Super excited to finally release our latest work where we show that AI can help humans in verifying Wikipedia claims. Everything is open-sourced: data, models, code and our new Sphere retrieval white-box engine. A step forward in general trustworthiness of information online 🚀

I'm thrilled to announce that our paper on the (in)completeness of debug information will appear in ASPLOS '23. We spot clang & gcc bugs behind debug information loss when compiling with optimizations. Preprint: arxiv.org/abs/2211.09568 The code will be open-sourced.

Apple has released a software update limiting the use of AirDrop in China in light of recent protests. AirDrop was being used by protesters to transmit info directly phone to phone, bypassing the Great Firewall of China.

The submission deadline for the special issue on "Benefits and Outlook of Program Analysis for Systems Security" has been extended. Sharpen your pencils and submit your best work by Dec 20!✍️ 📘: Computers & Security Guest editors: Daniele Cono D'Elia & Lorenzo Cavallaro journals.elsevier.com/computers-and-…

Brendan Dolan-Gavitt We provide a real bug dataset that is much larger than PyPIBugs and does not need reconstruction. Paper: proceedings.mlr.press/v162/he22a.html Repo: github.com/eth-sri/learni…

Our deep learning book goes 1.0 beta🎉 📕Forthcoming on Cambridge Univ Press (w PyTorch) 🆕JAX implementation 🆕Reinforcement Learning 🆕Gaussian Processes 🆕Hyperparameter Optimization - Thank 500 contributors & 400 univs for choosing! - Free at D2L.ai (1/5)

Announcing a holiday gift: 🎅SantaCoder - a 1.1B multilingual LM for code that outperforms much larger open-source models on both left-to-right generation and infilling! Demo: hf.co/spaces/bigcode… Paper: hf.co/datasets/bigco… Attribution: hf.co/spaces/bigcode… A🧵:

Gave Over. cacm.acm.org/magazines/2023…

Holiday hacking: a script that attempts to summarize a Ghidra-decompiled function/program using GPT-3 by recursively summarizing its dependencies in the call graph! github.com/moyix/gpt-wpre

Update to AI textbook: Understanding Deep Learning udlbook.github.io/udlbook/ Added early version of RL chapter. Improved backpropagation section so it relies less on matrix calculus. Fixed loads of small things. Thanks to everyone who sent comments!

Saturdays are for porting models! github.com/huggingface/tr…

I very quickly threw together some code for fine-tuning LLaMA. One version using PEFT+8bit, and another using (simple) pipeline parallelism for full fine-tuning. github.com/zphang/minimal…

«A distributed system is one in which the failure of a bank you didn't even know existed can render your paycheck uncashable.»

Are the efforts of the #EU sufficient to achieve gender equality? Last week, a thought-provoking discussion Helena Dalli Pina Picierno #Obiettivo5 LeContemporanee Corriere della Sera Sapienza Università di Roma

#EuroSys23 Test-of-Time AWARD Omega: flexible, scalable schedulers for large compute clusters Malte Schwarzkopf Andy Konwinski Michael Abd-El-Malek John Wilkes Congratulations! Malte Schwarzkopf

#EuroSys23 Jochen Liedtke Young Researcher AWARD Prof. Pramod Bhatotia Congratulations @pramod_bhatotia !

Just wrapped up an amazing experience as a student volunteer at EuroSys 2023! Grateful for the occasion to learn from excellent speakers and contribute to a successful event. Thanks to the general chairs Giuseppe Antonio Di Luna Leonardo Querzoni for the opportunity!

Today at WWDC we introduced a new static linker. It is a ground-up rewrite that’s up to 5x faster than ld64. The new linker is written with multicore in mind, and it’s the first production ready parallel linker officially supported for iOS development. (1/n)