Contrastive Latent Variable Models for Neural Text Generation
Deep latent variable models such as variational autoencoders and energy-based models are widely used for neural text generation. Most of them focus on matching the prior distribution with the posterior distribution of the latent variable for text reconstruction. In addition to instance-level reconstruction, this paper aims to integrate contrastive learning in the latent space, forcing the latent variables to learn high-level semantics by exploring inter-instance relationships. Experiments on various text generation benchmarks show the effectiveness of our proposed method. We also empirically show that our method can mitigate the posterior collapse issue for latent variable based text generation models.
Code (1)
Tasks
Contrastive LearningText GenerationMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Guiding Energy-based Models via Contrastive Latent Variables
An energy-based model (EBM) is a popular generative framework that offers both explicit density and architectural flexibility, but training them is difficult since it is often unstable and time-consuming. In recent years…
Representation LearningEnhancing Personalized Dialogue Generation with Contrastive Latent Variables: Combining Sparse and Dense Persona
The personalized dialogue explores the consistent relationship between dialogue generation and personality. Existing personalized dialogue agents model persona profiles from three resources: sparse or dense persona descr…
Dialogue GenerationTowards Zero-Shot Personalized Table-to-Text Generation with Contrastive Persona Distillation
Existing neural methods have shown great potentials towards generating informative text from structured tabular data as well as maintaining high content fidelity. However, few of them shed light on generating personalize…
Table-to-Text GenerationText GenerationContrastive Conditional Latent Diffusion for Audio-visual Segmentation
We propose a latent diffusion model with contrastive learning for audio-visual segmentation (AVS) to extensively explore the contribution of audio. We interpret AVS as a conditional generation task, where audio is define…
Contrastive LearningDenoisingRepresentation LearningSegmentationControlled Text Generation Using Dictionary Prior in Variational Autoencoders
While variational autoencoders (VAEs) have been widely applied in text generation tasks, they are troubled by two challenges: insufficient representation capacity and poor controllability. The former results from the pos…
Contrastive LearningLanguage ModelingLanguage ModellingRepresentation Learning+1