Zeitpunkt Nutzer Delta Tröts TNR Titel Version maxTL So 04.08.2024 00:01:07 0 -25.561 0 0,0 0 Sa 03.08.2024 00:00:18 25.561 +8 788.532 30,8 det.social 4.2.10 500 Fr 02.08.2024 00:01:13 25.553 +5 788.077 30,8 det.social 4.2.10 500 Do 01.08.2024 00:01:11 25.548 +5 787.576 30,8 det.social 4.2.10 500 Mi 31.07.2024 00:00:50 25.543 +6 787.041 30,8 det.social 4.2.10 500 Di 30.07.2024 00:00:03 25.537 +6 786.549 30,8 det.social 4.2.10 500 Mo 29.07.2024 00:00:56 25.531 +5 785.970 30,8 det.social 4.2.10 500 So 28.07.2024 00:00:36 25.526 +5 785.486 30,8 det.social 4.2.10 500 Sa 27.07.2024 00:00:00 25.521 +10 784.946 30,8 det.social 4.2.10 500 Fr 26.07.2024 00:01:08 25.511 0 784.393 30,7 det.social 4.2.10 500
Harald Klinke (@HxxxKxxx) · 10/2022 · Tröts: 1.636 · Folger: 1.486
So 04.08.2024 21:02
I was waiting for this point:
New research paper explores the risk of generative models creating "Model Autophagy Disorder" (MAD) when training on their own synthetic data. Without fresh real data, future models may see a decline in quality and diversity, leading to amplified artifacts. Read the full study for insights on how to mitigate these effects. #AI #MachineLearning #GenerativeModels #DataScience #Research #Tech
Link to the paper: https://arxiv.org/pdf/2307.01850
Research paper titled "Self-Consuming Generative Models Go MAD" by Sina Alemohemmad, Josue Casco-Rodriguez, Lorenzo Luzi, Ahmed Imtiaz Humayun, Hossein Babaei, Daniel LeJe
[Öffentlich] Antw.: 0 Wtrl.: 0 Fav.: 0 · via IceCubesApp