Tural Mammadov (@turikmammadov) 's Twitter Profile
Tural Mammadov

@turikmammadov

Doctoral Researcher at CISPA 🇩🇪 – Helmholtz Center for Information Security

ID: 1189716860410978304

linkhttps://cispa.de/en/people/c01tuma calendar_today31-10-2019 01:34:48

14 Tweet

25 Followers

44 Following

Andreas Zeller (@andreaszeller) 's Twitter Profile Photo

I got a 2.5 Million EUR European Research Council (ERC) Advanced Grant #ERCAdG (my 2nd)! In the project "Semantics of Software Systems" (S3), we will research massive generation of tests and oracles for software. Details (and proposal!) here: cispa.de/s3 Come work with me!

Tural Mammadov (@turikmammadov) 's Twitter Profile Photo

"Harry Potter and the Cursed Child" musical at the Princess Theater Melbourne is one of the best performances I have seen so far. You should definitely see it, because it's hard to express the emotions with words.

"Harry Potter and the Cursed Child" musical at the Princess Theater Melbourne is one of the best performances I have seen so far. You should definitely see it, because it's hard to express the emotions with words.
Andreas Zeller (@andreaszeller) 's Twitter Profile Photo

Back in the days, we didn’t count papers, we counted ideas. If an applicant had ONE good idea, we’d invite them for an interview. If they had TWO good ideas, we’d give them tenure. And if they had THREE good ideas… well, no one ever had three good ideas.

Tural Mammadov (@turikmammadov) 's Twitter Profile Photo

I was happy to attend the tutorial on "Larger-scale model training on multi-GPU systems" by Giuseppe Fiameni from Nvidia at ELLIS Summer School on Large-Scale AI for Research and Industry. Lots of insights about capabilities of the CUDA backend of the PyTorch framework.

Andreas Zeller (@andreaszeller) 's Twitter Profile Photo

Learning models from programs! Given a program P, our MODELIZER learns a model M that mocks P's behavior, producing P's output for a given input. But M is also reversible, predicting inputs for which P produces a given output, with up to 95.4% accuracy: arxiv.org/abs/2407.08597 🧵

Learning models from programs! Given a program P, our MODELIZER learns a model M that mocks P's behavior, producing P's output for a given input. But M is also reversible, predicting inputs for which P produces a given output, with up to 95.4% accuracy: arxiv.org/abs/2407.08597 🧵
Alexander Koller (@alkoller) 's Twitter Profile Photo

It was fun to apply #NLProc methods to software engineering with my brilliant colleague Andreas Zeller and his student Tural Mammadov. The coolest part, to me, is that you can backtranslate program outputs into program inputs. Let's see where this will go!