’Accountability’. There’s a lot of discussion about it, but it seems elusive in many real world situations. In fact, there is an alarming trend of people using automation in systems as an excuse to justify unfair treatment of others. Example: ”Oh, we can’t do anything, it’s in the system”. Or, ”I don’t have the authority […]
Avainsana: bias
Kategoriat
Algorithmic Scapegoating
Algorithm scapegoating = blaming an ”algorithm” for a social problem instead of attributing the blame appropriately to humans. This is often done with some vague and incorrect notion of ”algorithm”, not specifying what algorithm and how exactly the algorithm was at fault at carrying out its task. This misattribution is often associated with not properly […]
Just spent 1.5hrs talking to a journalist about algorithms. Sharing my notes, containing many ”unpopular opinions” that I nonetheless believe should be part of the public discussion about these topics. TL;DR: It’s easy to blame algorithms, hard to take individual responsibility. Here’s what’s wrong with the debate on algorithms: (1) algorithms are used as scapegoats […]
Introduction. Hal Daumé wrote an interesting blog post about language bias and the black sheep problem. In the post, he defines the problem as follows: The ”black sheep problem” is that if you were to try to guess what color most sheep were by looking and language data, it would be very difficult for you to conclude that […]
This report was created by Joni Salminen and Catherine R. Sloan. Publication date: December 10, 2017. Artificial intelligence (AI) and machine learning are becoming more influential in society, as more decision-making power is being shifted to algorithms either directly or indirectly. Because of this, several research organizations and initiatives studying fairness of AI and machine […]
Feature analysis could be employed for bias detection when evaluating the procedural fairness of algorithms. (This is an alternative to the ”Google approach” which emphasis evaluation of outcome fairness.) In brief, feature analysis reveals how well each feature (=variable) influenced the model’s decision. For example, see the following quote from Huang et al. (2014, p. 240): […]
Kategoriat
How to measure media bias?
Introduction Media bias is under heavy discussion at the moment, especially relating to the past presidential election in the US. However, the quality of discussion is not the way it should be; I mean, there should be objective analysis on the role of the media. Instead, most comments are politically motivated accusations or denials. This […]
Introduction. Here’s an argument: Most online disputes can be traced back to differences of premises. I’m observing this time and time again: two people disagree, but fail to see why that is, although for an outside it seems evident. Each party believes they are right, and so they keep on debating; it’s like a never-ending […]
1. Introduction Earlier, I had a brief exchange of tweets with @jonathanstray about algorithms. It started from his tweet: Perhaps the biggest technical problem in making fair algorithms is this: if they are designed to learn what humans do, they will. To which I replied: Yes, and that’s why learning is not the way to […]
Kategoriat
Belief systems and human action
Human-machine interaction on social platforms relies on human input. That input is driven by human bias. One form of bias is based on what I call ”belief systems”. Understanding different belief systems, e.g. by more advanced linguistic analysis, can yield solutions to various social problems on the Internet, such as online disputes and forming of […]
Algoritmien valta - Neutraalius ja puolueellisuus koneellisessa päätöksenteossa 