The article explores ways to estimate volatility in financial models using high-frequency data. Researchers look at different methods to calculate volatility based on the sum of absolute powers of log-returns. They find that adding certain processes to the model doesn't affect the accuracy of the volatility estimate, allowing for more flexibility in modeling. This can help in dealing with uncertainties or errors in the model. The study discusses how different choices of parameters can be used depending on the assumptions made in the model and the type of data available.