Related papers: Serialized Output Training by Learned Dominance

Self-supervised learning (SSL) learns knowledge from a large amount of unlabeled data, and then transfers the knowledge to a specific problem with a limited number of labeled data. SSL has achieved promising results in various domains. This…

Recent developments in unsupervised representation learning have successfully established the concept of transfer learning in NLP. Mainly three forces are driving the improvements in this area of research: More elaborated architectures are…

Self-supervised learning (SSL) has emerged as a promising paradigm for learning flexible speech representations from unlabeled data. By designing pretext tasks that exploit statistical regularities, SSL models can capture useful…

The performance of optimization algorithms relies crucially on their parameterizations. Finding good parameter settings is called algorithm tuning. The sequential parameter optimization (SPOT) package for R is a toolbox for tuning and…

Structured outputs are essential for large language models (LLMs) in critical applications like agents and information extraction. Despite their capabilities, LLMs often generate outputs that deviate from predefined schemas, significantly…

Many recent source separation systems are designed to separate a fixed number of sources out of a mixture. In the cases where the source activation patterns are unknown, such systems have to either adjust the number of outputs or to…

Unsupervised sentence representation learning aims to transform input sentences into fixed-length vectors enriched with intricate semantic information while obviating the reliance on labeled data. Recent strides within this domain have been…

This work proposes a novel and simple sequential learning strategy to train models on videos and texts for multimodal sentiment analysis. To estimate sentiment polarities on unseen out-of-distribution data, we introduce a multimodal model…

Task-oriented dialogue (TOD) system is designed to accomplish user-defined tasks through dialogues. The TOD system has progressed towards end-to-end modeling by leveraging pre-trained large language models. Fine-tuning the pre-trained…

The ability to learn from limited data, or few-shot learning, is a desirable and often critical requirement for NLP systems. While many existing methods do poorly at learning from a handful of examples, large pretrained language models have…

In a noisy environment, a lossy speech signal can be automatically restored by a listener if he/she knows the language well. That is, with the built-in knowledge of a "language model", a listener may effectively suppress noise interference…

This paper introduces a novel training framework called Focused Discriminative Training (FDT) to further improve streaming word-piece end-to-end (E2E) automatic speech recognition (ASR) models trained using either CTC or an interpolation of…

Spoken language understanding (SLU) tasks involve mapping from speech audio signals to semantic labels. Given the complexity of such tasks, good performance might be expected to require large labeled datasets, which are difficult to collect…

This paper proposes a textless training method for many-to-many multilingual speech-to-speech translation that can also benefit the transfer of pre-trained knowledge to text-based systems, text-to-speech synthesis and text-to-speech…

Overlapping speech diarization is always treated as a multi-label classification problem. In this paper, we reformulate this task as a single-label prediction problem by encoding the multi-speaker labels with power set. Specifically, we…

Speaking Style Recognition (SSR) identifies a speaker's speaking style characteristics from speech. Existing style recognition approaches primarily rely on linguistic information, with limited integration of acoustic information, which…

Topic classification systems on spoken documents usually consist of two modules: an automatic speech recognition (ASR) module to convert speech into text and a text topic classification (TTC) module to predict the topic class from the…

Amos et al. (2024) showed that the accuracy of Transformer models in sequence classification can be significantly improved by first pretraining with a masked token prediction objective without external data or augmentation, a procedure…

Speech-to-text translation (ST), which directly translates the source language speech to the target language text, has attracted intensive attention recently. However, the combination of speech recognition and machine translation in a…

Many recent studies have focused on fine-tuning pre-trained models for speech emotion recognition (SER), resulting in promising performance compared to traditional methods that rely largely on low-level, knowledge-inspired acoustic…