Ali H

Past: Signals, Circuits, Microwaves

Now: Model , Data, Intelligent solvers

[Ali H @007Alihaider]

@hsntalal_tiwana

[Jia-Bin Huang @jbhuang0604]

Wow! Such a cool paper on fixing softmax and cross-entropy!

I enjoyed reading this paper today, so I made an explainer to make the ideas more accessible to a broader audience.

Check it out: https://t.co/cQXPuQ2bd3

[Jon Barron @jon_barron]

Hi all, I have a new result I'm excited to share. It turns out that *all existing implementations* of softmax and cross-entropy (that's JAX, PyTorch, textbooks, etc) have a subtle bug --- two bugs, actually. A simple rewrite fixes both, and eliminates "slingshots" when grokking:

[Carles Domingo-Enrich @cdomingoenrich]

We're hiring research interns in machine learning, generative modeling, and their applications!

Our research scope is broad, so if you're interested, feel free to reach out and tell us a bit about your research interests!

Apply here: https://t.co/rhDPO8ezE6

[Oğuzhan Fatih Kar @oguzhanthefatih]

 🇨🇭 We're hiring a research intern at Apple Zurich! Come work on multimodal pretraining and scaling laws with us (9+ months, could start asap).

📩 Send your CV and representative work to [email protected]

[Arash Vahdat @ArashVahdat]

Please help spread the word!

We’re hiring Research Scientists and Research Interns at NVIDIA Research to work on some of the most exciting problems in generative AI:

🌎 World Models & Video Diffusion
🧠 Diffusion Language Models
🧬 Biomolecule Design

If your research aligns with any of these areas, please apply directly through the links below.

Open positions + application links:
https://t.co/MUAt4mcewx

[Ali H @007Alihaider]

@thsottiaux Intelligence

[Akshay 🚀 @akshay_pachaar]

https://t.co/haL8IGhx3h

[Lydia Hallie ✨ @lydiahallie]

One of our most requested updates! The team has been working on it for a while, much nicer experience when you hit your limit mid-task 👏

[Ali H @007Alihaider]

@angipathy Nah,
Doesn't seems like AI

[XY Han @XYHan_]

My PhD advisor read everything I wrote line-by-line. Re-derived all my proofs himself on principle.

I recheck and rewrite everything AI hands to me: regardless of how correct-looking or lean-certified it is. I can answer for anything with my name on it. On the same principle.

[Ali H @007Alihaider]

@thsottiaux Main*

[Ali H @007Alihaider]

@thsottiaux Well,
Sol for me is unpredictable due to frequent side stepping from the mail goal

[Asad Memon @_asadmemon]

Mighty can now build 3D occupancy maps from a single camera! 😁

Mighty runs VIO realtime on-device, giving us accurate camera motion. Using that + the camera feed, the SDK estimates depth and builds a 3D map of obstacles around it.

This means a robot or drone can use Mighty for things like:

- Collision avoidance
- Motion planning
- Autonomous navigation

No stereo camera or depth sensor needed. Just Mighty’s global shutter camera + IMU.

[Peyman Milanfar @docmilanfar]

“Thrilled to announce 14 papers accepted from our lab”

The lab:

[Peyman Milanfar @docmilanfar]

Classifier-Free Guidance

[Ali H @007Alihaider]

@thsottiaux Well I would agree with that

[Ali H @007Alihaider]

@claudeai Thanks GPT 5.6

[zed @zmkzmkz]

dude this ICML tutorial by Prof Mark Schmidt on optimization was so good I’m gonna download the presentation and rewatch it like a billion times. completely reworked my intuition on optimizers and neural net training in general (not that my intuition was good in the first place)

[Victoria X Lin @VictoriaLinML]

The video of my Stanford CS25 guest lecture, From Language Models to Native Multimodal Intelligence, is now online.

I discussed how the core ideas behind LLMs has shaped multimodal AI, from architectures to training paradigms and scaling, and where the next challenges may lie. 🧠🌐
🎥: https://t.co/sUVxSEjrVf