Google Research says diffusion creativity is a byproduct of smooth score learning
dl_weekly · x · 2026-07-21
Google Research argues that diffusion-model creativity can be understood as a mathematical side effect of training.
- Regularization smooths the learned score function.
- That smoothing encourages interpolation between training points instead of rote memorization.
- The result is a framework for thinking about why diffusion models can generate novel-looking outputs even when they are trained on finite data.
More from Research
- Project APE finds verifier reliability drops when papers contain multiple errors — soumitrashukla9 · 2026-07-22
- Project APE says verifier costs fell about 90x in a year as Chinese open models lead — soumitrashukla9 · 2026-07-22
- OpenAI-linked paper says capability RL can make models more reward-seeking — MariusHobbhahn · 2026-07-22
- Project APE builds its verifier benchmark from 100 AI-written papers with injected errors — soumitrashukla9 · 2026-07-22
- Paper proposes a CRED taxonomy and benchmark to measure research-error detectors — soumitrashukla9 · 2026-07-22
- OpenAI says long-horizon models need safety and alignment checks across full action sequences — rhiever · 2026-07-22