DIKU Bits: Is most published research wrong?


Maria Maistro, postdoc in the Machine Learning section at DIKU.


Science is not the static knowledge written in textbooks, but it is a dynamic process carried out by researchers all over the world, who are continuously publishing their latest discoveries. New published papers can motivate future research, inspire new products, and inform governmental policy makers, therefore it is of fundamental importance to rely on and validate published results, if their conclusions are wrong, we risk to waste time and money resources and to have serious consequences on our everyday life.

In this talk, I will share with you some data about the replicability and reproducibility of experimental results, why reproducibility has become a primary concern in most areas of science, and how we are facing this open challenge.

Zooming in on Maria Maistro

Which courses do you teach?
I teach “Web Science” and “Information Retrieval”, both on the MSc programme. 

Which technology/research/projects/startup are you excited to see the evolution of?
I am looking forward to see the evolution of conversational AI and how it will be integrated in daily life, as for example with home assistants and robots. See Jibo, the first social robot.

What is your favorite sketch from the DIKUrevy?  
I actually do not know what DIKUrevy is.