Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Improving Translational Accuracy02:07

Improving Translational Accuracy

2.7K
2.7K
Pre-mRNA Processing: RNA Splicing01:36

Pre-mRNA Processing: RNA Splicing

5.4K
5.4K
Additional Subnuclear Structures02:10

Additional Subnuclear Structures

2.2K
2.2K
Upsampling01:22

Upsampling

317
Managing signal sampling rates is essential in digital signal processing to maintain signal integrity. A decimated signal, characterized by a reduced frequency range due to its lower sampling rate, can be upsampled by inserting zeros between each sample. This upsampling process expands the original spectrum and introduces repeated spectral replicas at intervals dictated by the new Nyquist frequency. To refine this zero-inserted sequence, it is passed through a lowpass filter with a cutoff...
317
Alternative RNA Splicing02:18

Alternative RNA Splicing

3.9K
3.9K
Long-term Potentiation01:35

Long-term Potentiation

55.8K
Long-term potentiation, or LTP, is one of the ways by which synaptic plasticity—changes in the strength of chemical synapses—can occur in the brain. LTP is the process of synaptic strengthening that occurs over time between pre- and postsynaptic neuronal connections. The synaptic strengthening of LTP works in opposition to the synaptic weakening of long-term depression (LTD) and together are the main mechanisms that underlie learning and memory.
55.8K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Emotion-Aware RoBERTa enhanced with emotion-specific attention and TF-IDF gating for fine-grained emotion recognition.

Scientific reports·2025
Same author

Deep LSTM-Based Transfer Learning Approach for Coherent Forecasts in Hierarchical Time Series.

Sensors (Basel, Switzerland)·2021
Same author

A Cloud-Based IoT Platform for Precision Control of Soilless Greenhouse Cultivation.

Sensors (Basel, Switzerland)·2021
Same author

A Novel Autonomous Perceptron Model for Pattern Classification Applications.

Entropy (Basel, Switzerland)·2020
Same author

A natural evolution optimization based deep learning algorithm for neurological disorder classification.

Bio-medical materials and engineering·2020
Same author

Unsupervised Pre-training of a Deep LSTM-based Stacked Autoencoder for Multivariate Time Series Forecasting Problems.

Scientific reports·2019

Related Experiment Video

Updated: Sep 14, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

690

PEGASUS-XL with saliency-guided scoring and long-input encoding for multi-document abstractive summarization.

Rawan Alsultan1, Alaa Sagheer2, Hala Hamdoun1

  • 1Department of Computer Science, College of Computer Sciences and Information Technology, King Faisal University, Hofuf, Saudi Arabia.

Scientific Reports
|July 21, 2025
PubMed
Summary

PEGASUS-XL enhances multi-document summarization by integrating saliency modeling and long-input encoding. This framework generates more coherent and informative summaries, outperforming existing methods.

Keywords:
Abstractive summarizationBARTLongformerMaximal marginal relevanceMulti-document summarizationNatural language processingPEGASUSPRIMERASBERT embeddingsSaliency modelingTF-IDF

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

527
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

642

Related Experiment Videos

Last Updated: Sep 14, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

690
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

527
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

642

Area of Science:

  • Natural Language Processing
  • Artificial Intelligence
  • Information Retrieval

Background:

  • The exponential growth of digital content necessitates efficient methods for synthesizing information.
  • Multi-Document Summarization (MDS) aims to create coherent summaries from multiple sources.
  • Existing methods face challenges in salient content selection, redundancy reduction, factual consistency, and input length limitations.

Purpose of the Study:

  • To introduce PEGASUS-XL, an enhanced abstractive summarization framework for MDS.
  • To address key challenges in MDS, including information synthesis and handling long inputs.
  • To improve the quality, coherence, and faithfulness of generated summaries.

Main Methods:

  • Developed a structured enhancement pipeline integrating lexical-semantic saliency modeling and long-input encoding.
  • Employed a hybrid scoring mechanism (TF-IDF, SBERT) with adaptive weighting for content selection.
  • Utilized Maximal Marginal Relevance (MMR) for diversity and redundancy reduction.
  • Incorporated Longformer to overcome input length limitations and fine-tuned PEGASUS for abstractive summarization.

Main Results:

  • PEGASUS-XL consistently outperformed strong baselines (BART, PRIMERA) on Multi-News and XSum datasets.
  • Achieved superior performance across multiple evaluation metrics (ROUGE, METEOR, BERTScore, SBERT similarity).
  • Human evaluations confirmed enhanced coherence, informativeness, and faithfulness of summaries.

Conclusions:

  • PEGASUS-XL offers a robust, scalable, and extensible solution for high-quality abstractive summarization in multi-document scenarios.
  • The framework demonstrates substantial quality gains without significant computational overhead.
  • Further research can address remaining issues like factual drift and residual redundancy.