Timezone: »
Oral
Yes, but Did It Work?: Evaluating Variational Inference
Yuling Yao · Aki Vehtari · Daniel Simpson · Andrew Gelman
While it's always possible to compute a variational approximation to a posterior distribution, it can be difficult to discover problems with this approximation". We propose two diagnostic algorithms to alleviate this problem. The Pareto-smoothed importance sampling (PSIS) diagnostic gives a goodness of fit measurement for joint distributions, while simultaneously improving the error in the estimate. The variational simulation-based calibration (VSBC) assesses the average performance of point estimates.
Author Information
Yuling Yao (Columbia University)
Aki Vehtari (Aalto University)
Daniel Simpson (University of Toronto)
Andrew Gelman (Columbia University)
Related Events (a corresponding poster, oral, or spotlight)
-
2018 Poster: Yes, but Did It Work?: Evaluating Variational Inference »
Wed. Jul 11th 04:15 -- 07:00 PM Room Hall B
More from the Same Authors
-
2019 Poster: Bayesian leave-one-out cross-validation for large data »
Måns Magnusson · Michael Andersen · Johan Jonasson · Aki Vehtari -
2019 Poster: Active Learning for Decision-Making from Imbalanced Observational Data »
Iiris Sundin · Peter Schulam · Eero Siivola · Aki Vehtari · Suchi Saria · Samuel Kaski -
2019 Oral: Active Learning for Decision-Making from Imbalanced Observational Data »
Iiris Sundin · Peter Schulam · Eero Siivola · Aki Vehtari · Suchi Saria · Samuel Kaski -
2019 Oral: Bayesian leave-one-out cross-validation for large data »
Måns Magnusson · Michael Andersen · Johan Jonasson · Aki Vehtari