An important result on PCA recovery thresholds in low-rank matrix inference using the replica method, from our own @urteado and @PierpaoloVivo in collaboration with @gabrielescr . Check it out on the arXiv [arxiv.org/abs/2511.11927] 👇
Pierpaolo Vivo@PierpaoloVivo
🚨 NEW PHASE TRANSITION ALERT 🚨 We just posted a paper on arXiv that extends the famous BBP transition to the case of sparse noise! The "ASV transition" shows when PCA can recover signals from sparse corruption - beyond classical random matrix theory. [arxiv.org/abs/2511.11927]
English
