Learning Nonsymmetric Determinantal Point Processes

Authors: Mike Gartrell, Victor-Emmanuel Brunel, Elvis Dohmatob, Syrine Krichene

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our model on synthetic and real-world datasets, demonstrating improved predictive performance compared to symmetric DPPs, which have previously shown strong performance on modeling tasks associated with these datasets. We run extensive experiments on several synthetic and real-world datasets. We compare the performance of all methods using a standard recommender system metric: mean percentile rank (MPR). Table 1 shows the results of our performance evaluation on the Amazon and UK datasets.
Researcher Affiliation Collaboration Mike Gartrell Criteo AI Lab m.gartrell@criteo.com Victor-Emmanuel Brunel ENSAE Paris Tech victor.emmanuel.brunel@ensae.fr Elvis Dohmatob Criteo AI Lab e.dohmatob@criteo.com Syrine Krichene Criteo AI Lab syrinekrichene@google.com
Pseudocode No The paper describes algorithmic steps but does not include any formally labeled 'Pseudocode' or 'Algorithm' block.
Open Source Code Yes Our code is available at https://github.com/cgartrel/nonsymmetric-DPP-learning
Open Datasets Yes Amazon Baby Registries: This public dataset... UK Retail: This is a public dataset [8] that contains 25,898 baskets...
Dataset Splits Yes For all experiments, a random selection of 80% of the baskets are used for training, and the remaining 20% are used for testing. We use a small held-out validation set for tracking convergence and tuning hyperparameters.
Hardware Specification No The paper mentions implementing models and optimization algorithms but does not specify any particular hardware (e.g., GPU/CPU models, memory) used for the experiments.
Software Dependencies Yes We implement our models using Py Torch 2, and use the Adam [16] optimization algorithm to train our models.
Experiment Setup Yes Numerically, to prevent such singularities, in our implementation we add a small I correction to each LYi when optimizing Eq. 12 (we set = 10 5 in our experiments). We use a small held-out validation set for tracking convergence and tuning hyperparameters. Convergence is reached during training when the relative change in validation log-likelihood is below a pre-determined threshold, which is set identically for all models. We use D = 30, = 0 for both Amazon datasets; D0 = 100 for the Amazon 3-category dataset; D0 = 30 for the Amazon apparel dataset; D = 100, D0 = 20, = 1 for the UK dataset; and β = γ = 0 for all datasets.