Emerging Papers
Rank papers by semantic fit, citation velocity, and local topic growth instead of raw popularity alone.
Research Radar
V1 is a ranking and explainability product for MIR + audio representation learning. Neural audio effects and music generation are held back as a controlled edge slice so the corpus stays coherent while bridge-paper logic stays meaningful.
Rank papers by semantic fit, citation velocity, and local topic growth instead of raw popularity alone.
Surface work connecting nearby but distinct audio-ML clusters without turning the product into a graph toy.
Expose score contributions, cluster context, and baseline comparisons so each recommendation feels engineered.
Ship benchmark, diversity, and temporal backtest pages as part of MVP rather than as an afterthought.