|
|
--- |
|
|
title: The Synthetic Nomological Net |
|
|
emoji: 🔍 |
|
|
colorFrom: purple |
|
|
colorTo: blue |
|
|
sdk: docker |
|
|
suggested_hardware: "cpu-basic" |
|
|
app_port: 7860 |
|
|
pinned: true |
|
|
models: |
|
|
- "magnolia-psychometrics/surveybot3000" |
|
|
- "sentence-transformers/all-mpnet-base-v2" |
|
|
license: mit |
|
|
--- |
|
|
|
|
|
# SynthNet Search |
|
|
|
|
|
Mapping over 74,000 scales from more than 31,500 APA PsycTests Questionnaires, Surveys, and Tests. |
|
|
|
|
|
## Setup |
|
|
|
|
|
This application requires the following environment variables: |
|
|
|
|
|
- `MODEL__ITEM_MODEL_PATH`: HuggingFace model respository path |
|
|
- `MODEL__ITEM_MODEL_ACCESS_TOKEN`: Access token for private HuggingFace model |
|
|
- `MODEL__SCALE_MODEL_PATH`: HuggingFace model respository path |
|
|
- `MODEL__SCALE_MODEL_ACCESS_TOKEN`: Access token for private HuggingFace model |
|
|
- `DATA__DATASET_PATH`: HuggingFace dataset respository path |
|
|
- `DATA__ENCRYPTION_KEY`: Encryption key for database |
|
|
- `TRACKING_DB__API_ENDPOINT`: Usage tracking database endpoint |
|
|
- `TRACKING_DB__ANON_KEY`: Anonymous key for usage tracking database endpoint |
|
|
- `TRACKING_DB__JWT_TOKEN`: Access token for usage tracking database endpoint |
|
|
|
|
|
Configure these in the Space settings under "Repository secrets". |
|
|
|
|
|
## Citation |
|
|
``` |
|
|
Hommel, B. E., Külpmann, A. I., & Arslan, R. C. (2025). The Synthetic Nomological Net: A search engine to identify conceptual overlap in measures in the behavioral sciences. PsyArXiv. Manuscript in preparation. https://osf.io/preprints/psyarxiv/nfgmv_v1/ |
|
|
``` |
|
|
|
|
|
## About |
|
|
|
|
|
This research is part of the [SYNTH research project (#546323839)](https://gepris.dfg.de/gepris/projekt/546323839?language=en), which is conducted as part of the [META-REP Priority Program](https://gepris.dfg.de/gepris/projekt/441890184) aiming to improve the replicability, reproducibility, and generalizability of empirical research in the behavioral sciences. SYNTH contributes to these goals by integrating large language models into current research workflows to reduce burdens on scientists while improving transparency and replicability. |
