Matthew Yang retweeted
it's hilarious how many have taken the bitter lesson to mean just scaling up transformers. it's about scaling, but innovations in algorithmic primitives can lead to much more efficient scaling
Matthew Yang retweeted
Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHO
Matthew Yang retweeted
Never felt closer to a fundamental change in how we build retrievers, and a different kind of memory for agents.
ARI is basically test-time training for retrievers, where gradients derived from a reasoning model's preferences teach the retriever where to look *at search time*.
Lots more to do here to get oblique search queries as manageable as non-oblique ones, and that will include first figuring out sub-linear (late-interaction-esque) search.
But the path ahead has never been clearer. Read @dianetc_'s blog!
Is there a way to scale a reasoning model’s judgment for searching a massive corpus?
We introduce the Amortized Reasoning Index (ARI), a new kind of retriever that updates its weights during search, using gradients derived at inference time from the ranking preferences of a reasoning model.
Every session, current agents spend lots of reasoning deciding which search results matter, and then throw all of that judgment away upon compaction. ARI is our attempt to “amortize” that judgment, so that reasoning spent on past queries keeps improving search on future ones.
With no relevance labels, ARI takes a 400M cross-encoder that starts out weak and grows its Recall@100 on held-out queries by up to 7×, ending above Qwen3-Embedding-8B on various oblique/reasoning-intensive search tasks.
Read our blogpost: dianetc.com/musings/ari/
Preliminary results and discussions are below!
Matthew Yang retweeted
Is there a way to scale a reasoning model’s judgment for searching a massive corpus?
We introduce the Amortized Reasoning Index (ARI), a new kind of retriever that updates its weights during search, using gradients derived at inference time from the ranking preferences of a reasoning model.
Every session, current agents spend lots of reasoning deciding which search results matter, and then throw all of that judgment away upon compaction. ARI is our attempt to “amortize” that judgment, so that reasoning spent on past queries keeps improving search on future ones.
With no relevance labels, ARI takes a 400M cross-encoder that starts out weak and grows its Recall@100 on held-out queries by up to 7×, ending above Qwen3-Embedding-8B on various oblique/reasoning-intensive search tasks.
Read our blogpost: dianetc.com/musings/ari/
Preliminary results and discussions are below!
Matthew Yang retweeted
Extremely honored and humbled to be selected as a Rising Star in Data Science, a joint initiative of @UChicago, @Stanford, @Harvard, and @UCSanDiego!
I feel incredibly fortunate to learn from my current advisor at @MIT, @lateinteraction, and to be part of the wonderful OASYS lab, as well as to have been guided by my PhD advisors at @umdcs , @furongh and @dmanocha.
I'm also immensely grateful for the wonderful collaborations and discussions with some of my favorite mentors, @amritsinghbedi3, @MengdiWang10, @aldopacchiano, @anas__barakat, @hamidpalangi and so many others who have shaped my research journey.
Excited to meet, connect with, and learn from an inspiring community of researchers at the workshop!
Ref: datascience.uchicago.edu/res…
Matthew Yang retweeted
I am overwhelmed by the amount of attention this got.
Above all, I want to shine the spotlight on my brilliant co-first author @jayden_teoh_ . This would not have been possible without him.
Huge thanks also to my lab mate @a1zhang and advisor @lateinteraction for their incredible support!
Matthew Yang retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
Matthew Yang retweeted
Solving Continual Learning is widely viewed as the biggest opportunity in AI.
But are we defining it correctly?
CL is primarily concerned with one question:
How do you learn tasks B and C without forgetting task A?
Machine Studying instead asks:
How do you learn from a corpus with no task distribution at all? 🤔
The former assumes tasks are given. The latter starts before they exist. 🗺️
What should 24/7, "always on" agents actually be doing when nobody is asking them for anything? 👾♻️
I am SUPER EXCITED to publish the 148th episode of the Weaviate Podcast featuring Jacob X. Li (@jacobli99), a Ph.D. student @MIT_CSAIL! 🔥
Jacob's paper frames the answer as Machine Studying. Similarly to how a student prepares for an exam they haven't seen, an agent studies a corpus (code base, drug discovery papers, system logs, ....) without knowing the downstream task! 📚📝
Instead of leaning entirely on inference-time compute, can an agent build arbitrarily deep understanding ahead of time?
I loved the paper's definition of expertise. An expert efficiently uses inference-time compute, measured as the area under the performance-vs-compute curve of a ReAct agent loop. Studying intelligence! 📐
Jacob lays out three bets every studying method makes: synthetic data, weight updates (LoRA, KV-cache cartridges, a trained retriever), and amortized context like notes and cheat sheets. 🛠️
We also discussed why evaluation as a science matters once recursive self-improvement happens in domains with no benchmark at all. 🧪
This was a super fun conversation, I learned so much from Jacob, and I really hope you find it interesting!
YouTube: youtu.be/9iZIl0vZvvk
Spotify: spotifycreators-web.app.link…
Matthew Yang retweeted
Multi-Token Prediction (MTP) can be done much better.
🚀 Turn out,you can recover the gradient for KL self-distillation to next-token prediction (NTP) for free without ever computing the KL. You just need a one-line code change:
h_ntp += h_mtp.detach() - h_mtp
Soft Multi-Token Prediction is pretty much a free lunch for MTP.
📜 Paper: arxiv.org/abs/2610.38331
💻 Code: github.com/JaydenTeoh/Soft-M…
🤗 HuggingFace: huggingface.co/papers/2610.3…
ALT Soft Multi-Token Prediction Illustration
Matthew Yang retweeted
Machine Studying on the Weaviate Podcast 🎙️ thanks @CShorten30!
(yall can tell I had a great time there from the thumbnail!)
What should 24/7, "always on" agents actually be doing when nobody is asking them for anything? 👾♻️
I am SUPER EXCITED to publish the 148th episode of the Weaviate Podcast featuring Jacob X. Li (@jacobli99), a Ph.D. student @MIT_CSAIL! 🔥
Jacob's paper frames the answer as Machine Studying. Similarly to how a student prepares for an exam they haven't seen, an agent studies a corpus (code base, drug discovery papers, system logs, ....) without knowing the downstream task! 📚📝
Instead of leaning entirely on inference-time compute, can an agent build arbitrarily deep understanding ahead of time?
I loved the paper's definition of expertise. An expert efficiently uses inference-time compute, measured as the area under the performance-vs-compute curve of a ReAct agent loop. Studying intelligence! 📐
Jacob lays out three bets every studying method makes: synthetic data, weight updates (LoRA, KV-cache cartridges, a trained retriever), and amortized context like notes and cheat sheets. 🛠️
We also discussed why evaluation as a science matters once recursive self-improvement happens in domains with no benchmark at all. 🧪
This was a super fun conversation, I learned so much from Jacob, and I really hope you find it interesting!
YouTube: youtu.be/9iZIl0vZvvk
Spotify: spotifycreators-web.app.link…
Matthew Yang retweeted
This is Genie 4, our most realistic and interactive world model yet. 🌐
Explore dynamic environments in real time to see how the scene responds, sounds, and how the people in it move, react, and speak.
Matthew Yang retweeted
We found a way to scale humanoid control that reduces GPU-hours by more than 1000×.
Across 129,514 motions, SuperMimic surpasses the state of the art (128 GPUs, 7 days) with 1 GPU in 3 hours.
Project: supermimic.github.io
Demo: supermimic.github.io/demo/
Matthew Yang retweeted
How about having some fun with this "𝐋𝐨𝐨𝐩𝐞𝐝 𝐌𝐨𝐝𝐞𝐥 𝐀𝐭𝐥𝐚𝐬"? 📖
Fold depth into temporal recurrence, watch connections flow like a magic circle, and see how each paper handles its KV cache.
Safe travels! Don't get lost in the Sea of Recurrence 🌀
Link below 👇
Matthew Yang retweeted
We're open-sourcing RL360, the RL infrastructure behind K2 Horizon's six domain experts: 8,192 concurrent sandboxes, 150 steps in 7 days, ~1.3M rollouts.
Long-horizon agentic RL runs thousands of agents in real sandboxes on tasks that take minutes. Some of those sandboxes crash mid-task, and a crash arrives as the same low score as a wrong answer.
Without infrastructure that can keep up, GPUs are underutilized and the model gets penalized for system crashes rather than its own mistakes.
RL360 is built for this. We share how it works, what broke at scale, and what we learned in the thread.
Code: github.com/ifm-ai/RL360
Run: wandb.ai/mbzuai-llm/rl360-pu…
Matthew Yang retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
Matthew Yang retweeted
1/11 Can a single-token importance weight preserve the target of full-sequence correction? With Future-KL, we prove that both updates share the same unique stationary policy. We also contribute the promising empirical results we observed.
arxiv.org/abs/2610.18738