ML decision-aid systems are increasingly common on the web, but their successful integration relies on people trusting them appropriately: they should use the system to fill in gaps in their ability, but recognize signals that the system might be incorrect. We measured how people's trust in ML recommendations differs by expertise and with more system information through a task-based study of 175 adults. We used two tasks that are difficult for humans: comparing large crowd sizes and identifying similar-looking animals. Our results provide three key insights: (1) People trust incorrect ML recommendations for tasks that they perform correctly the majority of the time, even if they have high prior knowledge about ML or are given information indicating the system is not confident in its prediction; (2) Four different types of system information all increased people's trust in recommendations; and (3) Math and logic skills may be as important as ML for decision-makers working with ML recommendations.
Despite a rapid expansion of machine learning (ML) across fields and industries, little is known about how to best prepare students for work in ML. Most ML courses today are taught at the college or professional level using a theoretical programming approach. Existing educational resources may not be sufficient or preferable for many audiences, particularly those who do not have a strong computer science background, who wish to attain basic understanding of and ability to use ML. We present learnings from a beginner level, semester-long actionable machine learning course taught at Massachusetts Institute of Technology meant to be accessible for students with minimal computer science knowledge. Based on analysis of survey responses and student projects, we find 5 core concepts (Multilayer Networks, Convolutional Neural Networks, Transfer Learning, Recurrent Neural Networks, and Embeddings & Generative Models) and 8 core skills (scoping a problem, choosing datasets, creating datasets, choosing models, modifying models, creating models, modifying learning rates, and training & testing) that helped lead to student self-efficacy as independent ML developers when mastered. We conclude by discussing implications of this research on effective course design and educational efforts for beginner level university courses in ML.
Purpose
The purpose of this paper is to look into the topic of IP category extensions in an entertainment setting. The main goal of the study is to explore the reciprocal spillover effect of customer experience (CX) ratings with an intellectual property (IP) in one medium (i.e. film) on the sales of the same IP in other media (i.e. comic books).
Design/methodology/approach
The study is based on 21-years of monthly top 300 comic book direct market sales data linked to the release schedule and domestic box office gross figures for films featuring Marvel and DC comic book IP appearing in the weekly top 50 films over the same time period. The analysis is based on a hierarchical linear (i.e. mixed) model to account for the nested structure of the data.
Findings
The analysis reveals that CX ratings of weekly top 50 films featuring comic book IP have a quadratic relationship with comic book sales by the two major publishers. Films receiving very good but not excellent ratings are associated with the highest levels of incremental comic book sales.
Research limitations/implications
The model is based on sales of periodical comic books in the direct market only (i.e. specialty shops) and does not account for sales of digital comics or collected editions through other channels. The analysis is also limited to IP for the two major publishers (Marvel and DC comics).
Originality/value
This study expands current knowledge on CX spillover effects between different media, contributing to entertainment and CX-literature alike.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.