DABA- Behavior Analysis Tools Leveraging Local AI Platform

Understanding animal behavior is critical for advancing various scientific fields, including neuroscience, ecology, and evolutionary biology. Traditional methods for analyzing animal behavior are often labor-intensive and subjective. While pose-tracking tools like DeepLabCut (DLC) and SLEAP automate the extraction of pose-estimation, translating this data into meaningful behavioral insights remains challenging. To address this gap, I introduce DABA (Dynamic Animal Behavior Analysis), an open-source software that empowers researchers with a powerful and flexible toolkit for behavioral analysis. DABA uniquely combines a set of pre-built modules for standard analyses (such as spatial analysis, movement analysis, and event duration) with a dynamic code generation system powered by a local Large Language Model (LLM). This dual approach allows researchers to quickly quantify common behavioral metrics while also providing the flexibility to create custom analyses tailored to specific research questions, even without coding expertise. By leveraging local processing with an LLM, DABA ensures data privacy and security while enabling users to customize analyses. Adaptation of DABA by the community and its customizability have the potential to significantly streamline and enhance the analysis of animal behaviors workflows. Introduction Paper: https://github.com/farhanaugustine/DABA-Dynamic_Animal_Behavior_Analysis/blob/main/DABA_.pdf Github Repo: https://github.com/farhanaugustine/DABA-Dynamic_Animal_Behavior_Analysis Use with caution! LLMs are prone to errors. User discretion is advised.