Academic Software and Datasets
CESSC: Causal and Non-Causal Sentence Classification
The CESSC project provides a curated dataset and a fine-tuned BERT-based model for binary classification of causal and non-causal sentences within social science texts.
Dataset and Codebase Highlights:
- 1,000 manually annotated sentences (500 causal + 500 non-causal)
- 12,000 machine-labeled sentences (supplementary general dataset)
- Scripts for model fine-tuning and evaluation
- Benchmarks and performance metrics included
The AI SRT Translator is a browser-based application designed to translate .srt subtitle files into multiple languages using Google Gemini models, while preserving formatting and timing.
Key Features:
- Subtitle translation powered by Gemini 1.5/2.0 language models
- Operates entirely on the client-side via Cloudflare Workers
- Preserves original subtitle format (timestamps and numbering)
- Adjustable translation settings (temperature, model type, delay)
Requirements:
- A valid Google Gemini API key (see app interface for setup instructions)
Supported Languages:
English, Persian, Spanish, French, Arabic, Korean, German, Italian, Russian, Turkish, Chinese (Simplified), Dutch
Upcoming Projects
Several tools and research resources are currently in development, including:
- DAG Generator: Automated generation of directed acyclic graphs (DAGs) from causal texts
- Meta-Science RAG: A custom retrieval-augmented generation (RAG) pipeline for supporting theory-building in the social sciences