Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

How does test-time scaling impact robots? We find that larger models, more thinking, and more context help significantly for some prompts but not others. Like LLMs, we can also train a router to for a better performance/latency tradeoff! Paper:

Chelsea Finn

96,660 subscribers

23,121 Aufrufe • vor 11 Tagen •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

90% of AI products fail in the first month. It's not bad models but bad prompts. Adaline AI is a platform to create prompts, test across models and modalities, integrate tools, evaluate performance, and deploy with real-time monitoring. Find the best prompt for your LLMs...👇

90% of AI products fail in the first month. It's not bad models but bad prompts. Adaline AI is a platform to create prompts, test across models and modalities, integrate tools, evaluate performance, and deploy with real-time monitoring. Find the best prompt for your LLMs...👇

Akshay 🚀

46,821 Aufrufe • vor 1 Jahr

🧵1/n LLMs significantly improve Evolutionary Algorithms for molecular discovery! For 18 different molecular optimization tasks, we demonstrate how to achieve SOTA performance by incorporating different LLMs! Learn more in our new paper! Website: Code)

🧵1/n LLMs significantly improve Evolutionary Algorithms for molecular discovery! For 18 different molecular optimization tasks, we demonstrate how to achieve SOTA performance by incorporating different LLMs! Learn more in our new paper! Website: Code)

Yuanqi Du

17,895 Aufrufe • vor 2 Jahren

Today we announce TRINDs, a dataset and benchmarking pipeline that uses synthetic personas to train and optimize performance of LLMs for tropical and infectious diseases, which are out-of-distribution for most models. Learn more →

Today we announce TRINDs, a dataset and benchmarking pipeline that uses synthetic personas to train and optimize performance of LLMs for tropical and infectious diseases, which are out-of-distribution for most models. Learn more →

Google AI

39,237 Aufrufe • vor 1 Jahr

Speech-native models like Moshi sound great and answer fast, but aren’t as smart as text LLMs. In our new paper, MoshiRAG, we show how Moshi can ask for advice from a text LLM or a knowledge base. The tricky part is how to do this in real time without adding latency. 🧵

Speech-native models like Moshi sound great and answer fast, but aren’t as smart as text LLMs. In our new paper, MoshiRAG, we show how Moshi can ask for advice from a text LLM or a knowledge base. The tricky part is how to do this in real time without adding latency. 🧵

kyutai

52,598 Aufrufe • vor 1 Monat

As we face cruelty, corruption, and chaos, may we fight. Fight hard. And fight even harder. But let us not lose ourselves. In fact, let’s find ways to show up more fully and more significantly for others.

As we face cruelty, corruption, and chaos, may we fight. Fight hard. And fight even harder. But let us not lose ourselves. In fact, let’s find ways to show up more fully and more significantly for others.

Cory Booker

48,783 Aufrufe • vor 10 Monaten

For 3D pose some use different keypoints, others SMPL and other models. It's a mess! With Neural Localizer Fields, we can choose the output at test time! allowing to train using any. Results are real time and SOTA across the board. István Sárándi

For 3D pose some use different keypoints, others SMPL and other models. It's a mess! With Neural Localizer Fields, we can choose the output at test time! allowing to train using any. Results are real time and SOTA across the board. István Sárándi

Gerard Pons-Moll

43,472 Aufrufe • vor 1 Jahr

We couldn’t help but wonder… are you ready for more, more, more? And Just Like That...

We couldn’t help but wonder… are you ready for more, more, more? And Just Like That...

Sexy Unique Podcast

15,973 Aufrufe • vor 1 Jahr

It's time the British Museum addresses slavery. We want to see a room at the British Museum for slavery, so it can tell a better story of how we got here and who we are. Find out more:

It's time the British Museum addresses slavery. We want to see a room at the British Museum for slavery, so it can tell a better story of how we got here and who we are. Find out more:

Good Law Project

70,812 Aufrufe • vor 1 Jahr

Robot AI brains, aka Vision-Language-Action models, cannot adapt to new tasks as easily as LLMs like Gemini, ChatGPT, or Grok. LLMs can adapt quickly with their in-context learning (ICL) capabilities. But can we inject ICL abilities into a pre-trained VLA like pi0? Yes! Introducing RICL (Retraining for In-Context Learning), our Conference on Robot Learning (CoRL) 2025 paper. Our RICL-pi0 model can adapt to unseen objects, novel motions, and new scenes with just ICL and RAG (retrieval-augmented generation). RICL-pi0 also boosts performance on the long-tail of tasks. A quick 1 minute video summary:

Robot AI brains, aka Vision-Language-Action models, cannot adapt to new tasks as easily as LLMs like Gemini, ChatGPT, or Grok. LLMs can adapt quickly with their in-context learning (ICL) capabilities. But can we inject ICL abilities into a pre-trained VLA like pi0? Yes! Introducing RICL (Retraining for In-Context Learning), our Conference on Robot Learning (CoRL) 2025 paper. Our RICL-pi0 model can adapt to unseen objects, novel motions, and new scenes with just ICL and RAG (retrieval-augmented generation). RICL-pi0 also boosts performance on the long-tail of tasks. A quick 1 minute video summary:

Kaustubh Sridhar

52,158 Aufrufe • vor 10 Monaten

NeurIPS 2025 Paper: LLMs are Reinforcement Learners 🤯! Surprisingly, we show that LLMs can solve RL tasks without any external component! We introduce Prompted Policy Search (ProPS), an RL method based only LLMs and in-context learning. [Paper]

NeurIPS 2025 Paper: LLMs are Reinforcement Learners 🤯! Surprisingly, we show that LLMs can solve RL tasks without any external component! We introduce Prompted Policy Search (ProPS), an RL method based only LLMs and in-context learning. [Paper]

Heni Ben Amor

51,248 Aufrufe • vor 6 Monaten

when we first started Feastables, we had a lot to learn. we went all in to fix the problems and bring you a better product that not only tastes amazing, but also does good for the world!

when we first started Feastables, we had a lot to learn. we went all in to fix the problems and bring you a better product that not only tastes amazing, but also does good for the world!

Feastables

28,857 Aufrufe • vor 11 Monaten

📁 Yann LeCun says that scaling models will not get us to human intelligence. He explains that the industry remains obsessed with making LLMs bigger, but that this path is fundamentally broken. It does not matter how many parameters we add or how many clusters we build, because the models only imitate language patterns. Human intelligence does not emerge from size, it emerges from understanding the world.

📁 Yann LeCun says that scaling models will not get us to human intelligence. He explains that the industry remains obsessed with making LLMs bigger, but that this path is fundamentally broken. It does not matter how many parameters we add or how many clusters we build, because the models only imitate language patterns. Human intelligence does not emerge from size, it emerges from understanding the world.

Jon Hernandez

156,266 Aufrufe • vor 6 Monaten

From text to reality: MIT researchers find new ways to use LLMs to help in design & manufacturing. These models can convert text prompts to CAD, generate manufacturing instructions, and search for optimal designs:

From text to reality: MIT researchers find new ways to use LLMs to help in design & manufacturing. These models can convert text prompts to CAD, generate manufacturing instructions, and search for optimal designs:

MIT CSAIL

59,152 Aufrufe • vor 2 Jahren

Every home is different. That means that to build a useful home robot, we must be able to perform zero-shot generalization on a wide range of tasks. Humanoid company 1X has a solution: world models. 1X Director of Evaluations Daniel Ho joins us on RoboPapers to talk about: - why world models are the future for scaling robot learning - how to use world models for robot control - what world models unlock for evaluating robot model performance - how we can hill-climb from here to general purpose robots Watch Episode #61 of RoboPapers, with Michael Cho - Rbt/Acc and Chris Paxton, now!

Every home is different. That means that to build a useful home robot, we must be able to perform zero-shot generalization on a wide range of tasks. Humanoid company 1X has a solution: world models. 1X Director of Evaluations Daniel Ho joins us on RoboPapers to talk about: - why world models are the future for scaling robot learning - how to use world models for robot control - what world models unlock for evaluating robot model performance - how we can hill-climb from here to general purpose robots Watch Episode #61 of RoboPapers, with Michael Cho - Rbt/Acc and Chris Paxton, now!

RoboPapers

27,567 Aufrufe • vor 4 Monaten

Question we get constantly: "How does Cline handle context limits in long-running tasks?" Here's how users can manage their context: /newtask: Creates a detailed handoff summary and starts fresh context. Like handing off work to a new engineer with full background. more 👇

Question we get constantly: "How does Cline handle context limits in long-running tasks?" Here's how users can manage their context: /newtask: Creates a detailed handoff summary and starts fresh context. Like handing off work to a new engineer with full background. more 👇

Cline

21,739 Aufrufe • vor 11 Monaten

Tonight I’m at a dinner with Mark Crossley and he has just done this quick video with me, he really is a gentleman, we all agree that homelessness should not be a thing in 2025, but the more awareness we have on the subject the more we can help. Please help me to help others by donating in any way you can ❤️❤️❤️

Tonight I’m at a dinner with Mark Crossley and he has just done this quick video with me, he really is a gentleman, we all agree that homelessness should not be a thing in 2025, but the more awareness we have on the subject the more we can help. Please help me to help others by donating in any way you can ❤️❤️❤️

Streetwise Stu

21,985 Aufrufe • vor 9 Monaten

We not only need cognitive test for President, but we also need IQ tests for Congress. How can anyone be this sTupid🤷🏻‍♀️🤦🏻‍♀️😂

We not only need cognitive test for President, but we also need IQ tests for Congress. How can anyone be this sTupid🤷🏻‍♀️🤦🏻‍♀️😂

ꪻꫝể ꪻꫝể

927,606 Aufrufe • vor 2 Jahren

btw can we also talk about how the rocket monkeys pilot looked if not better but more appealing than the actual show we got

btw can we also talk about how the rocket monkeys pilot looked if not better but more appealing than the actual show we got

guil🎙

53,595 Aufrufe • vor 6 Monaten

Our infra lets us steer trillion-parameter frontier models in real time: - live, mid-CoT edits to internal activations - directly altering how the model reasons (not just outputs) - stackable edits - no added latency We can make models more Gen Z, more concise, etc.

Our infra lets us steer trillion-parameter frontier models in real time: - live, mid-CoT edits to internal activations - directly altering how the model reasons (not just outputs) - stackable edits - no added latency We can make models more Gen Z, more concise, etc.

Goodfire

29,399 Aufrufe • vor 6 Monaten

Should we design a new programming language for LLMs, that LLMs can use more efficiently? Chris Lattner, the creator of Swift and Mojo doesn't think so, and here's why: (cont'd)

Should we design a new programming language for LLMs, that LLMs can use more efficiently? Chris Lattner, the creator of Swift and Mojo doesn't think so, and here's why: (cont'd)

Gergely Orosz

15,416 Aufrufe • vor 7 Monaten