I have been lazy with posting research updates lately (or the last two years). So what happened? I was able to publish my research on using LLMs for hacking AD networks as TOSEM (dl.acm.org/doi/full/10.1145/…) and my research on LLMs for Linux PrivEsc at EMSE (link.springer.com/article/10…)
Andreas Happe retweeted
🚀 Exciting AppSec News! The OWASP Top Ten Proactive Controls 2024 is live! 🎉 Big thanks to the amazing contributors like @KatyAnton and @falkandreas for their hard work. Check out the details at top10proactive.owasp.org!
#AppSec #OWASP #SecureCoding #ProactiveControls
Andreas Happe retweeted
We had a great IPA Lab Summer closing party @tu_wien
Thank you to all students in my lab working on cool research projects and supporting our teaching efforts 🎉 – at Vienna, Austria
Andreas Happe retweeted
Will release TotalRecall in a few days. Loads to play with and to work on.
Thank you @GossiTheDog for the inspiration!
#WindowsRecall #CyberSecurity #Microsoft #TotalRecall
Andreas Happe retweeted
Last but (I hope) not least! My project, Web Check. The all-in-one security and OSINT tool, for analyzing any website. We give you x-ray vision, to see the inner workings of any website, server or app.
github.com/lissy93/web-check
Andreas Happe retweeted
@VMesel is simplify LLM deployments with a unified LLM Chat API, producing a simple Docker container
github.com/talkdai/dialog
Andreas Happe retweeted
@AkshayKagrawal and @themylesfiles developed Marimo, the next generation reactive notebook - like Jupyter, but reproducible, git-friendly, and deployable as scripts or apps.
github.com/marimo-team/marim…
Andreas Happe retweeted
@michael_vandi has built LangDrive. It enables you to train, deploy and query open source LLMs using your private data, all from one library - very nice.
github.com/addy-ai/langdrive
Andreas Happe retweeted
Namee (@nydoc53333) is the brains behind LLMWare, the unified framework for building enterprise RAG pipelines with small, specialized models
github.com/llmware-ai/llmwar…
Andreas Happe retweeted
@TimothyJBaek is building a really functional and beutiful UI for chat-based LLMs. @OpenWebUI is everything that's missing from ChatGPT's UI, plus you can use it with Ollama as well as OpenAI API.
github.com/open-webui/open-w…
Andreas Happe retweeted
@Alex_Combessie is building an awesome library for evaluating and testing AI models. It's able to automatically detect performance, bias & security issues in LLM-based applications.
github.com/Giskard-AI/giskar…
Andreas Happe retweeted
@DanielHanchen is making AI & LLMs more accessible + faster for everyone, with @UnslothAI
It's really quite an achievement to have fine-tuned these models to make them so much faster, using up to 80% less memory. Huge kudos!!
github.com/unslothai/unsloth
Andreas Happe retweeted
The legendary @dmarcos is taking @aframevr to the next-level, with the integration of AI workflows like 3D Gaussian Splatting and generative AI for images and environments
github.com/aframevr/aframe
Andreas Happe retweeted
@AndreasHappe and @citostyle are helping ethical hackers take pen testing to the next level, with an AI powered tool to discover new attack vectors and save the world
github.com/ipa-lab/hackingbu…
Andreas Happe retweeted
Good news everyone! Our project hackingBuddy.ai is part of this year’s @github accelerator 🎉🥳
hackingBuddyGPT is part of @andreashappe‘s PhD research in my lab, supported by students @neverbolt and Diana Strauß. We’re super excited what the accelerator will bring!!
AI innovation is not just thriving—it’s accelerating at a breathtaking pace across the globe and in the open. Meet the 11 projects in the 2024 cohort of GitHub Accelerator, who are shaping open source AI.
github.blog/2024-05-23-2024-…
Andreas Happe retweeted
AI innovation is not just thriving—it’s accelerating at a breathtaking pace across the globe and in the open. Meet the 11 projects in the 2024 cohort of GitHub Accelerator, who are shaping open source AI.
github.blog/2024-05-23-2024-…
Andreas Happe retweeted
If you are in Vienna tomorrow, and you want to learn about some of the nitty-gritty, but fun details about how MiniJinja works in Rust, come to Schikaneder tomorrow. codenugget.flobauer.at/
I will present a paper that's very dear to me at this year's FSE in SF (: when I came back to academia, I was surprised that there was almost no empirical research into how hackers work. Might come in handy, if you want to do research in that area.. arxiv.org/abs/2308.07057
I am very excited (and maybe a bit anxious) that I'll be presenting this paper and its preliminary results at this year's FSE'23 in San Francisco.. arxiv.org/abs/2308.00121 yeah, I've been a bit buzzword hunting lately.












