Negative Training for Neural Dialogue Response Generation

By Tianxing He, James R. Glass. 2019

74 citations recorded by the source corpus.

Although deep learning models have brought tremendous advancements to the field of open-domain dialogue response generation, recent research results have revealed that the trained models have undesirable generation behaviors, such as malicious responses and generic (boring) responses. In this work, we propose a framework named “Negative Training” to minimize such behaviors. Given a trained model, the framework will first find generated samples that exhibit the undesirable behavior, and then use them to feed negative training signals for fine-tuning the model. Our experiments show that negative training can significantly reduce the hit rate of malicious responses, or discourage frequent responses and improve response diversity.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Natural Language Processing

Citations in the corpus

Open-Domain Conversational Agents: Current Progress, Open Problems, and Future Directions (2020, background): “Other works have attempted to control for the toxicity of models by removing offensive content from the training data (Zhang et al., 2019; Adiwardana et al., 2020) or training objectives (He and Glass, 2019).”
Red Teaming Language Models with Language Models (2022, background): “For example, unlikelihood training (Welleck et al., 2020; He and Glass, 2020) could minimize the probability of the original, bad output given the test case.”
Don’t Say That! Making Inconsistent Dialogue Unlikely with Unlikelihood Training (2019, background): “Closer to our work is that of He and Glass (2019) which developed the approach of negative training to prevent generic and malicious responses in dialogue models.”
Mix-review: Alleviate Forgetting in the Pretrain-Finetune Framework for Neural Language Generation Models (2019, unknown): “As shown by He & Glass (2019a), it is relatively difficult to trigger the dialogue models trained from scratch to output malicious responses (note that the conversations from the Dailydialogue data tend to be very polite).”
Automatically Exposing Problems with Neural Dialog Models (2021, background): “The most similar work to ours is probably directing pre-trained models into generating a list of pre-defined tokens or sentences (He and Glass, 2018; Liu et al., 2019a, 2020; He and Glass, 2020).”
Exposure Bias versus Self-Recovery: Are Distortions Really Incremental for Autoregressive Text Generation? (2019, background): “Therefore, it is completely possible that a non-MLE training objec-tive can lead to better generation performance (Lu et al., 2018; Huszár, 2015; Welleck et al., 2020; He and Glass, 2020; Gu et al., 2017).”
Identifying Untrustworthy Samples: Data Filtering for Open-domain Dialogues with Bayesian Optimization (2021, methodology): “To accelerate the process, we design a training framework, named DiffMLE-NEG, which combines maximum likelihood estimation (MLE) and negative training method (NEG) [15].”
CLIFF: Contrastive Learning for Improving Faithfulness and Factuality in Abstractive Summarization (2021, background): “For text generation tasks, unlikelihood training is proposed to penalize the generation of negative tokens (e.g., repeated words) and sentences (e.g., contra-dictory responses in a dialogue system) (Welleck et al., 2020; Li et al., 2020; He and Glass, 2020).”
Analyzing the Forgetting Problem in Pretrain-Finetuning of Open-domain Dialogue Response Models (2021, background): “Recent research (Holtzman et al., 2019; Radford et al., 2019; Fan et al., 2018) has shown that a strategy called top-k sampling, in which the next word is sampled from the top k most probable choices, is a better choice than the traditional beam-search decoding, due to better diversity.”