Abstract page for arXiv paper 2610.02612: Learning When to Commit from Partial Speech for End-to-End Simultaneous Speech Translation