One-sided Matrix Completion from Two Observations Per Row

Authors: Steven Cao, Percy Liang, Gregory Valiant

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our algorithm on one-sided recovery of synthetic data and low-coverage genome sequencing. In these settings, our algorithm substantially outperforms standard matrix completion and a variety of direct factorization methods.
Researcher Affiliation Academia Steven Cao 1 Percy Liang 1 Gregory Valiant 1 1Stanford University. Correspondence to: Steven Cao <shcao@stanford.edu>.
Pseudocode No No explicit pseudocode or algorithm block was found.
Open Source Code No No statement regarding the release of source code or a link to a repository was found.
Open Datasets Yes Finally, we evaluate our method on synthetic data and the 1000genomes dataset (Fairley et al., 2019).
Dataset Splits No The paper describes varying the number of rows (m) and observations per row (k) for experiments, and a fixed number of columns (d), but does not specify explicit train/validation/test dataset splits with percentages or counts.
Hardware Specification Yes Experiments were run on TITAN RTX and RTX 3090 GPUs with 24 gigabytes of memory
Software Dependencies No The paper mentions 'Optimization was done via Adam' and 'implement via non-convex optimization' but does not list specific software packages or libraries with version numbers (e.g., PyTorch 1.9, TensorFlow 2.x).
Experiment Setup Yes Optimization was done via Adam (Kingma & Ba, 2015) with lr = 1e 10, β = (0.9, 0.999), and 10,000 steps.