Heavy-Tailed Noise
Heavy-tailed noise refers to noise whose probability distribution tails decay more slowly than an exponential distribution, with potentially infinite variance or only finite low-order moments. It is common in real-world data such as finance and sensor data, and poses challenges to the analysis of traditional statistical learning theory.
Timeline (last 90 days)
The paper clearly points out that achieving robustness under heavy-tailed noise requires a slightly stronger explicit sample size requirement than in the light-tailed case.
Although the paper's sample size requirement is 'mildly stronger', in extreme heavy-tailed cases (such as when only low-order moments exist), this requirement may still exceed the scale of many real datasets.
All Facts (2)
The paper clearly points out that achieving robustness under heavy-tailed noise requires a slightly stronger explicit sample size requirement than in the light-tailed case.
50%UnverifiedAlthough the paper's sample size requirement is 'mildly stronger', in extreme heavy-tailed cases (such as when only low-order moments exist), this requirement may still exceed the scale of many real datasets.
50%