Konrad Rieck 🌈

Konrad Rieck 🌈

@rieck.mlsec.org

Machine Learning and Security, Professor of Computer Science at TU Berlin,

415 followers114 following46 posts

Konrad Rieck 🌈 reposted

Yesterday, Lukas presented at #ICSE2026 results I didn't believe when he first showed them to me. A classifier based on simple code metrics matches frontier LLMs for vulnerability discovery in performance, as long as they are not agentic. 🧵 (1/4)

View full thread

This leaves us confused. Apparently, much of recent progress comes from the agent and the tools around the LLM rather than any inherent capability to understand vulnerable code. Which raises the question: how much of that huge model is actually needed? (3/4)

Yesterday, Lukas presented at #ICSE2026 results I didn't believe when he first showed them to me. A classifier based on simple code metrics matches frontier LLMs for vulnerability discovery in performance, as long as they are not agentic. 🧵 (1/4)

It's not just correlation. A causal analysis shows that when you change a metric value, LLM predictions shift proportionally. Despite their enormous size, these models are operating at the surface of the code, counting lines and loops 😢 (2/4)

Konrad Rieck 🌈 reposted

4 Jahre lang haben wir im Bayer. Forschungsverbund ForDaySec.de untersucht, wie IT-Sicherheit im Alltag funktioniert. Informatik, Soziologie und Recht, 5 Unis, 8 Teilprojekte. Am Mittwoch (25.3.) stellen wir die Ergebnisse in München vor. 1/2

Konrad Rieck 🌈 reposted

📣 Our paper “On the (In)Security of Loading Machine Learning Models” has been accepted at IEEE S&P 2026 (13% acceptance rate this cycle). (1/5) 👇 Preprint: arxiv.org/abs/2509.06703 #ieeesp #ieeesp26 #ieee #cybersecurity #softwaresecurity #aisecurity #machinelearning #ml #zeroday

Konrad Rieck 🌈 reposted

Lands of Packets TTL exceeded. I would like to collect texts from the scene about FX in his memory. A collection of obituaries that will then be posted on phenoelit.de. If anyone would like to contribute, please contact me. Mail: joernchen@phenoelit.de Signal: jrn.07

Jeff Moss@thedarktangent.defcon.social.ap.brid.gy

RIP FX - You are a legend

Konrad Rieck 🌈 reposted

Although we had secretly planned a small extension, we realized that the initial deadline was off by one… month 😵‍💫 Thus, we are now extending the submission deadline for the "Explainable AI and Security" (XAISEC) workshop: 🌐 https://xaisec-workshop.org ⏱️Deadline: J̶a̶n̶ ̶2̶9̶t̶h̶ Feb 26th, 2026 (AoE)

Konrad Rieck 🌈 reposted

A researcher used more than 2,000 em-dashes in his papers, revealing AI-based manipulation in 400+ papers since 1985. Professor Zeller claims he "typed" these dashes into the paper by using "two hyphens" and a "typesetting" system.

Konrad Rieck 🌈 reposted

went back and grabbed the full video of her intro. absolutely incredible 🇮🇹🍝👩‍🍳🤌🇮🇹🍝👩‍🍳🤌🇮🇹🍝👩‍🍳🤌🇮🇹🍝👩‍🍳🤌

Ken Childs@thekenchilds.com

Folks, she brought a whole plate of spaghetti to the broadcast center to do her headshot. She then houses that spaghetti. Mattia sips on an espresso during his. Truly the heroes we need during these times.

Konrad Rieck 🌈 reposted

I have been named an ACM Distinguished Member for "contributions to software security and fuzz testing". Happy and honored! A heartfelt *Thank You* to my nominator and all of you who endorsed and supported me - today and throughout my entire career. 1/2

ACM has named 61 new Distinguished Members for 2025! 🌍💡 Recognized for breakthroughs in computing and dedicated service to our community, this diverse group represents top institutions across 15 countries. Congratulations to the 2025 class! buff.ly/rajUhe6 #ACMDistinguishedMembers #computing

Konrad Rieck 🌈 reposted

🎉 Herzlichen Glückwunsch, Klaus-Robert Müller! Der #TUBerlin- Wissenschaftler und Direktor von @bifold.berlin erhält den renommierten Leibniz-Preis der @dfg.de. Er gilt als Pionier des Maschinellen Lernens und prägt die interdisziplinäre #KI- Forschung seit Jahrzehnten. (1/2) Weiterer Infos 👇

tu.berlinLeibniz-Preis für Klaus-Robert MüllerTU-Forscher und Co-Direktor des BIFOLD wird mit dem renommiertesten Forschungsförderpreis in Deutschland geehrt / Fördersumme 2,5 Millionen Euro.

Konrad Rieck 🌈 reposted

Replying to @dilya.bsky.social

Our lightweight adversarial fine-tuning attack lets you bend a feature to visualize any arbitrary concept. Off-manifold, we impose a hyperbolic activation landscape with its optimum at the target, while preserving on-distribution activations through a weighted two-term loss. 🕵️‍♀️

Konrad Rieck 🌈 reposted

✈️🇲🇽 Next Wednesday (Dec 3), 1–4 p.m. CST, I’ll be presenting Manipulating Feature Visualizations with Gradient Slingshots at NeurIPS 2025 in Mexico City! Feature Visualization has long been a staple interpretability tool. Our work shows it’s far from reliable! 🚨

Konrad Rieck 🌈 reposted

On Wednesday, November 5th, as part of our KASTEL Distinguished Lecture Series Prof. Gianluca Stringhini, PhD from Boston University will talk about ‘Improving the Moderation of False Information and Influence Campaigns on Social Media’. Everybody is welcome and stay around for the Q&A afterwards.

Konrad Rieck 🌈 reposted

Today I turn 60 – and what a journey this has been! Grateful for family, friends, students, colleagues, and a scientific community that keeps me inspired. The big party? You’ll have to wait four years, when I hit my (binary) 1,000,000th birthday. 😉