Tag: data-science
Reproduce Research Methods and Benchmark Metrics
This skill automates the process of implementing a new research method into a structured Jupyter notebook, ensuring reproducibility by utilizing existing data splits and dependencies. It systematically records all required metrics and imple…
Systematic research method implementation and benchmarking
This skill automates the systematic implementation of novel research methods into a structured Jupyter notebook. It ensures fair benchmarking by using existing data splits and logging all dependencies and metrics into a structured JSON log …
Experiment Management and Setup Skill
Automates the initialization and bookkeeping of research experiments by setting up directory structures, materialising research sources, and managing JSON-based experiment logs.
Experiment Benchmarking and Metric Extraction
Defines comparison metrics and extracts baseline values from notebook outputs to update experiment logs for downstream evaluation.
Analyze Current Baseline Implementation
Extracts model architectures, preprocessing steps, and hyperparameters from existing Jupyter notebooks to create a structured experiment log. It also identifies dependencies and generates a requirements file from the notebook's import state…
Automated Experiment Benchmarking Skill
Defines comparison metrics and extracts baseline values from notebook outputs to record them in a structured JSON log for downstream evaluation.
Analyse Current Experiment Baseline
Analyses existing Jupyter notebooks to extract model architectures, preprocessing steps, and hyperparameters, recording the findings in a structured experiment log.