A new task designed to help language models better generalize information.
― 6 min read
Computer Science - Computation and Language
RSSNew dataset improves language models' understanding across languages.
― 7 min read
A study on the reasoning behind document-level relation extraction model predictions.
― 5 min read
A new method enhances text-to-speech quality and emotional expression.
― 5 min read
Boosting language models' performance for languages with limited data.
― 5 min read
A method to detect systematic failures in multimodal models before real-world use.
― 5 min read
New benchmark assesses language models' understanding of human thoughts and feelings.
― 6 min read
Techniques to reduce model size while preserving performance are emerging.
― 4 min read
This project uses NLP to assess the quality of classroom discussions automatically.
― 4 min read
A new framework simplifies data retrieval for language researchers on TalkBank.
― 5 min read
AI models struggle to assist researchers with paper revisions based on reviewer feedback.
― 7 min read
Examining how AI handles human-like reasoning and its biases.
― 5 min read
A closer look at the problems with the RVL-CDIP benchmark dataset.
― 5 min read
Research investigates stacked layers in language models for enhanced performance.
― 7 min read
VisoGender dataset examines gender bias in image and text models in professional contexts.
― 5 min read
A new model reduces size while improving multilingual speech recognition.
― 6 min read
This paper examines LLMs' challenges with semantic concepts like universal quantification.
― 5 min read
Investigating how language models process and attribute feature interactions for improved performance.
― 8 min read
A new method improves speech recognition accuracy for African accents.
― 5 min read
New methods improve translation accuracy while using specific vocabulary.
― 6 min read
Study reveals how to identify text samples vulnerable to NLP attacks.
― 5 min read
A new method for detecting fake news through focused analysis.
― 6 min read
A dataset offering insights into how the Russo-Ukrainian war is reported worldwide.
― 6 min read
A guide to enhancing AI's ability to write jokes.
― 7 min read
This study examines ways to enhance multiple-choice question evaluations using automated systems.
― 6 min read
This paper highlights the importance of explainable evaluation metrics in machine translation.
― 7 min read
A new model forecasts when online chats may become harmful.
― 6 min read
New methods aim to reduce outliers in transformer models, enhancing quantization efficiency.
― 6 min read
A new method for enhancing soft prompt tuning efficiency and performance.
― 6 min read
Efforts to improve multilingual metrics for dialogue systems showcased in recent challenge.
― 8 min read
A new approach for improving knowledge base accuracy and usability.
― 7 min read
A look at how NLG is changing ad creation methods.
― 6 min read
A new system improves speech recognition in multi-speaker settings.
― 6 min read
New methods improve efficiency in training smaller neural models.
― 5 min read
A framework to improve dialog systems using user feedback for better responses.
― 7 min read
An overview of MLLMs and their applications in AI.
― 7 min read
Examining changes in vaccine sentiment during the COVID-19 pandemic.
― 6 min read
New method improves predictions by considering multiple expert scores.
― 6 min read
A structured approach to enhance transformer models using knowledge graphs.
― 7 min read
LeakDistill enhances AMR parsing using structural information and knowledge distillation.
― 6 min read