ACM TiiS Special Issue

ACM Transactions on Interactive Intelligent Systems (TiiS)
Special Issue on Human-Centered Machine Learning

DEADLINE: 2 December 2016


Machine learning is one of the most important and successful techniques in contemporary computer science. It involves the statistical inference of models (such as classifiers) from data. It is often conceived in a very impersonal way, with algorithms working autonomously on passively collected data. However, this viewpoint hides considerable human work of tuning the algorithms, gathering the data, and even deciding what should be modeled in the first place. Examining machine learning from a human-centered perspective includes explicitly recognizing this human work, as well as reframing machine learning workflows based on situated human working practices, and exploring the co-adaptation of humans and intelligent systems. A human-centered understanding of machine learning in human context can lead not only to more usable machine learning tools, but to new ways of framing learning computationally. This journal issue will bring together research from different disciplines that aims to create a human-centered approach to machine learning.

We invite submissions presenting novel research concerning the role of humans in machine learning systems to a special issue on Human-Centered Machine Learning to be published in the ACM Transactions of Interactive Intelligent Systems (TiiS, see more These include both interactive machine learning systems and user studies that aim to understand the role of people in machine learning (or a combination of the two). The relevant topics are listed below. Since TiiS requires that every submission must demonstrate the two defining characteristics of an interactive intelligent system (, this special issue will consider only submissions that have such two defining characteristics. Hence, not every submission that falls into one of the listed topics is relevant to TiiS.

Advancing the state of the art in interactive machine learning
* Design of new machine learning systems that is grounded in user research
* New methods of interacting with the machine learning process (e.g. interacting during training, data collection, ideation, evaluation and adaptation)
* New applications of interactive machine learning
* Evaluation of new interactive machine learning systems
* New user interaction approaches to machine learning including graphical interfaces but also physical and audio-visual interfaces

Understanding the role of people in machine learning
* User studies of machine learning systems
* Case studies of the use of a machine learning tool
* Identifying difficulties that users have with machine learning
* Understanding users’ conceptual models of machine learning
* Identifying ways of using machine learning that do not conform to standard models

Supporting Effective Use of Machine Learning
* Methods to guide users to give useful information to machine learning systems
* Methods to support debugging machine learning systems
* Studies of the factors that influence the effectiveness of users’ interactions with machine learning systems

Visualizing and Explaining Machine Learning systems
* Visualizations of the output and internal functioning of machine learning systems
* Methods of giving feedback on why a machine learning system acted as it did
* Textual, audio and other non-graphical feedback methods
* Users studies on the role of feedback in machine learning

Beyond Labels
* Methods of guiding machine learning systems beyond the standard data labeling or reward signals used in supervised or reinforcement learning
* Studies of the types of information that users want to give to machine learning systems
* Interactive machine learning systems that allow people to give multiple, diverse forms of information

* Rebecca Fiebrink (Goldsmiths, University of London) (
* Marco Gillies (Goldsmiths, University of London) (

* By December 2nd, 2016: Submission of manuscripts
* By March, 2017: Notification about decisions on initial submissions
* By September, 2017: Targeting special issue publication

Except for the initial submission deadline, these dates are indicative rather than definitive. Some submissions will be processed more quickly, while others may require more reviewing and revision. Each accepted article will be available online soon after, even if other articles for the special issue are not yet ready for publication.


Please see the instructions for authors on the TiiS website (


TiiS (pronounced “T double-eye S”), is an ACM journal for research about intelligent systems that people interact with.