Historical linguists have long sought to understand how language evolves over time, leading to the formulation of various laws, including contradictory ones governing the evolution of synonymy. While recent computational work has attempted to evaluate these laws, limitations in methodologies and data have resulted in inconclusive and conflicting findings. To address these challenges, this project leverages advancements in natural language processing (NLP) to track changes in usage of individual word senses over time, and to thereby assess the validity of two long-standing linguistic laws governing the evolution of synonymy.

Scientific research uses statistical models to describe and understand real-world social and behavioral phenomena. As more data is being collected, these statistical models become larger and more complex. I will develop novel methods so that researchers can use these increasingly large and complicated statistical models to better understand the world.

Bayesian Evidence Synthesis (BES) is a method that aggregates levels of evidence from multiple studies. BES is highly flexible because it aggregates at the level of hypotheses instead of on the level of data or model parameters (as, for instance, in meta-analysis). It can therefore aggregate results of highly diverse studies (e.g., with different design, statistical model, and/or variables) as, for instance, obtained from conceptual replication. BES is however still in its infancy. The proposed (methodological) investigations will mature BES towards a broadly applicable tool, that allows for empirical quantification of the generalizability of scientific theories.

The increasing use of data driven analysis techniques in data science & AI requires a great deal of expert knowledge. When the data are incomplete, this knowledge must also include the exact specification of statistical imputation models to solve for the missingness. This project proposes mindless imputation, a data-driven automated technique for solving incomplete data problems that minimizes the need for user intervention. With mindless imputation, applied researchers can focus on the analysis, knowing that the used imputation technique allows for valid inferences and reliable predictions in the context of incomplete data.

Central in the meeting will be the question what is open science?. By inviting speakers with diverse background to share their vision on open science, we create an easy and inviting setting for discussion. Furthermore, the proceedings will be bundled and published by the open science community utrecht, serving as the first publication of the newly created OSCUpublishing platform