Learning Neural Sequence-to-Sequence Models from Weak Feedback with Bipolar Ramp Loss
This addresses the problem of training neural models in scenarios where gold supervision is unavailable, offering an incremental improvement for weakly supervised tasks like machine translation and semantic parsing.
The paper tackled training neural sequence-to-sequence models without gold labels by using weak feedback, introducing bipolar ramp loss objectives that actively discourage negative outputs. Results showed these objectives outperformed non-bipolar ramp loss and minimum risk training on weakly supervised machine translation and semantic parsing tasks, with a novel token-level ramp loss achieving the best performance.
In many machine learning scenarios, supervision by gold labels is not available and consequently neural models cannot be trained directly by maximum likelihood estimation (MLE). In a weak supervision scenario, metric-augmented objectives can be employed to assign feedback to model outputs, which can be used to extract a supervision signal for training. We present several objectives for two separate weakly supervised tasks, machine translation and semantic parsing. We show that objectives should actively discourage negative outputs in addition to promoting a surrogate gold structure. This notion of bipolarity is naturally present in ramp loss objectives, which we adapt to neural models. We show that bipolar ramp loss objectives outperform other non-bipolar ramp loss objectives and minimum risk training (MRT) on both weakly supervised tasks, as well as on a supervised machine translation task. Additionally, we introduce a novel token-level ramp loss objective, which is able to outperform even the best sequence-level ramp loss on both weakly supervised tasks.