Daniel Stratti
Rank Embeddings for Detecting Ranking Outliers
A ranking containing ordered items from a set is often required for events such as awarding prizes (ranking the participants), grants (ranking the grant applications), and accepting papers into a conference (ranking the papers). Unfortunately, we are usually unable to automate this ranking process due to the complexity of the process. Experts are required to rank the items, but rather than getting each expert to rank the entire set of items (e.g. assessing all papers submitted to a conference), the individual load is reduced by using many experts, each ranking a subset of items. The individual subrankings must then be aggregated to form an overall ranking.
In this AMSI Vacation Research Scholarship project, we will investigate the detection of ranking outliers and the novel concept of ranking embeddings, a vector representation of a given ranking that should expose outliers and allow computation of a robust aggregated ranking. To compute a ranking embedding, we must examine metrics for sub-rankings, and possible vector space projections associated the metric. Inspiration will be taken from current word embedding research.
Daniel Stratti
Western Sydney University
Daniel Stratti is currently finishing his combined Bachelor of Information and Communications Technology and Bachelor of Data Science degrees. He was drawn to these fields of study to understand how to utilise machines and the software executed, with a keen interest in the application of smart city systems. Daniel’s interest have lead him to study the areas of physics, mathematics & computer science with a thirst for knowledge and a desire to design practical applications that merge his two fields of study.
You may be interested in
Brenton Horne
Investigating spontaneous symmetry breaking of spatial Kerr solitons in fractional spatial dimensions using Fourier spectral methods.
