https://github.com/jtrecenti/captchaoracle

Modelo Usando Oraculo
https://github.com/jtrecenti/captchaoracle

Last synced: about 2 months ago
JSON representation

Modelo Usando Oraculo

• Host: GitHub
• URL: https://github.com/jtrecenti/captchaoracle
• Owner: jtrecenti
• License: other
• Created: 2022-06-08T14:50:31.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-02-21T19:43:51.000Z (almost 2 years ago)
• Last Synced: 2024-10-17T14:34:08.802Z (3 months ago)
• Language: R
• Homepage:
• Size: 66.4 KB
• Stars: 1
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.Rmd
  • License: LICENSE

Awesome Lists containing this project

README

        
---

output: github_document

---

```{r, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>",

  fig.path = "man/figures/README-",

  out.width = "100%"

)

```

# captchaOracle

[![Lifecycle: experimental](https://img.shields.io/badge/lifecycle-experimental-orange.svg)](https://lifecycle.r-lib.org/articles/stages.html#experimental)

This package implements new dataset, loss and model operations in a `torch`/`luz` framework to address the oracle data structure discussed in [my doctorate thesis](https://jtrecenti.github.io/doutorado) (in Portuguese).

The main objects of this package are:

- `captcha_oracle()`: creates annotated datasets from an initial model and functions to access the oracle of a website.

- `captcha_dataset_oracle()`: This object is the same as `captcha_dataset()` from the `{captcha}` package, but deals with a different data structure that incorporates incomplete information provided by the websites.

- `net_captcha_oracle()`: This object is the same as `net_captcha()` from the `{captcha}` package, but can consider an initial model as a parameter in its initialization method.

- `oracle_loss()`: This object implements the oracle loss proposed in the doctorate thesis, which incorporates the incomplete information provided by the oracle in the websites.

- `captcha_accuracy_oracle()`: calculates the accuracy of the model in the fitting process dealing with the new data structure provided by `captcha_dataset_oracle()`.

The package also has an experimental feature to learn the parameters online, by automatically accessing the oracle in the web. It is implemented with the `captcha_dataset_oracle_online()` function. There is a script [in this link](https://github.com/jtrecenti/captchaOracle/blob/main/data-raw/passo_extra_iterativo.R) providing an example of this experimental feature.

This package is experimental and it was built as a tool to make simulations for my doctorate thesis. It may have breaking changes before it is considered as stable.

## License

MIT