Computer Science Colloquium
Upcoming:
November 18 at noon in Jepson Hall 120
Speaker: Timothy Merino is a PhD Candidate at New York University Tandon Engineering in the Game Innovation Lab
Title: Dream-Cubed: Imagining New Minecraft Worlds with AI
Abstract: How can we use AI to help us create new Minecraft environments? Introducing Dream-Cubed, a dataset containing millions of chunks of Minecraft terrain, including procedurally generated landscapes and human-built cities. We use this data to train generative diffusion models, which can create new chunks of the game world conditioned on a desired biome using human-authored block patterns to generate things like volcanos and waterfalls that aren’t in the original game. We discuss how our dataset and trained diffusion models unlock new possibilities in the field of interactive 3D environment generation.
Past Events:
October 21 at noon in Jepson Hall 103B
Speaker: William Enck is the Goodnight Distinguished Professor in Security Sciences in the Department of Computer Science at the North Carolina State University where he is co-director of the Secure Computing Institute (SCI) and member of the Secure Software Supply Chain Center (S3C2). His research spans the broad area of systems security, applying a range of systems design, program analysis, and empirical studies. He is President of the USENIX Board of Directors and has previously co-chaired the program committees of flagship security conferences including USENIX Security and IEEE S&P.
Title: Securing the Software Supply Chain
Abstract: Open source software is an economic driving force behind nearly all software products. However, the past half decade has seen a surge of attacks targeting critical but often overlooked parts of this software supply chain. The response from industry and government has been a frenzy of frameworks, regulations, tools, and best practices. But where does academia fit in?
The term “software supply chain” does not capture the nuance of the space. For the vast majority of open source software, there is no formal “supplier” and the “chain” is a complex interconnected graph. Existing tools that help developers manage known vulnerabilities in their dependencies rely on imperfect and imprecise data. Build processes were created based on threat models that no longer hold. Development environments make it too easy for developers to make choices that open both their projects and their own workstations to attack. The problems that underlie these challenges are not foreign to computer security research, but they require a partnership with practitioners to solve. This talk will describe what we have learned through our research and our extensive interactions with >150 practitioners from >50 companies as part of the NSF-funded Secure Software Supply Chain Center (S3C2).