Roadmap: embeddings
ML milestone 1 fills semantic_score and retrieval; bridge-style scores follow once clusters are available.
Recommended
Ranked recommendations
Papers come from a materialized ranking run (paper_scores per family). The API derives plain-language explanations from the same weights stored on the run. The undercited family only scores works in the frozen low-cite candidate pool (docs/candidate-pool-low-cite.md v0), scoped to your corpus snapshot.
ml2-5a-qual-r2-k6-20260405 | run rank-83976f1097 | snapshot source-snapshot-20260329-170012 | 15 papers
Web is filtering runs with NEXT_PUBLIC_RANKING_VERSION=ml2-5a-qual-r2-k6-20260405.
Bridge | live results
Family: bridge | order by materialized final_score descending
GiantMIDI-Piano: A Large-Scale MIDI Dataset for Classical Piano Music
Music and Audio ProcessingMusic Technology and Sound StudiesNeuroscience and Music PerceptionNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=1.0000, topic_growth=0.3751, bridge=0.4582, diversity_penalty=0.0000
- Speech and Audio ProcessingBlind Source Separation TechniquesVehicle Noise and Vibration Control
Near a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.0521, topic_growth=0.7179, bridge=0.8062, diversity_penalty=0.0000
The Sound Demixing Challenge 2023 - Music Demixing Track
Music and Audio ProcessingSpeech and Audio ProcessingMusic Technology and Sound StudiesNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.7292, topic_growth=0.3052, bridge=0.5481, diversity_penalty=0.0000
Perceptual and automated estimates of infringement in 40 music copyright cases
Music and Audio ProcessingMusic Technology and Sound StudiesDiverse Musicological StudiesNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.4167, topic_growth=0.4622, bridge=0.6783, diversity_penalty=0.0000
Diversity by Design in Music Recommender Systems
Music and Audio ProcessingMusic Technology and Sound StudiesNeuroscience and Music PerceptionNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.5556, topic_growth=0.3751, bridge=0.6179, diversity_penalty=0.0000
On Evaluation of Inter- and Intra-Rater Agreement in Music Recommendation
Music and Audio ProcessingNeuroscience and Music PerceptionSpeech and Audio ProcessingNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.6944, topic_growth=0.2999, bridge=0.7473, diversity_penalty=0.0000
- Music and Audio ProcessingDiverse Musicological StudiesInformation Retrieval and Search Behavior
Near a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.0000, topic_growth=0.6691, bridge=0.5610, diversity_penalty=0.0000
ChoraleBricks: A Modular Multitrack Dataset for Wind Music Research
Music and Audio ProcessingMusic Technology and Sound StudiesAnimal Vocal Communication and BehaviorNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.1042, topic_growth=0.5872, bridge=0.7262, diversity_penalty=0.0000
PiJAMA: Piano Jazz with Automatic MIDI Annotations
Music and Audio ProcessingMusic Technology and Sound StudiesDiverse Musicological StudiesNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.2865, topic_growth=0.4622, bridge=0.4512, diversity_penalty=0.0000
Supervised Contrastive Models for Music Information Retrieval in Classical Persian Music
Music and Audio ProcessingMusic Technology and Sound StudiesDiverse Music Education InsightsNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.0000, topic_growth=0.5872, bridge=0.7376, diversity_penalty=0.0000
Cross‑Modal Approaches to Beat Tracking: A Case Study on Chopin Mazurkas
Music and Audio ProcessingMusic Technology and Sound StudiesMusicology and Musical AnalysisNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.0000, topic_growth=0.5872, bridge=0.6125, diversity_penalty=0.0000
- Music and Audio ProcessingDiverse Musicological StudiesNatural Language Processing Techniques
Near a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.1042, topic_growth=0.5025, bridge=0.7860, diversity_penalty=0.0000
CCMusic: An Open and Diverse Database for Chinese Music Information Retrieval Research
Music and Audio ProcessingDiverse Musicological StudiesMusic Technology and Sound StudiesNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.1563, topic_growth=0.4622, bridge=0.6529, diversity_penalty=0.0000
Wagner Ring Dataset: A Complex Opera Scenario for Music Processing and Computational Musicology
Music and Audio ProcessingMusic Technology and Sound StudiesNeuroscience and Music PerceptionNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.2344, topic_growth=0.3751, bridge=0.6707, diversity_penalty=0.0000
JSD: A Dataset for Structure Analysis in Jazz Music
Music and Audio ProcessingMusic Technology and Sound StudiesNeuroscience and Music PerceptionNear a boundary between k-means cluster centroids in embedding space (squared L2 vs other centroids); bridge family weight is zero so final_score does not use this signal. Semantic score not used.
Signals: citation_velocity=0.2083, topic_growth=0.3751, bridge=0.6268, diversity_penalty=0.0000
Heuristic baseline
The rule-only undercited list (/api/v1/recommendations/undercited) uses the same pool definition but is not tied to a corpus snapshot. For snapshot-scoped A/B against the ranked undercited family, use Evaluation.