Final Report: Reliability and Robustness for Fast Bayesian Inference of Complex Data

reportActive / Technical Report | Accesssion Number: AD1205765 | Open PDF

Abstract:

The overall goal of this project is to develop inference methods that practitioners can trust. Importantly, these methods should be able to run quickly on modern problems of interest, and they should be able to work on complex data. That is, practitioners should be able to trust these methods as they actually work in practice. Especially in light of existing heuristics that can yield arbitrarily wrong results, an important part of reliability comes from supporting theory and evaluation.

Author(s):
Broderick, Tamara A.
Author Organization(s):
MASSACHUSETTS INST OF TECH CAMBRIDGE
Funding Organization(s):
ARMY RESEARCH LAB RESEARCH TRIANGLE PARK NC, RESEARCH TRIANGLE PARK , NC
Document Type:
Technical Report/Final Report
Publication Date:
2021 Aug 05
Pagination:
17

Security Markings

DOCUMENT & CONTEXTUAL SUMMARY

Distribution Code:
A - Approved For Public Release
Distribution Statement: Public Release

RECORD

Collection: TRECMS
Identifying Numbers
Report Number(s):
71204-CS-YIP.6
Grant Number(s):
W911NF-18-1-0063
Program Element Number(s):
PE611102
Subject Terms
Modernization Areas:
Autonomy
Communities of Interest:
Autonomy
Descriptor(s):
monte carlo method, bayesian networks, computational science, machine learning, information science, probabilistic models, bayesian inference, data analysis, data mining, data science, data sets, estimators, markov chains, statistical algorithms, accuracy, algorithms, gaussian processes
Keyword(s):
inference methods
Subject Categories:
Mathematical and Computer Sciences; Behavioral and Social Sciences
Creation Date:
2023 Jul 18
Update Date:
2023 Sep 07