Видео

please help me get the problem set

save 53:00

Still usefull in 2024

Imagine a whole college failing to beat one man.

The scene graph + GNN approach could be really, really easily replaced by vision LLMs. The address of vision LLM at 24:18 wasn't really fair.

Do we need data science concept to understand this course or we can understand having command on python?

Thank you!

Good sh*t!

Question wasn't even answered she was just ranting

I will build something extraordinary in this discipline of Machine Learning and Chris will have a great bit of contribution in it. I feel it right now, he is that good.

what is the textbook of this course?

Grazie Stefano e Stanford

35:40 He said my dad’s name! 37:21 He and a tons of random people clapped for my dad!! @johanderuiter

this is straightup mindf***

One of the few, or maybe the only sensible leader in the western world.

I keep watching this. This is my fourth time.

By YouSum Live 00:02:00 Importance of studying change itself 00:03:34 Predicting future trajectory in AI research 00:08:01 Impact of exponentially cheaper compute power 00:10:10 Balancing structure and freedom in AI models 00:14:46 Historical analysis of Transformer architecture 00:17:29 Encoder-decoder architecture in Transformers 00:19:56 Cross-attention mechanism between decoder and encoder 00:20:10 All decoder layers attend to final encoder layer 00:20:50 Transition from sequence-to-sequence to classification labels 00:21:30 Simplifying problems for performance gains 00:22:24 Decoder-only architecture for supervised learning 00:22:59 Self-attention mechanism handling cross-attention 00:23:13 Sharing parameters between input and target sequences 00:24:03 Encoder-decoder vs. decoder-only architecture comparison 00:26:59 Revisiting assumptions in architecture design 00:33:10 Bidirectional vs. unidirectional attention necessity 00:35:16 Impact of scaling efforts on AI research By YouSum Live

14:11 パースペクティブが重要。同じことをしてても結果が変わる 24:03 既存事業とイノベーション 28:33 変化できなければ死ぬ 29:06 事業同士の共食い 31:21 利益は競争で決まって、参入するかはリソースと需給で決める 45:55 良いCEOは曖昧さを受け入れる 55:27 成功したものを捨てて再発明する必要

I think someone might need to redo the evaluation lecture/curriculum since it has become an alphabet soup lately.

Anyone starting in june 2024, Reply to my comment, Lets do it together...👍

never thought in a million years that i would be learning AI from a Stanford professor as a 13 year old

וT•N•G•× ×[..]× bina sarana informatika

🖤🤗

❤️🤍

okay so the superscript i, ( 1 to m) represents the number of features, right? Because here m = 2 and I don't understand why m = # training examples

20:24 How will you get negative values for J(Ø) as you’ve defined it, i.e. as a square of the error?

Very expensive, but I wanna learn

That's an incredible talk, it should get more attention. I think web3 should be called data centric web

i do not understand why some people explain e.g. what is a transaction AFTER they used the word before. Turn it around man. Happy to was not part of his stanford bla bla bla.

Incredible talk.

Watching this video again now, first watched it back in 2017 I think and it was because Intel looking for new CEO and there were a lot rumors that Jensen was considered for the job, and we all know Intel was king in this chip business, and I was curious about this guy, after watching this I was amazed about everything he said and I remember checking NVDA price around $40 back then I think, regretfully didn’t buy the stock.

12 years later= this is a brilliant talk- every word resonates, wish I had watched this video so much earlier

"In technology, reinventing the company every 10 years is almost a necessity", great insight- now trying to find companies who did not and are still around- Intel, Cisco come to my mind, others please- reply

No PyTorch embedding?

where to find slides ?

now nvidia has 3.34 trillion net worth

These lectures have been really awesome so far, congratulations!

show us the git ffs

Amazing talk! I was wondering why the field has moved closer to decoder-only models lately and whether there's an explanation to it.

I was having a hard week.. because I bought a course, and I was not able to understand things in deep. I came on u tube, and this came on my screen.... bro I felt like god just sent me something to make me happy. This is best thing happened. Thank you for the course.

this video is so underrated, her explanations were so clear

Thanks for the lectures ❤

When we talk about Monte Carlo, when we evaluate V^(pi)(s), in order to pickout the best policy, we have to evaluate all possible policy ? and then pick the best one? Im a bit confused on how to do control here thanks :D

This created new synapses in my brain-cells graph

Poorly explained

Great lecture and insights on LLM.

Great presentation -- thank you for making it available. I do think it's funny -- and maybe a reflection on Andrew's humility -- that you can go through many YT videos about him and not hear his full name pronounced. What other well-known communicators can you say that about?

12:41 here he mentions that traditional ML classifiers have a disadvantage of only being able to draw linear decision boundaries, SVM for example can be used to draw non-linear decision boundaries if used along with the kernel trick right?

Awesome series of lectures up till now.

This lecture had a lot of gaps.