Differential Performance Debugging With Discriminant Regression Trees

Authors: Saeid Tizpaz-Niari, Pavol Cerny, Bor-Yuh Evan Chang, Ashutosh Trivedi

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our approach on benchmarks consisting of Java programs where we are interested in debugging performance. We show that our algorithm outperforms other well-known regression tree learning algorithms in terms of running time and accuracy of classification.
Researcher Affiliation Academia Saeid Tizpaz-Niari, Pavol ˇCern y, Bor-Yuh Evan Chang, Ashutosh Trivedi University of Colorado Boulder
Pseudocode Yes Algorithm 1: LEARNDISCRIMINANTREGRESSIONTREE(T , B, Bε) and Algorithm 2: KLINEARCLUSTERINGALGORITHM
Open Source Code No We implement our approach in the tool DPDEBUGGER and evaluate it on benchmarks consisting of a suite of Java programs.
Open Datasets No The paper describes datasets that were collected or constructed for the specific experiments (e.g., 'inputs...are PNG and JPEG images of various sizes', 'benchmarks were constructed', 'input data set for our experiments consists of asking for different plots'), but does not provide concrete access information (links, DOIs, or formal citations to publicly available datasets with author/year attribution) for these datasets.
Dataset Splits Yes The metric we are interested in is accuracy of classification based on 10-fold cross-validation.
Hardware Specification No No specific hardware details (e.g., exact GPU/CPU models, memory amounts, or detailed computer specifications) used for running experiments were mentioned in the paper.
Software Dependencies No The paper mentions using Java programs and tools like M5Prime, GUIDE, and CART, but does not specify version numbers for any of the software dependencies.
Experiment Setup Yes Our tool applied spectral clustering with the number of clusters set to 3. (Charts4j) We applied K-linear clustering algorithm where we set the number of clusters to 5. (Snapbuddy) Our tool applied K-linear clustering algorithm with K = 2. (JFree Chart)