Latest Articles for Natural Language Processing

HQA-Attack creates high-quality adversarial examples in text while preserving meaning.

This article reviews techniques to enhance Large Language Models' efficiency and performance.

KB-Plugin improves how LLMs access and use lesser-known knowledge bases.

Research shows how style vectors can control text output in language models.

A method for speeding up large language models without sacrificing output quality.

A new method simplifies understanding of speech classification models.

This study analyzes how language models handle familiar and unfamiliar topics.

Introducing DE-BERT, a framework improving efficiency in language models through early exiting strategies.

Effective data selection enhances the performance of language models during instruction tuning.

This article discusses techniques to enhance LLMs' efficiency with lengthy text.

Using multilingual lexicons to improve sentiment analysis in low-resource languages.

New methods GliDe and CaPE boost language model response times significantly.

A new approach improves code review efficiency and quality through deep learning.

Leveraging LLMs to streamline patient recruitment for clinical trials.

This research enhances RNNs by using multiple perspectives for better text processing.

This article discusses a new method for K-VQA using LLMs for improved accuracy.

CompeteSMoE improves training efficiency and performance in Sparse Mixture of Experts models.

Creating POS taggers for languages with limited training data.

Learn methods to enhance BERT training speed and efficiency using GPUs.

LB-KBQA enhances intent recognition in financial question answering.

New indirect reasoning techniques significantly enhance language model performance in complex tasks.

A novel method enhances Vision-and-Language Navigation using training data from driving videos.

This paper discusses methods for efficiently using LLM services.

Discover how Mamba changes in-context learning for artificial intelligence applications.

A new method for better label projection in cross-lingual NLP.

SIDU-TXT sheds light on AI decisions in natural language processing.

This article presents strategies and a new method for effective sample selection in few-shot learning.

DeRa offers a method to adjust language model alignment without retraining.

Strategies for improving search engines in multiple languages with limited data.

This study investigates how retrieval-augmented models improve reliability in language generation.

A method for fine-tuning language models using fewer parameters.

NORMY framework enhances conversational QA by improving retrieval and understanding of context.

This research examines the efficiency of backward computation in training language models.

CAST improves the efficiency of self-attention in Transformer models for long sequences.

Examining the risks of data contamination in closed-source language models.

A new method combining language models with optimization strategies for improved performance.

A new method improves learning efficiency while retaining past knowledge.

Exploring the impact of attention layers on learning from text.

This study focuses on improving skill extraction using advanced language models.

A new method uses internal states for better accuracy in LLMs.