![StatQuest with Josh Starmer](/img/default-banner.jpg)
- Видео 278
- Просмотров 66 155 167
StatQuest with Josh Starmer
США
Добавлен 23 май 2011
Statistics, Machine Learning and Data Science can sometimes seem like very scary topics, but since each technique is really just a combination of small and simple steps, they are actually quite simple. My goal with StatQuest is to break down the major methodologies into easy to understand pieces. That said, I don't dumb down the material. Instead, I build up your understanding so that you are smarter.
Contact, Video Index, Etc: statquest.org
Contact, Video Index, Etc: statquest.org
Human Stories in AI: Amy Finnegan
In this episode we have special guest Dr. Amy Finnegan, the Deputy Director of Data Science at IntraHealth International. Amy is a demographer and data scientist with over 10 years of experience working in global health, development, and data science in emerging economies on four continents. Amy is also an adjunct faculty at Duke University’s Global Health Institute, and before these jobs, she was a research scholar at Duke.
If you'd like to support StatQuest, please consider...
Patreon: www.patreon.com/statquest
...or...
RUclips Membership: ruclips.net/channel/UCtYLUTtgS3k1Fg4y5tAhLbwjoin
...buying my book, a study guide, a t-shirt or hoodie, or a song from the StatQuest store...
statquest.org...
If you'd like to support StatQuest, please consider...
Patreon: www.patreon.com/statquest
...or...
RUclips Membership: ruclips.net/channel/UCtYLUTtgS3k1Fg4y5tAhLbwjoin
...buying my book, a study guide, a t-shirt or hoodie, or a song from the StatQuest store...
statquest.org...
Просмотров: 3 362
Видео
Human Stories in AI: Xavier Moyá
Просмотров 3,7 тыс.21 день назад
In this episode we have special guest Xavier Godoy, the director of customer experience and automation at HBX Group in Mallorca, Spain. Xavier has had a career driven by curiosity and a desire to learn more while simultaneously making sure that customer satisfaction is always is the focus of his efforts. If you'd like to support StatQuest, please consider... Patreon: www.patreon.com/statquest ....
Human Stories in AI: Tommy Tang
Просмотров 4,1 тыс.Месяц назад
In this episode we have special guest Tommy Tang, the Director of Computational Biology at Immunitas Therapeutics. Tommy is a computational biologist with over ten years of computational experience and six years' wet lab experience committed to reproducible research and open science. At Immunitas Therapeutics, Tommy employs a single-cell sequencing platform to dissect the biology of immune cell...
Human Stories in AI: Simon Stochholm
Просмотров 4 тыс.Месяц назад
In this episode we have special guest Simon Stochhom, a lecturer at UCL in Denmark. Simon applies machine learning, especially deep learning, to images, video and time series in wide variety of settings. And by “wide variety”, I really mean it. Simon is fearless when it comes to seizing opportunities that come up and somehow turns them all into success stories. If you'd like to support StatQues...
Log_e Song - Official Lyric Video
Просмотров 5 тыс.2 месяца назад
Check out the track on Spotify: open.spotify.com/track/4OcFh2yFOTUqzmjjwJF5QY When I first started making StatQuest videos it never dawned on me that people would try to re-do my math on their own. I was also new to explaining things and just assumed that everyone already knew that, in statistics and machine learning, when you use the log, you use base 'e'. Big rookie mistake! Ever since then, ...
Human Stories in AI: Brian Risk@devra.ai
Просмотров 4,5 тыс.2 месяца назад
In this episode we have special guest Brian Risk, a multi-talented data scientist and the President and Founder of devra.ai, a company specializing in automated coding. Brian is also a great personal friend of mine and an amazing musician. If you'd like to support StatQuest, please consider... Patreon: www.patreon.com/statquest ...or... RUclips Membership: ruclips.net/channel/UCtYLUTtgS3k1Fg4y5...
The matrix math behind transformer neural networks, one step at a time!!!
Просмотров 47 тыс.2 месяца назад
Transformers, the neural network architecture behind ChatGPT, do a lot of math. However, this math can be done quickly using matrix math because GPUs are optimized for it. Matrix math is also used when we code neural networks, so learning how ChatGPT does it will help you code your own. Thus, in this video, we go through the math one step at a time and explain what each step does so that you ca...
Human Stories in AI: Fabio Urbina
Просмотров 4,3 тыс.2 месяца назад
In this episode we have special guest Fabio Urbina, an Associate Director at Collaborations Pharmaceuticals. Fabio combines computational tools and machine learning with classical small-molecule, molecular, and cell biology techniques to address previously-difficult to probe scientific problems. Specifically, Fabio finds solutions to drug discovery with machine learning. If you'd like to suppor...
Human Stories in AI: Khushi Jain
Просмотров 8 тыс.3 месяца назад
In this episode we have special guest Kushi Jain, who works in Data Analytics Development at John Deere and in the Master’s in Computer Science - Data Science at the University of Illinois. Having recently graduated with her bachelor’s, Kushi participated in the data science club and also completed several internships at John Deere. If you'd like to support StatQuest, please consider... Patreon...
Human Stories in AI: Achal Dixit
Просмотров 8 тыс.3 месяца назад
In this episode we have special guest Achal Dixit, a Data Scientist at Delhivery, the largest fully integrated logistics services in India. Achal solves problems using Data, statistics, and machine learning with a focus on business and people. Before Delhivery, Achal was a Business Technology Analyst at ZS. And before that, Achal was a research assistant at Imperial College London. If you'd lik...
Human Stories in AI: Rick Marks
Просмотров 8 тыс.4 месяца назад
In this episode we have special guest Rick Marks, a professor at the University of North Carolina Chapel Hill School of Data Science and Society. Before UNC, Rick was a director at Google's Advanced Technology and Projects group, exploring new interaction approaches for ambient computing environments. And before that, Rick founded the PlayStation Magic Lab at PlayStation R&D. If you'd like to s...
Essential Matrix Algebra for Neural Networks, Clearly Explained!!!
Просмотров 45 тыс.6 месяцев назад
Although you don't need to know matrix algebra to understand the ideas behind neural networks, if you want to code them or read the latest manuscripts about the field, then you'll need to understand matrix algebra. This video teaches the essential topics in matrix algebra and shows how a neural network can be written as a matrix equation, and then shows how understand PyTorch documentation, err...
Word Embedding in PyTorch + Lightning
Просмотров 31 тыс.7 месяцев назад
Word embedding is the first step in lots of neural networks, including Transformers (like ChatGPT) and other state of the art models. Here we learn how to code a stand alone word embedding network from scratch and with nn.Linear. We then learn how to load and use pre-trained word embedding values with nn.Embedding. NOTE: This StatQuest assumes that you are already familiar with Word Embedding, ...
The Golden Play Button, Clearly Explained!!!’
Просмотров 24 тыс.8 месяцев назад
The Golden Play Button is usually super confusing. In this video, we break it down and walk you through it one-step-at-a-time. By the end of this StatQuest, you'll completely understand The Golden Play Button.
Decoder-Only Transformers, ChatGPTs specific Transformer, Clearly Explained!!!
Просмотров 103 тыс.9 месяцев назад
Decoder-Only Transformers, ChatGPTs specific Transformer, Clearly Explained!!!
Transformer Neural Networks, ChatGPT's foundation, Clearly Explained!!!
Просмотров 603 тыс.11 месяцев назад
Transformer Neural Networks, ChatGPT's foundation, Clearly Explained!!!
Attention for Neural Networks, Clearly Explained!!!
Просмотров 233 тыс.Год назад
Attention for Neural Networks, Clearly Explained!!!
Sequence-to-Sequence (seq2seq) Encoder-Decoder Neural Networks, Clearly Explained!!!
Просмотров 166 тыс.Год назад
Sequence-to-Sequence (seq2seq) Encoder-Decoder Neural Networks, Clearly Explained!!!
Word Embedding and Word2Vec, Clearly Explained!!!
Просмотров 268 тыс.Год назад
Word Embedding and Word2Vec, Clearly Explained!!!
The AI Buzz, Episode #5: A new wave of AI-based products and the resurgence of personal applications
Просмотров 11 тыс.Год назад
The AI Buzz, Episode #5: A new wave of AI-based products and the resurgence of personal applications
CatBoost Part 2: Building and Using Trees
Просмотров 17 тыс.Год назад
CatBoost Part 2: Building and Using Trees
CatBoost Part 1: Ordered Target Encoding
Просмотров 30 тыс.Год назад
CatBoost Part 1: Ordered Target Encoding
The AI Buzz, Episode #4: ChatGPT + Bing and How to start an AI company in 3 easy steps.
Просмотров 7 тыс.Год назад
The AI Buzz, Episode #4: ChatGPT Bing and How to start an AI company in 3 easy steps.
One-Hot, Label, Target and K-Fold Target Encoding, Clearly Explained!!!
Просмотров 45 тыс.Год назад
One-Hot, Label, Target and K-Fold Target Encoding, Clearly Explained!!!
The AI Buzz, Episode #3: Constitutional AI, Emergent Abilities and Foundation Models
Просмотров 5 тыс.Год назад
The AI Buzz, Episode #3: Constitutional AI, Emergent Abilities and Foundation Models
Mutual Information, Clearly Explained!!!
Просмотров 81 тыс.Год назад
Mutual Information, Clearly Explained!!!
Cosine Similarity, Clearly Explained!!!
Просмотров 79 тыс.Год назад
Cosine Similarity, Clearly Explained!!!
Long Short-Term Memory with PyTorch + Lightning
Просмотров 58 тыс.Год назад
Long Short-Term Memory with PyTorch Lightning
You are an absolute genius when it comes to explaining stuff. Every single time I come across a new concept and want to get a good solid basic understanding, I turn to your channel first. Thank you so very much for doing this fantastic work.
Thank you very much
Thank you sir for this amazing video, it helped me last year in my NLP exam and now i'm refreshing my information's about transformers hoping to land an interview soon!
this was the best explanation i've ever seen in my life, (i'm not even a english native speaker, i'm brazilian lol)
Bedankt
Yo! StatQuest please make a video about pipeline concept in machine learning.(like this one clear and true)
I'll keep that in mind.
We understood the calculation of weights and biases. But how would i know about the nodes...how do i understand the logic to connect all the inputs to activation function and to the output.?..and how many hidden layes we need? And no example for more than one hidden layer Could you please help me here
Designing neural networks is more of an art than a science - there are general guidelines, but generally speaking you find something that works on a related dataset and then train it with your own data. In other words, you rarely build your own neural network. However, if you are determined to build your own, the trade off is this - the more hidden layers and nodes within the hidden layers, the better your model will be able to fit any kind of data, no matter how complicated, but at the same time, you will increase the computation and training will be slow.
Can't we use correlation factor instead of Mutual information for continuous variable?
If you have continuous data, use R^squared.
Thank you so much for a very easy-and-nice-to-walk-through video! I really enjoyed the explanations and well-prepared slides! Also you made it very nicely paced. Thank you :))
Thank you!
BAM!
:)
awesome pca for dim reduction with vertical+horizontal+depth all in one 3-d rotates
:)
The amount of effort for some of these animations, especially in these videos on Attention and Transformers in insane. Thank you!
Glad you like them!
@statquest - Hi Josh - Does it make sense to perform PCA on categorical variables.
Not usually.
BAM!
:)
Hi all, with this seq2seq, can we apply for embedding a sentence, than using the output vector for semantic similarity ?
Maybe - I think it is more common to use an Encoder-Only Transformer like BERT. Encoder only transformers are just like Decoder-Only Transformers ( ruclips.net/video/bQ5BoolX9Ag/видео.html ), except they don't use Masked Attention.
I am on vacation in Hawaii but I am watching your neural network video. This video is so entertaining to watch :) Tai
BAM! Have a great vacation! :)
36:49
The explanation is so clean. I was clapping for him from my room. How can someone be so good at their job!
Thank you! :)
How these hidden states weights are determined, is it via backpropogation?
Yes. All neural networks are trained the same way, with backpropagation.
I feel like 15:15 deserved a triple bam, but maybe that's just me
I think you might be right on that one.
man was womp womping before womp womping was "cool"
:)
amazing thank you for the hard work!
My pleasure!
Thank you very much
You are welcome!
Where where I can apply the concept of principal component analysis 🙃
Whenever you have lots of things you are measuring.
goated intro
:)
Way way better than those taught in uni. Easily understood
Thanks!
BAM!!❤
:)
Too much biology in the beginning itself!
Sorry about that. I made this video for my colleagues at work - I used to work in a genetics laboratory - so I wanted them to understand the concepts in the context of the work that they did every day. I never expected anyone else to watch this video.
I did not understand at 8:30 minutes. Please explain. Why 0.9 * 100 * 0.15 ? Not 100*0.15 or log2(1/(0.9**100))?
100 = 100 flips. 0.9 = the probability of getting heads. 0.15 the amount of surprise for getting heads. Thus, the total surprise after 100 coin flips = 100 *0.9*0.15.
you did not show how the rest of the weights are updated? I need to understand how the derivative of the activation function affects the weight update
See this video for details and just replace the derivative of the softmax with 0 or 1 depending on the value for x: ruclips.net/video/GKZoOHXGcLo/видео.html
BAM BAM BAMMMM LOVED THE EXPLAINATION!!
Thank you!
I have a doubt..if the ROC plot is given, how to determine the threshold value used for classification for that confusion matrix?
I talk about that in this video: ruclips.net/video/qcvAqAH60Yw/видео.html
I saw so many videos before, IDK , what the fcuk they were teaching. This explaination is awesome.
Thanks!
8:24 The main part of the video! Everyone must watch!!!
BAM! :)
Amazing Video ✨
Thank you!!
statquest is the goat ngl
Thanks!
StatQuest never disappoints
BAM! :)
Thank you man. Appreciate such thorough but concise explanation.
Glad it was helpful!
Your videos are awesome! Makes things so much clearer! But I have a couple of questions: How do you handle the situation where a point has many identical points (ie. high-dim distance = 0)? How to calculate sigma_i? For example, if k = 10, but 7-8 of the neighbours are duplicates with Dij = 0, then sigma_i is undefined. Do I de-duplicate the data first and then add it back in at the end? And symmetrizing: Wij' = Wji' = Wij + Wji - Wij x Wji, yes? But aren't Wij and Wji only calculated for neighbours of i and j? What happens if Wij exists, but Wji does not? Do I add i as another neighbour of j's? (but then j would have more than k neighbours) I'm so confused.
To be honest, I would just try UMAP out and see what it does. It could treat duplicate points as a single point or do something else.
You are very interactive person😉
I try!
Are the conditions at each leaf random? Or is there a criteria for selecting them?
Gradient Boost usually uses regression trees. These are very similar to classification trees, but have a slightly different way to decide how to add branches and leaves. For more details, see: ruclips.net/video/g9c66TUylZ4/видео.html
So . In coding how can we calculate the derivative of loss function dont using math
It depends. For neural networks, tensors can calculate the derivatives for you. In other settings you have to do it by hand.
you are great .. learning the concept with music awesome..
Thanks a lot!
Heaven Of statistics!
Thanks!
Why don't we just equate d(sum of squared residuals)/d slope to 0, that point will give either the minima or maxima, if it is giving minima, then that will be the value of intercept, we need to take. Why can't we do this instead of checking using gradient descent?
A lot of people ask why we are using Gradient Descent to estimate the parameters in this video when we could just use least squares. We use least squares to produce a "gold standard estimate". This is the best possible estimate. We then attempt to derive the same estimate using Gradient Descent. This shows 1) how gradient descent works and 2) that the estimate is pretty good compared to the "gold standard".
@@statquestunderstood, thanks a lot!