Results for "data distribution"

Sampling from easier distribution with reweighting.

Updated belief after observing data.

Train/test environment mismatch.

Sum of independent variables converges to normal distribution.

Bayesian parameter estimation using the mode of the posterior distribution.

Models that learn to generate samples resembling training data.

Belief before observing data.

Autoencoder using probabilistic latent variables and KL regularization.

Differences between training and deployed patient populations.

Measures divergence between true and predicted probability distributions.

Measures how one probability distribution diverges from another.

Diffusion model trained to remove noise step by step.

Shift in model outputs.

How well a model performs on new data drawn from the same (or similar) distribution as training.

Learns the score (∇ log p(x)) for generative sampling.

Generator produces limited variety of outputs.

Describes likelihoods of random variable outcomes.

Eliminating variables by integrating over them.

Graphical model expressing factorization of a probability distribution.

Differences between training and inference conditions.

A mismatch between training and deployment data distributions that can degrade model performance.

Generative model that learns to reverse a gradual noise process.

Models that define an energy landscape rather than explicit probabilities.

Probabilistic model for sequential data with latent states.

Maintaining alignment under new conditions.

A measure of randomness or uncertainty in a probability distribution.

Controls amount of noise added at each diffusion step.

Stochastic generation strategies that trade determinism for diversity; key knobs include temperature and nucleus sampling.

Scales logits before sampling; higher increases randomness/diversity, lower increases determinism.

Converts logits to probabilities by exponentiation and normalization; common in classification and LMs.