Fine-Tuned PEGASUS: Exploring the Performance of the Transformer-Based Model on a Diverse Text Summarization Dataset
Bibliographic record
Abstract
Text summarization is the art of succinctly capturing the essence of a lengthy text document through a concise summary. The intricate craft of text summarization involves distilling a voluminous text document into a brief and elegant summation that conveys its core message. Ultimately, the goal of text summarization is to help on grasping the essence of a text without having to wade through its entire length. In this research, we propose to fine-tune and explore the quality performance of the deep learning and transformer-based PEGASUS model for abstractive text summarization on a diverse dataset. The diversity of the dataset is expected to challenge the model and test its capabilities in generating summaries for a wide range of text types and styles. Our experimental results indeed indicate that the model's performance varies based on the topic and category of the text reaching as high as 88.03 F1 (ROUGE-1) score with some topics and as low as 81.22 with others. This is crucial as texts, such as political, economic, literary, legal, and medical, have distinctive writing conventions and styles, and a model that performs well on a diverse dataset is more likely to adapt to other text types.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.009 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.003 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".