Papers › An efficient manifold density estimator for all recommendation systems

An efficient manifold density estimator for all recommendation systems

2 Jun 2020arXiv:2006.01894archive 2025-07-28

Jacek Dąbrowski, Barbara Rychalska, Michał Daniluk, Dominika Basaj, Konrad Gołuchowski, Piotr Babel, Andrzej Michałowski, Adam Jakubowski

Many unsupervised representation learning methods belong to the class of similarity learning models. While various modality-specific approaches exist for different types of data, a core property of many methods is that representations of similar inputs are close under some similarity function. We propose EMDE (Efficient Manifold Density Estimator) - a framework utilizing arbitrary vector representations with the property of local similarity to succinctly represent smooth probability densities on Riemannian manifolds. Our approximate representation has the desirable properties of being fixed-size and having simple additive compositionality, thus being especially amenable to treatment with neural networks - both as input and output format, producing efficient conditional estimators. We generalize and reformulate the problem of multi-modal recommendations as conditional, weighted density estimation on manifolds. Our approach allows for trivial inclusion of multiple interaction types, modalities of data as well as interaction strengths for any recommendation setting. Applying EMDE to both top-k and session-based recommendation settings, we establish new state-of-the-art results on multiple open datasets in both uni-modal and multi-modal settings.

PaperPDFConference PDFCode

Code

Synerise/booking-challenge mentioned on GitHubpytorch report
Synerise/kdd-cup-2021 mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AllDensity EstimationRecommendation SystemsRepresentation LearningSession-Based Recommendations

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Session-Based Recommendations Diginetica EMDE MM Hit@20 38.49 #10 of 13 Archive leaderboard report
Session-Based Recommendations Diginetica EMDE MM MRR@20 17.31 #10 of 13 Archive leaderboard report
Session-Based Recommendations Diginetica EMDE Hit@20 37.52 #11 of 13 Archive leaderboard report
Session-Based Recommendations Diginetica EMDE MRR@20 17.24 #11 of 13 Archive leaderboard report
Session-Based Recommendations Retailrocket EMDE MM Hit@20 0.5073 #1 of 2 Archive leaderboard report
Session-Based Recommendations Retailrocket EMDE MM MRR@20 0.3664 #1 of 2 Archive leaderboard report
Session-Based Recommendations Retailrocket EMDE Hit@20 0.4704 #2 of 2 Archive leaderboard report
Session-Based Recommendations Retailrocket EMDE MRR@20 0.3524 #2 of 2 Archive leaderboard report
Session-Based Recommendations yoochoose1 EMDE MM MRR@20 31.16 #1 of 4 Archive leaderboard report
Session-Based Recommendations yoochoose1 EMDE MM Precision@20 74.3 #1 of 4 Archive leaderboard report
Session-Based Recommendations yoochoose1 EMDE MRR@20 31.04 #2 of 4 Archive leaderboard report
Session-Based Recommendations yoochoose1 EMDE Precision@20 73.0 #2 of 4 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections