Qiang Fang: A Dual-Path Speaker-Independent Acoustic-Toarticulatory Inversion Model Based on Content and Speaker Information Disentanglement. APSIPA 2025: 712-717