We will give presenters 45 or 90 minutes for translation and implication tutorials, and 90 or 180 minutes for hands-on tutorials to address technical and/or policy/law aspects of ACM FAT* issues for a broad audience.
Hands-on tutorials should offer a broad audience the chance to experiment with new software packages designed to support ACM FAT* efforts. These tutorials should introduce the motivation for the tool, explain how the underlying technology works, and walk through a few example use cases of the presented software. We are also open to other, more experimental setups, if they are well justified.
Given our emphasis on accountability and transparency, tools used in the tutorials should be and should use open-source software (licensed under GPL, Apache 2.0, MIT, BSD etc.).
Presenters may assume that participants will bring their own laptop to the session.
The tutorial should be accessible to an audience with only a basic understanding of programming, as participants may not be all computer scientists.
We are interested in tutorials that aim to "translate" between disciplines; for instance, by explaining computer science concepts in a way that will be practically useful for lawyers, policy makers, and other practitioners, or by explaining legal, policy, or social science concepts in a way that will guide computer scientists in their future technical explorations.
These tutorials should be geared towards an interested, but beginning audience. Translation tutorials should situate the topic in the related literature and proceed to deeply explain that specific topic.
Implications tutorials should cover known legal, policy, medical, or socio-economic effects of unfair algorithmic systems, lack of interpretability of machine learning models, biases in the data, or other ACM FAT* related issues.
These tutorials should emphasize “real-world” implications with known examples. For instance, an implications tutorial may focus on specific case studies, walking the audience through the likely or known causes and effects of a particular ACM FAT* issue for specific individuals or communities.
We particularly encourage submissions by civil rights lawyers, policy advocates, civil society representatives, and others who work closely with individuals and communities affected by algorithmic systems and who can offer a more in-depth understanding of the processes around the use of these systems, including those for generating the datasets used by such systems.
To apply, please send a description of the proposed tutorial (max 3 pages). The tutorial description must include:
In addition, proposals for hands-on tutorials may use an additional +1 page (up to a total 4 pages) and must include include short code snippets, installation instructions, and pointers to existing sample datasets that the attendees may use. These instructions might be tested by the Tutorials PC members for selection purposes. Remember that the code should be open sourced, and that the tutorials should not be relying on commercial software that others won’t have free access to.
Submissions must be in PDF format and should be formatted according to the 2017 ACM Master Article Template.