Lightnews — Scholar-powered news

Reposted by Peter Gray

9to5Mac @9to5mac.com · 14d

Apple researchers develop SimpleFold, a lightweight AI for protein folding prediction

Google DeepMind’s work with AlphaFold has been nothing short of a miracle, but it is computationally expensive. With that in mind, Apple researchers set off to develop an alternative method to use AI to predict the 3D structure of proteins, and it shows promise. Here are the details. more…

9to5mac.com

4 20

Peter Gray @peteryugray.bsky.social · Aug 21

tl;dr: some parameters are much more important than others, and in some cases removing just 1 can turn an LLM's output to nonsense

Peter Gray @peteryugray.bsky.social · Aug 21

New Apple #ML Research Highlight: The "Super Weight:" How Even a Single Parameter can Determine an #LLM's Behavior machinelearning.apple.com/research/the...

The

A recent paper from Apple researchers,

machinelearning.apple.com

1 2 3

Peter Gray @peteryugray.bsky.social · Jul 23

The inference code, model checkpoints, and an iOS/macOS demo app based on MLX are available here: github.com/apple/ml-fas...

GitHub - apple/ml-fastvlm: This repository contains the official implementation of "FastVLM: Efficient Vision Encoding for Vision Language Models" - CVPR 2025

This repository contains the official implementation of "FastVLM: Efficient Vision Encoding for Vision Language Models" - CVPR 2025 - apple/ml-fastvlm

github.com

2

Peter Gray @peteryugray.bsky.social · Jul 23

How fast is it? Here's the demo app running FastVLM 0.5B model on iPhone 16 Pro. Time to first token is shown on the screen, highlighting near real-time performance.

1 1

Peter Gray @peteryugray.bsky.social · Jul 23

New Apple #ML Research Highlight: "FastVLM: Efficient Vision Encoding for Vision Language Models" machinelearning.apple.com/research/fas...

FastVLM: Efficient Vision Encoding for Vision Language Models

Vision Language Models (VLMs) enable visual understanding alongside textual inputs. They are typically built by passing visual tokens from a…

machinelearning.apple.com

1 1 3

Peter Gray @peteryugray.bsky.social · Jul 17

New paper: 'Apple Intelligence Foundation Language Models Tech Report 2025' provides technical details for two multilingual, multimodal foundation language models that power Apple Intelligence features across Apple devices and services:
machinelearning.apple.com/research/app...

Apple Intelligence Foundation Language Models Tech Report 2025

We introduce two multilingual, multimodal foundation language models that power Apple Intelligence features across Apple devices and…

machinelearning.apple.com

1

Peter Gray @peteryugray.bsky.social · Jul 11

And for a comprehensive overview of Apple research at the conference - including the complete schedule of orals, posters, workshops, booth programming and more - see this post: machinelearning.apple.com/updates/appl...

This link will take you to a page that’s not on LinkedIn

lnkd.in

Peter Gray @peteryugray.bsky.social · Jul 11

Next week at #ICML2025, Apple researchers will present many new papers across a range of topics in #AI & #ML - check out this post for some highlights: machinelearning.apple.com/research/icm...

1

Peter Gray @peteryugray.bsky.social · Jun 9

New post: "Updates to Apple's On-Device and Server Foundation Language Models" - details the architectures, training data and recipes, inference optimization techniques, and evaluation results compared to comparable models: machinelearning.apple.com/research/app...

Updates to Apple's On-Device and Server Foundation Language Models

With Apple Intelligence, we're integrating powerful generative AI right into the apps and experiences people use every day, all while…

machinelearning.apple.com

1

Peter Gray @peteryugray.bsky.social · Apr 24

New Apple #ML Research Highlight: "An LLM-Based Approach to Review Summarization on the App Store" - details the multi-step #LLM workflow for generating high quality summaries from diverse crowdsourced reviews in a dynamic environment: machinelearning.apple.com/research/app...

An LLM-Based Approach to Review Summarization on the App Store

Ratings and reviews are an invaluable resource for users exploring an app on the App Store, providing insights into how others have…

machinelearning.apple.com

Peter Gray @peteryugray.bsky.social · Apr 21

New post: "Apple Machine Learning Research at #ICLR 2025" - highlighting a selection of the many Apple #ML research papers to be presented at @iclr-conf.bsky.social this week: machinelearning.apple.com/research/icl...

Apple Machine Learning Research at ICLR 2025

Apple researchers are advancing machine learning (ML) and AI through fundamental research that improves the world’s understanding of this…

machinelearning.apple.com

3 7

Peter Gray @peteryugray.bsky.social · Apr 10

Accepted as a Spotlight at @iclr-conf.bsky.social the work shares a new method for fine-grained control over #genAI output - without the computational overhead, complexity, and volume of data needed by #RLHF or fine-tuning, and with more reliable results than prompt engineering.

1 3

Peter Gray @peteryugray.bsky.social · Apr 10

New Apple #ML Research Highlight: "Controlling Language and Diffusion Models by Transporting Activations” machinelearning.apple.com/research/tra...

Controlling Language and Diffusion Models by Transporting Activations

Large generative models are becoming increasingly capable and more widely deployed to power production applications, but getting these…

machinelearning.apple.com

1 1

Reposted by Peter Gray

WIRED @wired.com · Mar 27

Modern science wouldn’t exist without the online research repository known as arXiv. Three decades in, its creator still can’t let it go.

Inside arXiv—the Most Transformative Platform in All of Science

Modern science wouldn’t exist without the online research repository known as arXiv. Three decades in, its creator still can’t let it go.

wrd.cm

9 210 740

Peter Gray @peteryugray.bsky.social · Jan 23

Very cool work from an international team of researchers including @www.helmholtz-munich.de and Apple: applying optimal transport to give scientists a new ability to observe millions of cells simultaneously, as they develop across time and in space

Nature @nature.com · Jan 23

Nature research paper: Mapping cells through time and space with moscot

https://go.nature.com/40pqpvw

Mapping cells through time and space with moscot - Nature

Moscot is an optimal transport approach that overcomes current limitations of similar methods to enable multimodal, scalable and consistent single-cell analyses of datasets across spatial and temporal dimensions.

go.nature.com

2

Reposted by Peter Gray

Marco Cuturi @marcocuturi.bsky.social · Jan 22

Today is a great day for optimal transport 🎉! Lots of gratitude 🙏 for all folks who contributed to ott-jax.readthedocs.io and pushed for the MOSCOT (now @ nature!) paper, from visionaries @dominik1klein.bsky.social, G. Palla, Z. Piran to the magician, Michal Klein! ❤️

www.nature.com/articles/s41...

7 22

Peter Gray @peteryugray.bsky.social · Jan 16

Congratulations!

1

Peter Gray @peteryugray.bsky.social · Dec 29

Aging hipster targeted marketing

1

Peter Gray @peteryugray.bsky.social · Dec 22

Supermarket bot on patrol in suburban NJ

A tall robot with googly eyes in the produce section of a supermarket

2

Peter Gray @peteryugray.bsky.social · Dec 18

Devs can now benefit from faster inference for their production LLMs on NVIDIA GPUs - benchmarking shows 2.7x acceleration in token generation 5/5

Peter Gray @peteryugray.bsky.social · Dec 18

o make this advancement production-ready for NVIDIA GPUs, the team collaborated with NVIDIA to integrate ReDrafter into the NVIDIA TensorRT-LLM framework: developer.nvidia.com/blog/nvidia-... 4/5

NVIDIA TensorRT-LLM Now Supports Recurrent Drafting for Optimizing LLM Inference | NVIDIA Technical Blog

Recurrent drafting (referred as ReDrafter) is a novel speculative decoding technique developed and open-sourced by Apple for large language model (LLM) inference now available with NVIDIA TensorRT-LLM...

developer.nvidia.com

1

Peter Gray @peteryugray.bsky.social · Dec 18

Earlier this year, Apple Machine Learning researchers published & open sourced ReDrafter, a novel approach to speculative decoding: machinelearning.apple.com/research/rec... 2/5

1

Peter Gray @peteryugray.bsky.social · Dec 18

New #Apple Machine Learning research highlight: “Accelerating LLM Inference on #NVIDIA GPUs with ReDrafter” - a story of impactful, production-relevant #ML research: machinelearning.apple.com/research/red... 1/5

Accelerating LLM Inference on NVIDIA GPUs with ReDrafter

Accelerating LLM inference is an important ML research problem, as auto-regressive token generation is computationally expensive and…

machinelearning.apple.com

1