Sandra Siby: SINBAD: Saliency-informed detection of breakage caused by ad-blocking at
Users rely on privacy-enhancing blocking tools to protect themselves against the security, privacy and usability issues caused by online advertising and tracking. An unfortunate side effect of blocking tools is that they can inadvertently cause breakage, i.e., legitimate parts of a webpage from functioning. In this talk, Sandra Siby describes SINBAD, an automated breakage detection system. Her team first conducted a study of popular forums to understand common issues in breakage detection that are not addressed by existing systems. They then used the concept of saliency -- the most relevant parts of a page -- to develop SINBAD. SINBAD's use of saliency allows it to not only trigger breakage caused by user interactions, but also shows good performance in identifying breakage. SINBAD can be used by filter-list maintainers during rule creation and updates.
Sandra Siby is a postdoctoral research associate at Imperial College London.
✅ About Ad-Filtering Dev Summit:
From exploring the latest advancements in machine learning and AI to shaping the future of online advertising, we engage in thought-provoking discussions and discover how technology can be harnessed to enhance user experience, foster sustainability, and promote transparency in the digital ecosystem.
Learn more about AFDS here: https://adfilteringdevsummit.com/publ...
✅ About eyeo
We transform the internet into a trusted, sustainable and accessible place where people and businesses regain control over their experience, content creators and publishers are rewarded for their content, and advertisers and consumers can connect on mutually agreed terms.
With eyeo …
✔️ People regain control over their online experience.
✔️ Content creators and publishers are rewarded for their content.
✔️ Advertisers and consumers connect on mutually agreed terms.
Find out more about eyeo at https://eyeo.com/about.