Marcel Böhme's Avatar

Marcel Böhme

@mboehme.bsky.social

Software Security @ MPI for Security and Privacy Spokesperson for Max Planck Research Group Leads at CPTS PhD @NUS, Dipl.-Inf. @TUDresden Research Group: http://mpi-softsec.github.io

898 Followers  |  504 Following  |  160 Posts  |  Joined: 17.11.2024
Posts Following

Posts by Marcel Böhme (@mboehme.bsky.social)

So true!

18.02.2026 07:07 — 👍 2    🔁 0    💬 0    📌 0
Post image

Thrilled to give a keynote at ACM India's ISEC'26 in Jaipur this Friday!

How do we know whether our program has no bugs if we have never seen it have any, and if we don't even have anything (i.e., an oracle) that can tell us whether a behavior is a bug or feature?

Stick around till Friday!

17.02.2026 16:30 — 👍 12    🔁 2    💬 1    📌 0
Preview
The Age of Academic Slop is Upon Us what happens when AI automates "normal science"?

Hopefully, it will allow us to refocus on what truly matters in science, on big problems, on intuition and insights, on theory building, and away from the brute mechanics of publishing and the bean counting that has put the breaks on true progress in science.
hegemon.substack.com/p/the-age-of...

14.01.2026 22:39 — 👍 100    🔁 18    💬 12    📌 1

Very interesting! You are suggesting there may be other domains where the proxy carries "more signal" about the expected value of the proxied random variable than the random variable itself, right?

26.12.2025 07:27 — 👍 1    🔁 0    💬 1    📌 0

Yes. For the special case where the benchmark set is size 1 (i.e., 1 program), what you describe is what we found. For Fuzzbench (primarily a coverage-based benchmark), your intuition extends to larger benchmark sets, too.

26.12.2025 07:21 — 👍 0    🔁 0    💬 0    📌 0

In fact, a ranking of fuzzers by the coverage achieved on a bunch of programs is much more representative of the ranking of fuzzers by the number of bugs found than a ranking of fuzzers by the number of bugs found itself.

Don't believe me? Wait until we release our preprint 😄.

Merry Xmas everyone!

25.12.2025 14:51 — 👍 6    🔁 0    💬 1    📌 0
Post image

Which is better? Asking your distant Uncle Barry for the Top10 restaurants in NY or consulting the Michelin Guide? Well, turns out that bug-based fuzzer benchmarking is much like Uncle Barry. Random and noisy.

Accepted at #FSE26. Led by Ardi Madadi, @is-eqv.bsky.social, and @nimgnoeseel.bsky.social

25.12.2025 14:51 — 👍 12    🔁 3    💬 1    📌 1

Thanks Carlo!

19.12.2025 11:13 — 👍 0    🔁 0    💬 0    📌 0

Many thanks, Rahul :)

19.12.2025 11:13 — 👍 0    🔁 0    💬 0    📌 0

Thank you, Konrad!

19.12.2025 11:13 — 👍 0    🔁 0    💬 0    📌 0

Thanks Adolfo!

18.12.2025 09:35 — 👍 1    🔁 0    💬 0    📌 0

Thank you to all of my former and current students, to my friends, and particularly to my dear family who have been my greatest support in my life. Thank y'all so much – I would not be where I am without you. ❤️

2/2

18.12.2025 08:25 — 👍 7    🔁 0    💬 0    📌 0

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

18.12.2025 08:25 — 👍 30    🔁 4    💬 5    📌 0
Starter Pack: Max Planck scientists on Bluesky

Starter Pack: Max Planck scientists on Bluesky

Many Max Planck scientists have started sharing their #research findings on #BlueSky. Follow their posts and join the conversation! 👋 go.bsky.app/BYcBy6R #StarterPack

17.11.2025 11:15 — 👍 116    🔁 41    💬 5    📌 2
Preview
ISSTA 2026 - Research papers - ISSTA 2026 Welcome to the website of the ISSTA 2026 conference. The ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) is the leading research symposium on software testing and analysis...

📢 Call for Papers for ISSTA 2026

We invite high-quality submissions on software testing and analysis from industry and academia, incl.
* research papers
* experience papers, and
* replicability studies.

📆 29th January 2026
🖊️ issta2026.hotcrp.com
🌐 conf.researchr.org/track/issta-...

17.12.2025 08:33 — 👍 5    🔁 2    💬 0    📌 0
Böhme Research Group: Home

@mboehme.bsky.social 👋

MPI for Security and Privacy (Software Security group)
Spokesperson for MPRGL at CPTS
mpi-softsec.github.io

16.12.2025 14:46 — 👍 1    🔁 0    💬 1    📌 0

⏱️ 9 days until submission deadline (Dec 11, 23:59 AoE).

Organized by: @yannicnoller.bsky.social, @rohan.padhye.org, @ruijiemeng.bsky.social, and Laszlo (@lszekeres.bsky.social) Szekeres.

03.12.2025 10:59 — 👍 4    🔁 5    💬 0    📌 0
Post image Post image Post image

Dearly beloved, we are gathered here today to celebrate this thing called ASE 2025 ;) @aseconf.bsky.social @mboehme.bsky.social @llingming.bsky.social

17.11.2025 03:18 — 👍 15    🔁 3    💬 0    📌 0
Post image

🎙️ #ASE2025 Keynote Speaker Series (1 of 3)

What do symbolic model checking, path profiling, and quantum simulation have in common? 🤔

Find out from Prof. Reps (University of Wisconsin-Madison) in his ASE2025 Keynote “We Will Publish No Algorithm Before Its Time”!

conf.researchr.org/track/ase-20...

22.10.2025 11:39 — 👍 10    🔁 3    💬 0    📌 1
Post image

🎙️ ASE 2025 Keynote Speaker Series (3 of 3)

Prof. Taesoo Kim (Georgia Tech)
“Hyperscale Bug Finding and Fixing: DARPA AIxCC”

conf.researchr.org/track/ase-20...

28.10.2025 07:44 — 👍 4    🔁 2    💬 1    📌 0
Post image

🎙️ #ASE2025 Keynote Speaker Series (2 of 3)

Dr. Cristina Cifuentes, Vice President @ Oracle Software Assurance

“Oracle Parfait – Detecting Application Vulnerabilities at Scale – Past, Present and Future”

26.10.2025 03:19 — 👍 6    🔁 2    💬 1    📌 2

Awesome! Also, I'll be happy to catch up in Seoul in the week after next if you are around for ASE :)

09.11.2025 13:29 — 👍 0    🔁 0    💬 1    📌 0

On the negative side, the AI reviewer seems to be worse at setting priorities, i.e., distinguishing between critical and insubstantial problems w.r.t. to the main claims. Moreover, it was convincingly incorrect whereas a human reviewer might be incorrect and detectably "silent" on the rationale.
2/2

09.11.2025 08:59 — 👍 0    🔁 0    💬 0    📌 0

Great question!

On the positive side, I found the AI reviewer *way* more elaborate in eliciting both the positive and negative points. The review is more objective, less/not opinionated. It is more constructive and for every weakness makes suggestions for improvements.

1/

09.11.2025 08:50 — 👍 0    🔁 0    💬 1    📌 0

Exactly. This is our assumption. Also, there can be infinitely many ways to implement that function.

09.11.2025 07:44 — 👍 1    🔁 0    💬 2    📌 0

bsky.app/profile/mboe...

08.11.2025 20:04 — 👍 0    🔁 0    💬 0    📌 0

Overall, the AI reviewer is super impressive! I think, it would help me tremendously during the preparation of our submission to identify points to improve before the paper is submitted.

However, it does make errors, and I wouldn't trust it as an actual (co)-reviewer.

12/12

08.11.2025 19:51 — 👍 1    🔁 0    💬 1    📌 0
Post image

The AI reviewer lists several other items as weaknesses and the corresponding suggestions for improvement. These are summarily deemed to be fixable. Yay!

11/

08.11.2025 19:51 — 👍 0    🔁 0    💬 1    📌 0
Post image

The fourth weakness is a set of presentation issues. These are helpful but easily fixed.

10/

08.11.2025 19:51 — 👍 0    🔁 0    💬 1    📌 0
Post image

The third weakness is a matter of preference.

Our theorem expresses what (and how efficiently) we can learn about detecting non-zero incoherence given the alg. output: "If after n(δ,ε) samples we detect no disagreement, then incoherence is at most ε with prob. at least 1-δ".

9/

08.11.2025 19:51 — 👍 0    🔁 0    💬 1    📌 0