Instruction
Please complete the research discussion assignment in a Jupyter or R Markdown notebook. You should post the GitHub link to your research in a new discussion thread.
Read one or more of the articles below and consider how to counter the radicalizing effects of recommender systems or ways to prevent algorithmic discrimination.
Renee Diresta, Wired.com (2018): Up Next: A Better Recommendation System
Zeynep Tufekci, The New York Times (2018): YouTube, the Great Radicalizer
Sanjay Krishnan, Jay Patel, Michael J. Franklin, Ken Goldberg (n/a): Social Influence Bias in Recommender Systems: A Methodology for Learning, Analyzing, and Mitigating Bias in Ratings
Research Discussion Assignment 4
Mitigating the Harm of Recommender Systems
According to Diresta (2018) in one of the articles provide for this task, the need to rethink the ethics of recommendation engines is only growing more urgent as curatorial systems and AI crop up in increasingly more sensitive places. Adomavicius, Bockstedt, Curley and Zhang (2014) explored two approaches to removing anchoring biases from consumer ratings.
The first approach uses post-hoc adjustment rules to systematically sanitize user-submitted ratings that are known to be biased. The investigation demonstrates the advantage of unbiased ratings over biased ratings on recommender systems’ predictive performance. Removing biases from submitted ratings using a global rule or user-specific rule was found to be problematic, most likely due to the fact that the anchoring effects can manifest themselves very differently for different users and items.
The second approach was a user-interface-based solution that tries to minimize anchoring biases at rating collection time. The researchers provided several ideas for recommender systems interface design and demonstrated that using alternative representations can reduce the anchoring biases in consumer preference ratings.
At the end of the experimental approach, they demonstrated that some interfaces are more advantageous for minimizing anchoring biases. For example, using graphic, binary, and star-only rating displays can help reduce anchoring biases when compared to using the popular numerical forms.
Companies that build recommendation systems needs to prioritize ethics over profit making because algorithms and technology has taken so much influence in our daily lives. Recommendation bias can radicalize people who binge on a certain type of movies to satisfy their desire. I believe consumer education to make consumers more cognizant of the potential decision-making biases introduced through online recommendations might help to reduce recommendation bias.
References
Adomavicius, G., Bockstedt, J., Curley, S., and Zhang, J. (2014). De-Biasing User Preference Ratings in Recommender Systems. In RecSys 2014 Workshop on Interfaces and Human Decision Making for Recommender Systems (IntRS 2014) (Foster City, CA, USA, 2014), 2–9. Retrieved from http://ceur-ws.org/Vol-1253/paper1.pdf
Renee Diresta, Wired.com (2018): Up Next: A Better Recommendation System