PsiZ encompasses a set of tools designed to help researchers infer psychological embeddings. PsiZ is organized into three major components: data collection, data analysis, and dataset curation.
In order to infer a psychological embedding, one requires behavioral data. Source code for the scripts and webpages are provided so that researchers can collect their own behavioral data without having to invest substantial time in development. Given the nature of hosting webpages, some assembly is required, but substantial effort has been made to make the process as painless as possible. This same source code is used to collect behavioral data at psiz.org.
The analysis code constitutes the heart of the Psiz project. It contains the algorithms that infer embeddings from behavioral data. All analysis code is available via the psiz python package hosted on GitHub.
Once behavioral data has been collected, it is pre-processed and packaged into easily digestable files. A number of datasets are hosted on the OSF psiz-datasets page. These datasets can be manually downloaded from the OSF website or using the psiz python package.
| 2020 |
|---|
| Roads, B. D., & Mozer, M. C. (accepted). Predicting the Difficulty of Human Category Learning Using Exemplar-Based Neural Networks. [ PDF ] |
| 2019 |
| Roads, B. D., & Mozer, M. C. (2019). Obtaining psychological embeddings through joint kernel and metric learning. Behavior Research Methods. doi: 10.3758/s13428-019-01285-3 [ Open Access ] |
| 2018 |
| Roads, B. D., Xu, B., Robinson, J. K., & Tanaka, J. W. (2018). The easy-to-hard training advantage with real-world medical images. Cognitive Research: Principles and Implications, 3(38). doi: 10.1186/s41235-018-0131-6 [ Open Access ] |