Absolutely amazing video course. Especially after looking at other sources I notice how valuable this is. Every video achieves to combine the intuition and math in a concise was. I recommend the videos to anyone who wants to learn about ML.
23:20 perplexity - adjust sigma for each i so that we reach perplexity=30. may be small in dense group, but big in sparse group. 43:40 crowding problem
t-SNE is 1) non-linear 2) non-parametric (aka stochastic, non-deterministic) 3:28-4:20 8:46 MNIST 9:22 PCA's visual 17:1417:57 18:45 t-SNE's visual 31:29❗2 separate blue clusters cannot get together 32:41 the fix: increase "Early Exaggeration" temporarily to increase the attraction force and then decrease back
Great lesson. How can you use t-SNE not just for visualization but also for classification? Does t-SNE take into account that some variables are more related with the formation of the cluster and other just add noise? I mean, in some moedls you can calculate the p-value and the SHAP for each variable. Can you get this kind of information here?
Absolutely amazing video course. Especially after looking at other sources I notice how valuable this is. Every video achieves to combine the intuition and math in a concise was.
I recommend the videos to anyone who wants to learn about ML.
The best video on this topic I have found so far by a large margin. Excellent work!
Worth every second. You are a blessing to humanity.
what an amazing explanations.......................well done............BRAVO!
So well explained! The best video resource I have seen on t-SNE so far!
It is an amazing course, worth the time to watch and learn from it.
Amazingly explained, It's such a great resource.
23:20 perplexity - adjust sigma for each i so that we reach perplexity=30. may be small in dense group, but big in sparse group.
43:40 crowding problem
Excellent talk with spot on visuals and explanations. Thanks!
Amazing Lecture, very well explained! Thank you for sharing!
Wonderful job. Really enjoy watching this.
Amazing course with great vizualisations ! thank you very much
Great explanation with both details and good examples
Incredibly explained. Congratulations!
Cool explanation and visualizations!
Top quality lecture, thanks for sharing
t-SNE is 1) non-linear 2) non-parametric (aka stochastic, non-deterministic) 3:28-4:20
8:46 MNIST
9:22 PCA's visual
17:14 17:57
18:45 t-SNE's visual
31:29❗2 separate blue clusters cannot get together
32:41 the fix: increase "Early Exaggeration" temporarily to increase the attraction force and then decrease back
Excellent presentation
Fabulous video! This was really helpful, thank you!
Thank you for your awesome explanation and illustrations nive thank you very much
Thank you for this course!
Beautiful explanations!
amazing lecture. Please post more videos.
Awesome explanations. Thank you very much.
Amazing! Super interesting and understandable!
Great lesson.
How can you use t-SNE not just for visualization but also for classification?
Does t-SNE take into account that some variables are more related with the formation of the cluster and other just add noise?
I mean, in some moedls you can calculate the p-value and the SHAP for each variable. Can you get this kind of information here?
how can one get good results with PCA init as don't we lose valuable non-linear information?
Excellent lecture, thanks
Thanks a lot for greay content
Bravo! Thank you very much.
Amazing!
amazing, thx.
amazing content
Greetings from Spain
Very good lesson
Respect!
This video is the bees knees
The dogs bollocks
Where can we find lecture notes?
Use the subtitles/closed captions?
It's like your cup when u add the coffee powder into water