Approximate Posterior Inference For Latent Variable Discovery

via “scalable-posterior-inference-via-variational-approximation”

* 🏆 2006: [Reducing the Dimensionality of Data with Neural Networks (Autoencoder)](https://www.science.org/doi/abs/10.1126/science.1127647)

Unique: Introduces mean-field variational inference to topic modeling (Blei et al. 2003), replacing expensive Gibbs sampling with coordinate ascent optimization over variational parameters — enabling orders-of-magnitude speedup while maintaining interpretability through explicit posterior approximation

vs others: Dramatically faster than Gibbs sampling on large corpora (hours vs days) while providing explicit uncertainty estimates unlike deterministic LSA; trades some accuracy for scalability but remains more principled than heuristic approximations

* 🏆 2014: [Generative Adversarial Networks (GAN)](https://papers.nips.cc/paper/2014/hash/5ca3e9b122f61f8f06494c97b1afccf3-Abstract.html)

Unique: Learns an amortized inference network that maps observations directly to posterior parameters, avoiding the need to optimize separate variational parameters for each data point. This amortization enables fast inference at test time and allows the inference network to generalize to unseen data. Prior variational inference methods required optimizing per-datapoint parameters, making inference slow and preventing generalization.

vs others: Provides orders of magnitude faster inference than sampling-based methods (Gibbs sampling, Hamiltonian Monte Carlo) because the encoder is a single forward pass; enables generalization to new data unlike per-datapoint variational parameters; provides deterministic posterior estimates (via mean) unlike sampling methods which require multiple samples for low-variance estimates.