The Database Filter That Outsmarts Adversarial Queries?! [SIGMOD 2024] with Marco Costa скачать в хорошем качестве

The Database Filter That Outsmarts Adversarial Queries?! [SIGMOD 2024] with Marco Costa

Welcome back to Disseminate – The Computer Science Research Podcast, where we make world-class research accessible for practitioners, researchers, and students alike. In this episode, Jack Waudby talks with Marco Costa (Huawei Research) about his SIGMOD 2024 paper “Graffiti: Taming Adversarial Queries with Optimal Range Filters.” They dive into: ⚙️ What are range queries and adversarial workloads in modern databases? 💥 How existing heuristic range filters break under adversarial access patterns. 🧩 How Graffiti achieves optimal space and time guarantees while remaining simple to implement. 🪣 The new “Bucketing” heuristic and its surprising performance gains. 🧠 Lessons learned from turning a master’s thesis into a SIGMOD paper. 👉 Watch to learn how this breakthrough bridges theory and practice — and why “simple solutions” often win in systems research. 📘 Referenced Paper: Graffiti: Taming Adversarial Queries with Optimal Range Filters (SIGMOD 2024) https://arxiv.org/pdf/2311.15380 🎓 Guest: Marco Costa, Research Engineer at Huawei 🎙️ Host: Jack Waudby 📅 Series: Cutting-Edge Computer Science Research If you enjoy discussions that narrow the gap between academia and industry, subscribe for more deep-dives into the latest database, AI, and systems research. #ComputerScience #DatabaseResearch #SIGMOD #Graffiti #AdversarialQueries #RangeFilters #BloomFilter #DataStructures #ResearchPodcast #AIResearch #SystemsEngineering #DisseminatePodcast #AcademicToIndustry