Latest Articles for Natural Language Processing

U-DiT TTS system enhances natural speech generation through innovative architecture.

Combining multiple models can lead to clearer and more accurate text generation.

A new method enhances retrieval quality by refining dual encoder models.

A new method improves computer understanding of spoken commands with fewer examples.

A novel method enhances language models' commonsense reasoning capabilities.

A new approach to compress Transformer models into modular layers for enhanced efficiency.

Research highlights how images improve understanding of language in models.

This article examines how model scaling impacts parameter-efficient tuning methods.

This research presents a method to balance creativity and factual accuracy in language models.

Polyglot-Ko aims to create high-quality language models for Korean speakers.

V-GLOSS improves visual descriptions for better image classification and generation.

This study highlights the effectiveness of tailored language models using Stack Overflow data.

A simple method improves efficiency in retrieval systems by simplifying query processing.

A new method boosts information retrieval efficiency with fewer data needs.

This article discusses tackling inconsistencies in natural language explanations for AI.

A call for better research practices in developing language models.

This study assesses middle training methods for improved information retrieval in language models.

New model enhances emotional understanding in dialogues.

A method to reduce GPT-2's size while keeping performance intact.

A new method improves text classification accuracy with minimal labeled data.

Innovative framework enhances knowledge graphs for better task performance.

New approaches improve grammar-based AI models for sequence generation.

This research addresses forgetting in AI through continual learning in spoken language understanding.

This study investigates word alignment challenges and optimal transport solutions.

A novel approach combines cloze tasks and data augmentation for improved QA performance.

This article examines efficient methods for debiasing language models.

Improving language models' ability to answer complex questions using knowledge graphs.

This method improves VQA performance using unlabeled images for question generation.

This study introduces a model for more accurate assessment of text generation systems.

This article compares BERT and CLIP in text comprehension and image connections.

New method enhances inclusion of important medical terms in care instructions.

CUE framework improves reliability of PLM predictions by addressing uncertainty.

A look into improving image captioning with user input and data techniques.

Introducing an unsupervised approach for creating concise text summaries efficiently.

Discover how in-context learning reshapes language model capabilities.

A new method enhances response selection in dialogue systems through better context understanding.

Study adapts adversarial attacks from English to Chinese NLP systems.

A new method improves neural network efficiency and performance in machine learning.

ScienceBenchmark offers a new benchmark for complex scientific databases.

This article presents a modular design to improve language models' efficiency and adaptability.