DeepAI AI Chat
Log In Sign Up

How to "DODGE" Complex Software Analytics?

by   Amritanshu Agrawal, et al.
NC State University

AI software is still software. Software engineers need better tools to make better use of AI software. For example, for software defect prediction and software text mining, the default tunings for software analytics tools can be improved with "hyperparameter optimization" tools that decide (e.g.,) how many trees are needed in a random forest. Hyperparameter optimization is unnecessarily slow when optimizers waste time exploring redundant options (i.e., pairs of tunings with indistinguishably different results). By ignoring redundant tunings, the Dodge(E) hyperparameter optimization tool can run orders of magnitude faster, yet still find better tunings than prior state-of-the-art algorithms (for software defect prediction and software text mining).


page 7

page 8

page 10

page 12


While Tuning is Good, No Tuner is Best

Hyperparameter tuning is the black art of automatically finding a good c...

Is One Hyperparameter Optimizer Enough?

Hyperparameter tuning is the black art of automatically finding a good c...

Mining Documentation to Extract Hyperparameter Schemas

AI automation tools need machine-readable hyperparameter schemas to defi...

Better Software Analytics via "DUO": Data Mining Algorithms Using/Used-by Optimizers

This paper claims that a new field of empirical software engineering res...

Hyperparameter Optimization for AST Differencing

Computing the differences between two versions of the same program is an...

HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperparameter Optimization

We present a new software, HYPPO, that enables the automatic tuning of h...

Applications of Psychological Science for Actionable Analytics

Actionable analytics are those that humans can understand, and operation...