RESUMEN
Identifying key statements in large volumes of short, user-generated texts is essential for decision-makers to quickly grasp their key content. To address this need, this research introduces a novel abstractive key point generation (KPG) approach applicable to unlabeled text corpora, using an unsupervised approach, a feature not yet seen in existing abstractive KPG methods. The proposed method uniquely combines topic modeling for unsupervised data space segmentation with abstractive summarization techniques to efficiently generate semantically representative key points from text collections. This is further enhanced by hyperparameter tuning to optimize both the topic modeling and abstractive summarization processes. The hyperparameter tuning of the topic modeling aims at making the cluster assignment more deterministic as the probabilistic nature of the process would otherwise lead to high variability in the output. The abstractive summarization process is optimized using a Davies-Bouldin Index specifically adapted to this use case, so that the generated key points more accurately reflect the characteristic properties of this cluster. In addition, our research recommends an automated evaluation that provides a quantitative complement to the traditional qualitative analysis of KPG. This method regards KPG as a specialized form of Multidocument summarization (MDS) and employs both word-based and word-embedding-based metrics for evaluation. These criteria allow for a comprehensive and nuanced analysis of the KPG output. Demonstrated through application to a political debate on Twitter, the versatility of this approach extends to various domains, such as product review analysis and survey evaluation. This research not only paves the way for innovative development in abstractive KPG methods but also sets a benchmark for their evaluation.
RESUMEN
Brain organoids represent a useful tool for modeling of neurodevelopmental disorders and can recapitulate brain volume alterations such as microcephaly. To monitor organoid growth, brightfield microscopy images are frequently used and evaluated manually which is time-consuming and prone to observer-bias. Recent software applications for organoid evaluation address this issue using classical or AI-based methods. These pipelines have distinct strengths and weaknesses that are not evident to external observers. We provide a dataset of more than 1,400 images of 64 trackable brain organoids from four clones differentiated from healthy and diseased patients. This dataset is especially powerful to test and compare organoid analysis pipelines because of (1) trackable organoids (2) frequent imaging during development (3) clone diversity (4) distinct clone development (5) cross sample imaging by two different labs (6) common imaging distractors, and (6) pixel-level ground truth organoid annotations. Therefore, this dataset allows to perform differentiated analyses to delineate strengths, weaknesses, and generalizability of automated organoid analysis pipelines as well as analysis of clone diversity and similarity.