To fill this gap, we propose two attention-mechanism-based encoder–decoder models that incorporate multisource ... with a success rate of nearly 50%. Our model achieved competitive results, mainly ...