Name
Title | ||
|---|---|---|
Streaming Bayesian inference: theoretical limits and mini-batch approximate message-passing. |
Abstract | ||
|---|---|---|
In statistical learning for real-world large-scale data problems, one must often resort to "streaming" algorithms which operate sequentially on small batches of data. In this work, we present an analysis of the information-theoretic limits of mini-batch inference in the context of generalized linear models and low-rank matrix factorization. In a controlled Bayes-optimal setting, we characterize the optimal performance and phase transitions as a function of mini-batch size. We base part of our results on a detailed analysis of a mini-batch version of the approximate message-passing algorithm (Mini-AMP), which we introduce. Additionally, we show that this theoretical optimality carries over into real-data problems by illustrating that Mini-AMP is competitive with standard streaming algorithms for clustering. |
Year | DOI | Venue |
|---|---|---|
2017 | 10.1109/allerton.2017.8262853 | 2017 55TH ANNUAL ALLERTON CONFERENCE ON COMMUNICATION, CONTROL, AND COMPUTING (ALLERTON) |
DocType | Volume | ISSN |
Conference | abs/1706.00705 | 2474-0195 |
Citations | PageRank | References |
1 | 0.35 | 4 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Andre Manoel | 1 | 64 | 5.11 |
| Florent Krzakala | 2 | 977 | 67.30 |
| Eric W. Tramel | 3 | 365 | 17.03 |
| Lenka Zdeborová | 4 | 1190 | 78.62 |