--
AI and Machine Learning are increasingly used to detect patterns in large and complex data sets allowing them to make recommendations, detect anomalies, automate physical systems, recognize speech and images, and classify, among others. From national defense to business to personal entertainment, AI/ML is ubiquitous and increasingly requires a multidisciplinary approach.
For this workshop, we are specifically concerned with the experiences of humans when interacting with Interactive Machine Learning (IML) systems with the aim of speeding it up and improving the quality of the training period as a continuous process of learning. IML is a new way to include human inputs to improve the performance and predictions of ML model. Although human engagement with ML model helps with the training of the model and accuracy of decision-making process, it also comes with its own challenges.
Challenges can be related to the time of labeling, the speed of interaction with the system, interaction modalities and emotional state of the person labeling data, trust to the system and the different world views of humans and machines, to name a few.
In this workshop we suggest that to improve the engagement and interaction between users and the ML models, further contributions from the HCI community are needed. That is, beyond the computational issues, IML systems will also benefit from a sustained focus on understanding user needs, their emotional and physical states, and the design of IML interfaces and interactions in general.