Kareem H. El-Safty

2K posts

Kareem H. El-Safty banner
Kareem H. El-Safty

Kareem H. El-Safty

@Kero_qml

Quantum machine learning Researcher @ Wigner Research Centre for Physics | @qiskit Advocate | MSc Candidate @La_UPM | Communications & Electronics Engineer

Munich, Bavaria Katılım Kasım 2014
843 Takip Edilen486 Takipçiler
Sabitlenmiş Tweet
Kareem H. El-Safty
Kareem H. El-Safty@Kero_qml·
1) In this new research, that has been also accepted at @IEEEQuantumWeek "qce", we provide a new way to use a lesser number of qubits to encode graph colouring problems. We managed to achieve n*log(k) qubits instead of n*k where n is the number of nodes and k is the colours
English
1
1
16
0
Kareem H. El-Safty retweetledi
🇹🇷Perde Arkası خلف الستار
🇹🇷Perde Arkası خلف الستار@perde_arkasi1·
" وتبين أن الهجوم الذي أصاب يهوديين في لندن لم يكن حادثا ارهابيا ضد معاداة الساميه . فبالرغم من تفرد الإعلام عن طعن يهوديين . " تبين أن المهاجم طعن شخصا ثالثا مسلم أيضا لم يتم ذكره في أي وسائل الإعلام ، لاظهار الحادثه معاديه للساميه . . لقد تمكنوا بسبب الحادثه من جني 25 مليون جنيه إسترليني. وتمكنوا من حظر الاحتجاجات المناهضة للإبادة الجماعية.
🇹🇷Perde Arkası خلف الستار tweet media🇹🇷Perde Arkası خلف الستار tweet media
العربية
11
648
1.9K
86.8K
Amal
Amal@Amal69084155·
انت متخيل إن دكتور من صغره كان في مدارس المتفوقين، من أقوى وأعلى مدارس وقتها، الأول على مدرسته طول فترة دراسته، وطالع من ثانوية عامة بنسبة 101% والأول على محافظة القاهرة واتكرم من المحافظ وقتها. دخل كلية الطب وكان من الأوائل على دفعته طول فترة الجامعة، وكمل ماجستير ودكتوراه 👇
Amal tweet media
العربية
86
56
261
89.3K
Kareem H. El-Safty retweetledi
nic carter
nic carter@nic_carter·
Many are wondering "what Google saw" that caused them to revise their post-quantum cryptography transition deadline to 2029 last week. It was this: research.google/blog/safeguard…
English
97
958
7.5K
7M
Kareem H. El-Safty retweetledi
alphaXiv
alphaXiv@askalphaxiv·
Yann LeCun and his team can't stop cooking "LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels" One of the biggest bottlenecks of JEPA is they are hard to train, and this new research changes that. They propose LeWorldModel, which shows that a small model can learn a usable world model directly from raw pixels end-to-end. Sitting at 15M parameters, they made it without needing heuristics and avoiding anti-collapse hacks while staying competitive and planning up to 48x faster. Making JEPA based modeling much more accessible, cheaper, and stabler.
alphaXiv tweet media
English
41
242
1.8K
196.1K
Kareem H. El-Safty retweetledi
Oliver Prompts
Oliver Prompts@oliviscusAI·
🚨 BREAKING: China has open-sourced a massive Python framework for building AI agents. It’s called AgentScope, a python framework built around Agent-Oriented Programming that lets you build AI agents visually with MCP tools, memory, rag, and reasoning capabilities. 100% Open Source.
Oliver Prompts tweet media
English
36
192
961
56.9K
Kareem H. El-Safty retweetledi
Matt Dancho (Business Science)
Matt Dancho (Business Science)@mdancho84·
This is huge. A group of 50 AI researchers (ByteDance, Alibaba, Tencent + universities) just dropped a 303 page field guide on code models + coding agents. And the takeaways are not what most people assume. Here are the highlights I’m thinking about (as someone who lives in Python + agents):
Matt Dancho (Business Science) tweet media
English
25
162
918
86.2K
Kareem H. El-Safty retweetledi
Utkarsh Sharma
Utkarsh Sharma@techxutkarsh·
A senior Google engineer just dropped a 421-page doc called Agentic Design Patterns. Every chapter is code-backed and covers the frontier of AI systems: → Prompt chaining, routing, memory → MCP & multi-agent coordination → Guardrails, reasoning, planning This isn’t a blog post. It’s a curriculum. And it’s free.
Utkarsh Sharma tweet media
English
1.7K
821
4.7K
646.4K
Kareem H. El-Safty retweetledi
Donald J. Trump
Donald J. Trump@realDonaldTrump·
Now that Obama’s poll numbers are in tailspin – watch for him to launch a strike in Libya or Iran. He is desperate.
English
3K
44.3K
63.6K
0
Kareem H. El-Safty retweetledi
Sakana AI
Sakana AI@SakanaAILabs·
We’re excited to introduce Doc-to-LoRA and Text-to-LoRA, two related research exploring how to make LLM customization faster and more accessible. pub.sakana.ai/doc-to-lora/ By training a Hypernetwork to generate LoRA adapters on the fly, these methods allow models to instantly internalize new information or adapt to new tasks. Biological systems naturally rely on two key cognitive abilities: durable long-term memory to store facts, and rapid adaptation to handle new tasks given limited sensory cues. While modern LLMs are highly capable, they still lack this flexibility. Traditionally, adding long-term memory or adapting an LLM to a specific downstream task requires an expensive and time-consuming model update, such as fine-tuning or context distillation, or relies on memory-intensive long prompts. To bypass these limitations, our work focuses on the concept of cost amortization. We pay the meta-training cost once to train a hypernetwork capable of producing tasks or document specific LoRAs on demand. This turns what used to be a heavy engineering pipeline into a single, inexpensive forward pass. Instead of performing per-task optimization, the hypernetwork meta-learns update rules to instantly modify an LLM given a new task description or a long document. In our experiments, Text-to-LoRA successfully specializes models to unseen tasks using just a natural language description. Building on this, Doc-to-LoRA is able to internalize factual documents. On a needle-in-a-haystack task, Doc-to-LoRA achieves near-perfect accuracy on instances five times longer than the base model's context window. It can even generalize to transfer visual information from a vision-language model into a text-only LLM, allowing it to classify images purely through internalized weights. Importantly, both methods run with sub-second latency, enabling rapid experimentation while avoiding the overhead of traditional model updates. This approach is a step towards lowering the technical barriers of model customization, allowing end-users to specialize foundation models via simple text inputs. We have released our code and papers for the community to explore. Doc-to-LoRA Paper: arxiv.org/abs/2602.15902 Code: github.com/SakanaAI/Doc-t… Text-to-LoRA Paper: arxiv.org/abs/2506.06105 Code: github.com/SakanaAI/Text-…
GIF
English
74
349
2.2K
604.7K
Kareem H. El-Safty retweetledi
Ryan Grim
Ryan Grim@ryangrim·
Pro-Israel advocate is angry that Oman’s foreign minister went on TV to describe the unprecedented deal the Iranians are offering to avoid war. He does not want the public to know that peace was an option and Trump chose needless war.
Richard Goldberg@rich_goldberg

This is an outrageous attempt to meddle in American politics and decision-making by Oman. It might as well be the Iranian foreign minister talking. Did someone in our government even know this was going to happen? Trying to box in POTUS and sell Americans on a false narrative.

English
87
3.3K
11K
390.6K
Kareem H. El-Safty retweetledi
HatsOff
HatsOff@HatsOffff·
Oman’s foreign minister says Iran agreed to “zero enriched uranium stockpiling.” Within hours, Israel hits Tehran—and the U.S. joined. Was the deal ever real?
English
54
1.5K
5.1K
465.7K
Kareem H. El-Safty retweetledi
Cenk Uygur
Cenk Uygur@cenkuygur·
Great job by our negotiators! Iran has agreed to give up its entire enriched uranium stockpile. Now, there is absolutely no reason to go to war. If we still attack, that shows you this was always a war of choice and it was always to do regime change as Israel has demanded.
Trita Parsi@tparsi

The Omanis are famously cautious. The Omani FM going on CBS to reveal what has actually been achieved in the negotiations is quite unprecedented. And what has been achieved is significant - Trump can indeed declare victory. Listen to this segment - it goes way beyond what Obama achieved. But everything indicates that Trump won't take yes for an answer. That he will start a war of choice very soon. Which is probably why the Omani FM decided to go public. So that the American people knew that peace was within reach when Trump instead opted for war.

English
1K
1.5K
7.4K
783.4K
Kareem H. El-Safty retweetledi
Badr Albusaidi - بدر البوسعيدي
Badr Albusaidi - بدر البوسعيدي@badralbusaidi·
I met Vice President JD Vance today and shared details of the ongoing negotiation between the United States and Iran and the progress achieved so far. I am grateful for their engagement and look forward to further and decisive progress in the coming days. Peace is within our reach.
Badr Albusaidi - بدر البوسعيدي tweet media
English
8.1K
1.1K
6.3K
1.5M
Kareem H. El-Safty retweetledi
Danny Polishchuk
Danny Polishchuk@Dannyjokes·
Hey did you accidentally bomb Iraq instead of Iran??? ChatGPT: You’re absolutely right good catch!
Sam Altman@sama

Tonight, we reached an agreement with the Department of War to deploy our models in their classified network. In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome. AI safety and wide distribution of benefits are the core of our mission. Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems. The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement. We also will build technical safeguards to ensure our models behave as they should, which the DoW also wanted. We will deploy FDEs to help with our models and to ensure their safety, we will deploy on cloud networks only. We are asking the DoW to offer these same terms to all AI companies, which in our opinion we think everyone should be willing to accept. We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements. We remain committed to serve all of humanity as best we can. The world is a complicated, messy, and sometimes dangerous place.

English
286
4.6K
65.3K
3.2M
Kareem H. El-Safty retweetledi
Jens Eisert
Jens Eisert@jenseisert·
Photons are indistinguishable, and this makes certification - as it is relevant for photonic state preparations and computing - a subtle issue. We introduce methods and a mindset for certification of linear optical quantum state preparation that work in the light of this and explore them experimentally. scirate.com/arxiv/2602.122… In detail, certification is important to guarantee the correct functioning of quantum devices. A key certification task is verifying that a device has produced a desired output state. In this work, we study this task in the context of photonic platforms, where single photons are propagated through linear optical interferometers to create large, entangled resource states for metrology, communication, quantum advantage demonstrations and for so-called #linearoptical quantum computing (LOQC). This setting derives its computational power from the indistinguishability of the #photons, i.e., their relative overlap. Therefore, standard fidelity #witnesses developed for distinguishable particles (including qubits) do not apply directly, because they merely certify the closeness to some fixed target state. We introduce a measure of fidelity suitable for this setting and show several different ways to witness it, based on earlier proposals for measuring genuine multi-photon #indistinguishability. We argue that a witness based upon the discrete Fourier transform is an optimal choice. We #experimentally implement this witness and certify the fidelity of several multi-photon states. Warm thanks to the team for the wonderful collaboration, once again a @FU_Berlin-Eindhoven University of Technology one, involving Riko Schadow, Naomi Spier, Stefan N. van den Hoven, Malaquias Correa Anguita, Redlef B. G. Braamhaar, Sara Marzban, @Jelmer_Renema, and Nathan Walk.
English
4
8
42
2.3K

Keşfet

@Amal69084155 @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates @NASA @nikifrancismediavine