Towards scalable summarization and visualization of large text corpora (abstract only)
Bibliographic record
Abstract
Society is awash with problems requiring the analysis of vast quantities of text and data. From detecting flu trends out of twitter conversations to finding scholarly works answering specific questions, we rely more and more on computers to process text for us. Text analytics is the application of computational, mathematical, and statistical models to derive information from large quantities of data coming primarily as text. Our project provides fast and effective text-analytics tools for large document collections, such as the blogosphere. We use natural language processing and database techniques to extract, collect, analyze, visualize, and archive information extracted from text. We focus on discovering relationships between entities (people, places, organizations, etc.) mentioned in one or more sources (blog posts or news articles). We built a custom solution using mostly off-the-shelf, open-source tools to provide a scalable platform for users to search and analyze large text corpora. Currently, we provide two main outlets for users to discover these relations: (1) full-text search over the documents and (2) graph visualizations of the entities and their relationships. This provides the user with succinct and easily digestible information gleaned from the corpus as a whole. For example, we can easily pose queries like which companies were bought by Google? as entity:google relation:bought. The extracted data is stored on a combination of the noSQL database CouchDB and Apache's Lucene. This combination is justified as our work-flow consists of offline batch insertions with almost no updates. Because we support specialized queries, we can forgo the flexibility of traditional SQL solutions and materialize all necessary indices, which are used to quickly query large amounts of de-normalized data using MapReduce. Lucene provides a flexible and powerful query syntax to yield relevant ranked results to the user. Moreover, its indices are synchronized by a process subscribed to the list of database changes published by CouchDB. The graph visualizations rely on CouchDB's ability to export the data in any format: we currently use a customized graph visualization relying on XML data. Finally, we use memcached to further improve the performance, especially for queries involving popular entities.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".