Publications by authors named "Mayana Pereira"
Article Synopsis
- Research shows that search engine algorithms, like Bing's, don't typically expose users to unreliable websites.
- Most interactions with unreliable sources happen when users specifically search for those sites, which is a very small part of overall searches.
- This indicates that user preferences play a significant role in engaging with unreliable information, rather than the algorithms themselves.
Article Synopsis
- Differentially private synthetic datasets help people share data while keeping personal information safe, which is important in areas like health care and charity work.
- The study looks at how well synthetic data can work instead of real data in machine learning, checking which methods are best for creating useful and fair data.
- It finds that a specific type of synthetic data (marginal-based) is better than another type (GAN-based) for training machine learning models, showing that it can be just as good as using real data.
