Helios: A POWER Enhanced Music Recommendation Engine
summary
Helios is a patent‑pending, POWER‑accelerated music search engine that takes audio as the query and returns ranked matches from commercial catalogs in seconds. It’s deployed as a systemd server providing a REST API, performs DSP‑heavy sub-symbolic analysis on time and spectral domains, scales to catalogs of thousands-to-millions of tracks, and runs particularly fast on Power9 hardware (IBM collaboration). Built to live on a label’s own servers so catalogs never leave premises, the prototype works and shows substantial runtime gains over generic x86/ARM deployments.
AI-written summary · about
related talks
23mRevolutionizing API Search with Holon Concept Explorations
50mBetter Spotify Playlists through Music Theory & Discrete Optimization
35mDiversity and Data: Cases in the Music Industry
20mThe Pod: Agents Are Developers Too: The Helius Agentic Stack
15mPowering PlayStation personalization to millions
11mCrafting a Personalised Music App
25mAI-powered digital transformation
46mHelidon and AI - Building Intelligent Applications