[13] Adji Bousso Dieng - Deep Probabilistic Graphical Modeling
The Thesis Review
The Importance of Deep Learning for Graph Modeling
I was getting into probabilistic graphical modeling, remember the two courses I mentioned. So my very first paper was on an approach to doing approximate posterior inference with latent variable models. It's not just taking a latent per latent, one latent variable for observation and then parameterizing it with a neural network and doing inference. It's more about, okay, give me data, I have inductive biases, let me set up a generative process which gives me interpretivity from the get go. And let me see now how I can leverage those things within this formalism. There are overlaps, but it's a bit different.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
