생
리스트
좋아요
좋아요
39:38
I am very excited to announce the 1st episode of The Deep Learning Crowd Podcast - where I speak to Dr. Holger Bartel, CEO & Founder of RealRate In this episode, we get to know Holger and what RealRate do in the world of AI. We break down XAI and the concerns we are facing, also understanding Hybrid AI and the importance of it. Lastly we dive into …
저자 Mike Gray
The actual paper for DALL-E was released only a few days after we published our episode 2, so in this episode we re-visit DALL-E in its full published glory. Joining us as a special guest for this episode is Aditya Ramesh from OpenAI, the lead author of DALL-E. DALL-E blog post: https://openai.com/blog/dall-e/ DALL-E paper: https://arxiv.org/abs/21…
D
Deep Learning Deep Dive


1
Episode #2: DALL-E and friends in image generation
1:51:27
1:51:27
나중에 재생
나중에 재생
리스트
좋아요
좋아요
1:51:27
Andrej Karpathy and Justin Johnson deep dive into OpenAI's DALL-E and use it as an anchor point to recurse into some of the recent work in AI on image generation. Approximate agenda: DALL-E Blog Post: https://openai.com/blog/dall-e/ ImageGPT https://openai.com/blog/image-gpt/ VQ-VAE https://arxiv.org/abs/1711.00937 VQ-VAE-2 https://arxiv.org/abs/19…
Here is give a 4 min summary of Andrew Yang and why his ideas are correct--- Support this podcast: https://anchor.fm/ChloeRobotics/support
D
Deep Learning: Zero to One


1
Art Generation - Facebook AI Research, Google DeepDream, and Ruder's Style Transfer for Video - Deep Learning: Zero to One
7:49
7:49
나중에 재생
나중에 재생
리스트
좋아요
좋아요
7:49
Justin Johnson, now at Facebook, wrote the original Torch implementation of the Gatys 2015 paper, which combines the content of one image and the style of another image using convolutional neural networks. Manuel Ruder’s newer 2016 paper transfers the style of one image to a whole video sequence, and it uses a computer vision technique called optic…
D
Deep Learning: Zero to One


1
Music Generation - Google Magenta Best Demo NIPS 2016 LSTM RNN - Deep Learning: Zero to One
8:29
8:29
나중에 재생
나중에 재생
리스트
좋아요
좋아요
8:29
I talk through generating 10 melodies, two of which I play at the conclusion using a model trained on thousands of midi examples contained in a .mag Magenta file bundle. I used the Biaxial RNN (https://github.com/hexahedria/biaxial-rnn-music-composition) by a student named Daniel Johnson and the Basic RNN (https://github.com/tensorflow/magenta/tree…
D
Deep Learning: Zero to One


1
Image Generation - Google DeepMind paper with TensorFlow - Deep Learning: Zero to One
5:01
5:01
나중에 재생
나중에 재생
리스트
좋아요
좋아요
5:01
I talk through generating an image of IRS tax return characters using a model trained on the IRS tax return dataset - NMIST. The authors trained for 70 hours on 32 GPUs. I used unconditioned image generation to create an image in 6 hours on my MacBook Pro CPU. I used the TensorFlow implementation of Conditional Image Generation with PixelCNN Decode…