Yannic Kilcher Videos (Audio Only)

#nlp #sparsity #transformers This video is an interview with Barret Zoph and William Fedus of Google Brain about Sparse Expert Models. Sparse Expert models have been hugely successful at distributing parts of models, mostly Transformers, across large array of machines and use a routing function to effectively route signals between them. This means that even though these models have a huge number o...

#neuralsearch #interview #google This is an interview with the authors Yi Tay and Don Metzler. Paper Review Video: https://youtu.be/qlB0TPBQ7YY Search engines work by building an index and then looking up things in it. Usually, that index is a separate data structure. In keyword search, we build and store reverse indices. In neural search, we build nearest-neighbor indices. This paper does somethi...

#dsi #search #google Search engines work by building an index and then looking up things in it. Usually, that index is a separate data structure. In keyword search, we build and store reverse indices. In neural search, we build nearest-neighbor indices. This paper does something different: It directly trains a Transformer to return the ID of the most relevant document. No similarity search over em...

#mlnews #palm #dalle2 Google releases PaLM and OpenAI releases DALL-E 2 (and more news). Sponsor: Weights & BIases Start here: https://wandb.me/yannic Thumbnail credit: DALL-E 2 via Sam Altman OUTLINE 0:00 - Street interview w/ random stranger 2:25 - Intro 2:50 - PaLM - Google's 540B Pathways Language Model 7:50 - Sponsor: Weights & Biases 9:10 - OpenAI releases DALL-E 2 12:05 - Open Sourc...

#aiart #deeplearning #clip Since the release of CLIP, the world of AI art has seen an unprecedented level of acceleration in what's possible to do. Whereas image generation had previously been mostly in the domain of scientists, now a community of professional artists, researchers, and amateurs are sending around colab notebooks and sharing their creations via social media. How did this happen? Wh...

#reinforcementlearning #ai #explained This is an interview with Jesse Mu, first author of the paper. Original Paper Review: https://youtu.be/NeGJAUSQEJI Exploration is one of the oldest challenges for Reinforcement Learning algorithms, with no clear solution to date. Especially in environments with sparse rewards, agents face significant challenges in deciding which parts of the environment to exp...

#reinforcementlearning #ai #explained Exploration is one of the oldest challenges for Reinforcement Learning algorithms, with no clear solution to date. Especially in environments with sparse rewards, agents face significant challenges in deciding which parts of the environment to explore further. Providing intrinsic motivation in form of a pseudo-reward is sometimes used to overcome this challeng...

#mlnews #gpt3 #pathways Your updates on the latest and greatest from the depths of Machine Learning! Sponsor: Weights & Biases https://wandb.me/yannic OUTLINE: 0:00 - Intro 0:15 - Weights & Biases Report about Reports 2:45 - GPT-3 learns to edit 6:30 - Make-A-Scene: Text-to-Image with Human Priors 8:00 - Pathways: Google's new High-Performance ML scheduler 10:45 - DouBlind: Open Peer-Revie...

#nlp #gpt3 #prompt This is an interview with the authors of this work, Aman Madaan and Niket Tandon. Large language models such as GPT-3 have enabled many breakthroughs and new applications recently, but they come with an important downside: Training them is very expensive, and even fine-tuning is often difficult. This paper presents an adaptive method to improve performance of such models after d...

#nlp #gpt3 #prompt Large language models such as GPT-3 have enabled many breakthroughs and new applications recently, but they come with an important downside: Training them is very expensive, and even fine-tuning is often difficult. This paper presents an adaptive method to improve performance of such models after deployment, without ever changing the model itself. This is done by maintaining a m...

#deeplearning #nlp #sampling This is an interview with first author Clara Meister. Paper review video hereé https://youtu.be/_EDr3ryrT_Y Modern language models like T5 or GPT-3 achieve remarkably low perplexities on both training and validation data, yet when sampling from their output distributions, the generated text often seems dull and uninteresting. Various workarounds have been proposed, suc...

#deeplearning #nlp #sampling Modern language models like T5 or GPT-3 achieve remarkably low perplexities on both training and validation data, yet when sampling from their output distributions, the generated text often seems dull and uninteresting. Various workarounds have been proposed, such as top-k sampling and nucleus sampling, but while these manage to somewhat improve the generated samples,...

#blip #interview #salesforce Paper Review Video: https://youtu.be/X2k7n4FuI7c Sponsor: Assembly AI https://www.assemblyai.com/?utm_sourc... This is an interview with Junnan Li and Dongxu Li, authors of BLIP and members of Salesforce research. Cross-modal pre-training has been all the rage lately in deep learning, especially training vision and language models together. However, there are a number...

#blip #review #ai Cross-modal pre-training has been all the rage lately in deep learning, especially training vision and language models together. However, there are a number of issues, such as low quality datasets that limit the performance of any model trained on it, and also the fact that pure contrastive pre-training cannot be easily fine-tuned for most downstream tasks. BLIP unifies different...

#mlnews #gtc22 #ithaca GTC Registration Link: https://ykilcher.com/gtc Your regular updates on what's going on in the ML world! OUTLINE: 0:00 - Intro 0:20 - Register to Nvidia GTC and win a 3090! 4:15 - DeepMind's Ithaca deciphers Lost Ancient Texts 6:45 - Drug discovery model turns toxic 10:00 - Gary Marcus: Deep Learning is hitting a wall 19:40 - GopherCite: Backing up answers with citations 22:...

#multitasklearning #biology #neuralnetworks This is an interview with the paper's authors: Abhiram Iyer, Karan Grewal, and Akash Velu! Paper Review Video: https://youtu.be/O_dJ31T01i8 Check out Zak's course on Graph Neural Networks (discount with this link): https://www.graphneuralnets.com/p/int... Catastrophic forgetting is a big problem in mutli-task and continual learning. Gradients of differen...

#multitasklearning #biology #neuralnetworks Catastrophic forgetting is a big problem in mutli-task and continual learning. Gradients of different objectives tend to conflict, and new tasks tend to override past knowledge. In biological neural networks, each neuron carries a complex network of dendrites that mitigate such forgetting by recognizing the context of an input signal. This paper introduc...

#deeplearning #objectdetection #outliers An interview with the authors of "Virtual Outlier Synthesis". Watch the paper review video here: https://youtu.be/i-J4T3uLC9M Outliers are data points that are highly unlikely to be seen in the training distribution, and therefore deep neural networks have troubles when dealing with them. Many approaches to detecting outliers at inference time have been pro...

#vos #outliers #deeplearning Sponsor: Assembly AI Check them out here: https://www.assemblyai.com/?utm_sourc... Outliers are data points that are highly unlikely to be seen in the training distribution, and therefore deep neural networks have troubles when dealing with them. Many approaches to detecting outliers at inference time have been proposed, but most of them show limited success. This pape...

#deepmind #rl #society This is an in-depth paper review, followed by an interview with the papers' authors! Society is ruled by norms, and most of these norms are very useful, such as washing your hands before cooking. However, there also exist plenty of social norms which are essentially arbitrary, such as what hairstyles are acceptable, or what words are rude. These are called "silly rules". Thi...

#openai #math #imo This is an interview with Stanislas Polu, research engineer at OpenAI and first author of the paper "Formal Mathematics Statement Curriculum Learning". Watch the paper review here: https://youtu.be/lvYVuOmUVs8 OUTLINE: 0:00 - Intro 2:00 - How do you explain the big public reaction? 4:00 - What's the history behind the paper? 6:15 - How does algorithmic formal math work? 13:10 -...

#openai #math #imo Formal mathematics is a challenging area for both humans and machines. For humans, formal proofs require very tedious and meticulous specifications of every last detail and results in very long, overly cumbersome and verbose outputs. For machines, the discreteness and sparse reward nature of the problem presents a significant problem, which is classically tackled by brute force...

Updates on what's going on in the ML world! Check out w&b's alerts feature: https://wandb.me/yannic OUTLINE: 0:00 - Intro 0:20 - Sponsor: Weights & Biases 2:35 - DeepMind uses Reinforcement Learning to control nuclear fusion 4:35 - Google responds to carbon emission estimates 8:40 - Yann LeCun proposes new architecture for world models 11:05 - Fruit fly neurons may perform multiplication 1...

An interview with the creators of AlphaCode! Paper review video here: https://youtu.be/s9UAOmyah1A OUTLINE: 0:00 - Intro 1:10 - Media Reception 5:10 - How did the project go from start to finish? 9:15 - Does the model understand its own code? 14:45 - Are there plans to reduce the number of samples? 16:15 - Could one do smarter filtering of samples? 18:55 - How crucial are the public test cases? 21...

#ai #alphacode #deepmind AlphaCode is an automated system that can solve competitive programing exercises. The authors found an interesting combination of language models, large-scale sampling, and clever techniques to filter and subsequently cluster the resulting programs, which lets the system perform on the level of an average competitor in real competitions. In this video, we take a deep dive...