ABSTRACT
In many applications, developed deep-learning models need to be iteratively debugged and refined to improve the model efficiency over time. Debugging some models, such as temporal multilabel classification (TMLC) where each data point can simultaneously belong to multiple classes, can be especially more challenging due to the complexity of the analysis and instances that need to be reviewed. In this article, focusing on video activity recognition as an application of TMLC, we propose DETOXER, an interactive visual debugging system to support finding different error types and scopes through providing multiscope explanations.
ABSTRACT
The paper investigates parameterized approximate message-passing schemes that are based on bounded inference and are inspired by Pearl's belief propagation algorithm (BP). We start with the bounded inference mini-clustering algorithm and then move to the iterative scheme called Iterative Join-Graph Propagation (IJGP), that combines both iteration and bounded inference. Algorithm IJGP belongs to the class of Generalized Belief Propagation algorithms, a framework that allowed connections with approximate algorithms from statistical physics and is shown empirically to surpass the performance of mini-clustering and belief propagation, as well as a number of other state-of-the-art algorithms on several classes of networks. We also provide insight into the accuracy of iterative BP and IJGP by relating these algorithms to well known classes of constraint propagation schemes.
