Semi-Supervised Learning on Data Streams via Temporal Label Propagation

Authors: Tal Wagner, Sudipto Guha, Shiva Kasiviswanathan, Nina Mishra

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on real datasets validate that the algorithm can quickly and accurately classify points on a stream with a small quantity of labeled examples. We evaluate our solution on several real datasets.
Researcher Affiliation Collaboration 1CSAIL, MIT. Work done during an internship at Amazon. 2Amazon. Correspondence to: Tal Wagner <talw@mit.edu>.
Pseudocode Yes Algorithm TLP : Temporal Label Propagation
Open Source Code No The paper does not provide a link or explicit statement about the availability of the source code for the described methodology.
Open Datasets Yes (a) Incart-ECG (Goldberger et al., 2000): Dataset of ECG timeseries from Physio Net bank (b) Daphnet-Gait (Bachlin et al., 2010): Annotated readings of 9 accelerometer sensors (c) Caltech10-101 (Fei-Fei et al., 2006): Caltech-101 dataset consists of images annotated by 101 object classes (d) Cam Vid-Car (Brostow et al., 2009) (Cambridge-driving Labeled Video Database)
Dataset Splits No The paper does not provide specific training/test/validation dataset splits (e.g., percentages, sample counts, or explicit references to standard splits) needed to reproduce the data partitioning.
Hardware Specification Yes All experiments were performed on a 3.1 GHz Intel Core i7 machine with 16GB RAM.
Software Dependencies No The paper mentions using "standard RBF similarity" but does not specify any software libraries, frameworks, or solvers with version numbers that would be necessary for reproduction.
Experiment Setup Yes We set σ = 0.1 for Incart-ECG, Daphnet-Gait, and Cam Vid and σ = 10 for Caltech10-101.