Minsuk Chang
@minsuk.bsky.social
Research Scientist @ Google Deepmind. Opinions are my own.
minsukchang.com
minsukchang.com
clarification: I won't be able to reply to all your emails (thanks for sending those!). But I will try my best in making sure your application is seen by our recruiters.
re: non-us based candidates - tbh I don't know, but I assume yes.
re: non-us based candidates - tbh I don't know, but I assume yes.
November 7, 2025 at 11:36 PM
clarification: I won't be able to reply to all your emails (thanks for sending those!). But I will try my best in making sure your application is seen by our recruiters.
re: non-us based candidates - tbh I don't know, but I assume yes.
re: non-us based candidates - tbh I don't know, but I assume yes.
Reposted by Minsuk Chang
Yes!! Originally, generative ML modeled a joint distribution, e.g., P(X,Y), whereas discriminative ML modeled a conditional distribution, e.g., P(Y|X). A famous paper by Ng and Jordan (papers.nips.cc/paper_files/...) compared naive Bayes (generative) to logistic regression (discriminative). 1/
On Discriminative vs. Generative Classifiers: A comparison of logistic regression and naive Bayes
papers.nips.cc
June 3, 2025 at 4:42 AM
Yes!! Originally, generative ML modeled a joint distribution, e.g., P(X,Y), whereas discriminative ML modeled a conditional distribution, e.g., P(Y|X). A famous paper by Ng and Jordan (papers.nips.cc/paper_files/...) compared naive Bayes (generative) to logistic regression (discriminative). 1/