Outlier-Robust Convex Segmentation

Authors: Itamar Katz, Koby Crammer

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Robustness to outliers is evaluated on two real-world tasks related to speech segmentation. Our algorithms outperform baseline segmentation algorithms. We evaluate the performance of the proposed algorithms on two speech segmentation tasks, for both clean sources and sources contaminated with added non-stationary noise. Our algorithms outperform other algorithms in both the clean and outlier-contaminated setting.
Researcher Affiliation Academia Itamar Katz and Koby Crammer Department of Electrical Engineering The Technion Israel Institute of Technology Haifa, 32000 Israel
Pseudocode Yes Algorithm 1 Top-down outlier-robust hierarchical segmentation
Open Source Code No The paper does not provide any explicit statements about making the source code available or include links to a code repository.
Open Datasets Yes In this experiment we used the TIMIT corpus (Garofolo and others 1988).
Dataset Splits No The paper mentions using datasets like TIMIT and a hand-annotated audio recording, but it does not specify explicit training, validation, and test splits (e.g., percentages or sample counts) for reproduction.
Hardware Specification No The paper does not provide specific details regarding the hardware (e.g., GPU/CPU models, memory, or cloud platforms) used to run the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., programming languages, libraries, or solvers with their versions) that would be needed for replication.
Experiment Setup Yes Audio was divided into frames of 16ms duration and 1 ms hop-length, each represented with 13 MFCC coefficients. We used 13 MFCC coefficient with 25ms window length and 10ms hop length. Then a Gaussian Mixture Model (GMM) Tj with 10 components and a diagonal covariance matrix is fitted to the jth block Sj. These parameters of the GMM were selected using the Bayesian Information Criterion (BIC).