ArIA: Molecular Simulation Datasets and Code for AI Agent Development

A collection of code and data for reproducing results from the paper 'Molecular Simulations Assisted by an Artificial Intelligence Agent (ArIA)'. The dataset includes directories for model development, prompt generation, and application deployment. It was authored by Supphachok Chanmungkalakul and last updated on 2026-05-18.

Use Cases

• Fine-tuning language models for chemistry based on synthetic reasoning texts (CoT, CoVe, ToT, GoT, intrinsic reasoning) mentioned in the description.
• Developing LoRA adapters for molecular simulation tasks based on the dedicated model development directory.
• Deploying a chatbot for molecular simulation assistance based on the application deployment framework described.
• Generating synthetic prompts for chemistry AI from ORCA input files as described in the prompt generation directory.

Strengths

• Includes full code and datasets to reproduce the results of a specific research paper.
• Contains three distinct, organized directories for development, prompt generation, and deployment.
• Last updated on 2026-05-18, indicating recent maintenance.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

DataverseNL Harvested Dataverse

Collection Method

Created to accompany the paper 'Molecular Simulations Assisted by an Artificial Intelligence Agent (ArIA)'.

Freshness

Last updated 2026-05-18 06:13:41; freshness should be verified.