Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

🌀 Introducing 𝐄𝐪𝐮𝐢𝐥𝐢𝐛𝐫𝐢𝐮𝐦 𝐑𝐞𝐚𝐬𝐨𝐧𝐞𝐫𝐬 (𝐄𝐪𝐑) ! Feedforward models and weight-tied models behave very differently on hard reasoning generalization. EqR pushes this difference to the extreme by learning 𝐭𝐚𝐬𝐤-𝐜𝐨𝐧𝐝𝐢𝐭𝐢𝐨𝐧𝐞𝐝 𝐧𝐞𝐮𝐫𝐚𝐥 𝐚𝐭𝐭𝐫𝐚𝐜𝐭𝐨𝐫𝐬 . • Sudoku-Extreme: 99.8% • Maze: 93% #ICML2026

Benhao Huang ✈️ ICML

1,432 subscribers

89,808 görüntüleme • 1 ay önce •via X (Twitter)

Eğitim Bilim & Teknoloji

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

GPT-5 on Sudoku-Bench 🧩 Since releasing Sudoku-Bench in May 2025, when no LLM could solve a classic 9x9 puzzle, we've been evaluating the latest generation of models. GPT-5 now leads our leaderboard with 33% puzzles solved--approximately 2x the previous leader--and is the first LLM we've tested to solve a 9x9 Sudoku variant. However, with 67% of the much harder puzzles remaining unsolved, Sudoku-Bench continues to present significant challenges for AI reasoning. Modern Sudoku variants require models to first understand novel rulesets through meta-reasoning, then maintain global consistency across long reasoning chains. Our experiments with GRPO fine-tuning on Qwen2.5-7b and "Thought Cloning" (training on expert human reasoning from Cracking the Cryptic) show that current approaches still struggle with the spatial reasoning and creative "break-in" points that human solvers use naturally. We believe new approaches are required to solve our benchmark. These results highlight persistent gaps between computational problem-solving and human-like reasoning, particularly in tasks requiring integrated mathematical logic, spatial awareness, and creative insight. Read more about our update here: 🔗 Blogpost →

GPT-5 on Sudoku-Bench 🧩 Since releasing Sudoku-Bench in May 2025, when no LLM could solve a classic 9x9 puzzle, we've been evaluating the latest generation of models. GPT-5 now leads our leaderboard with 33% puzzles solved--approximately 2x the previous leader--and is the first LLM we've tested to solve a 9x9 Sudoku variant. However, with 67% of the much harder puzzles remaining unsolved, Sudoku-Bench continues to present significant challenges for AI reasoning. Modern Sudoku variants require models to first understand novel rulesets through meta-reasoning, then maintain global consistency across long reasoning chains. Our experiments with GRPO fine-tuning on Qwen2.5-7b and "Thought Cloning" (training on expert human reasoning from Cracking the Cryptic) show that current approaches still struggle with the spatial reasoning and creative "break-in" points that human solvers use naturally. We believe new approaches are required to solve our benchmark. These results highlight persistent gaps between computational problem-solving and human-like reasoning, particularly in tasks requiring integrated mathematical logic, spatial awareness, and creative insight. Read more about our update here: 🔗 Blogpost →

Sakana AI

154,621 görüntüleme • 7 ay önce

This is François Chollet discussing deep learning guided program synthesis, reasoning, o-series models and the ARC challenge. We will drop the full video hopefully later today!

This is François Chollet discussing deep learning guided program synthesis, reasoning, o-series models and the ARC challenge. We will drop the full video hopefully later today!

Machine Learning Street Talk

171,869 görüntüleme • 1 yıl önce

Introducing SYNTHETIC-1: Collaboratively generating the largest synthetic dataset of verified reasoning traces for math, coding and science using DeepSeek-R1. Join us to contribute compute towards state-of-the-art open reasoning models.

Introducing SYNTHETIC-1: Collaboratively generating the largest synthetic dataset of verified reasoning traces for math, coding and science using DeepSeek-R1. Join us to contribute compute towards state-of-the-art open reasoning models.

Prime Intellect

198,473 görüntüleme • 1 yıl önce

While others fight, we ship 🚀 Agno’s ReasoningTools are now live on the Playground — giving non-reasoning models the ability to "think", "analyze", and outperform Reasoning Models on Agentic workloads. If you haven't tried `ReasoningTools` yet - you definitely should 👇

While others fight, we ship 🚀 Agno’s ReasoningTools are now live on the Playground — giving non-reasoning models the ability to "think", "analyze", and outperform Reasoning Models on Agentic workloads. If you haven't tried `ReasoningTools` yet - you definitely should 👇

Ashpreet Bedi

27,864 görüntüleme • 1 yıl önce

Meet Gemini Robotics: our latest AI models designed for a new generation of helpful robots. 🤖 Based on Gemini 2.0, they bring capabilities such as better reasoning, interactivity, dexterity and generalization into the physical world. 🧵

Meet Gemini Robotics: our latest AI models designed for a new generation of helpful robots. 🤖 Based on Gemini 2.0, they bring capabilities such as better reasoning, interactivity, dexterity and generalization into the physical world. 🧵

Google DeepMind

642,242 görüntüleme • 1 yıl önce

Introducing OpenAI o3 and o4-mini—our smartest and most capable models to date. For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation.

Introducing OpenAI o3 and o4-mini—our smartest and most capable models to date. For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation.

OpenAI

3,720,457 görüntüleme • 1 yıl önce

Our paper “Learning Situated Awareness in the Real World” has been accepted to #ICML2026 as a Spotlight (top 2.2%)! Congrats to Chuhan Li ✈️ICML and the team! We introduce SAW-Bench, a real-world egocentric benchmark for evaluating situated (first-person) spatial reasoning in multimodal models, uncovering a large gap between humans and current systems. All videos recorded via Meta Ray-Ban 2 glasses!

Our paper “Learning Situated Awareness in the Real World” has been accepted to #ICML2026 as a Spotlight (top 2.2%)! Congrats to Chuhan Li ✈️ICML and the team! We introduce SAW-Bench, a real-world egocentric benchmark for evaluating situated (first-person) spatial reasoning in multimodal models, uncovering a large gap between humans and current systems. All videos recorded via Meta Ray-Ban 2 glasses!

Xin Eric Wang

11,503 görüntüleme • 2 ay önce

Today we are introducing Tara. Biological datasets are a source of insights and a means to train biological AI models. As the ability to reason at scale emerges, they take on a new role: the ground truth for testing what reasoning models produce, and the environment in which those models operate, get feedback, and improve. Tara, our autonomous research agent, is embedded in our ever-expanding datasets, lab-generated and synthetic, and built to test and evolve the hypotheses frontier models generate, matching the pace at which they produce new ideas. By keeping those models grounded in a vast space of high-precision biological data, we believe we can compound biological reasoning and close the impedance mismatch between hypothesis generation and validation.

Today we are introducing Tara. Biological datasets are a source of insights and a means to train biological AI models. As the ability to reason at scale emerges, they take on a new role: the ground truth for testing what reasoning models produce, and the environment in which those models operate, get feedback, and improve. Tara, our autonomous research agent, is embedded in our ever-expanding datasets, lab-generated and synthetic, and built to test and evolve the hypotheses frontier models generate, matching the pace at which they produce new ideas. By keeping those models grounded in a vast space of high-precision biological data, we believe we can compound biological reasoning and close the impedance mismatch between hypothesis generation and validation.

Nima Alidoust

25,625 görüntüleme • 8 gün önce

Humans learn and improve from failures. Similarly, foundation models adapt based on human feedback. Can we leverage this failure understanding to enhance robotics systems that use foundation models? Introducing AHA—a vision-language model for detecting and reasoning over failures in robotic manipulation. Project page: 🧵Thread👇 Aha!

Humans learn and improve from failures. Similarly, foundation models adapt based on human feedback. Can we leverage this failure understanding to enhance robotics systems that use foundation models? Introducing AHA—a vision-language model for detecting and reasoning over failures in robotic manipulation. Project page: 🧵Thread👇 Aha!

Jiafei Duan

48,739 görüntüleme • 1 yıl önce

AI Is Replacing Programmers and Mathematicians—Fast Eric Schmidt predicts that within one year, the majority of programmers will be replaced by AI agents, and graduate-level mathematicians will be rivaled by models trained on mathematical reasoning. Why? Because both code and math follow structured languages, which AI can learn to manipulate with extreme precision using techniques like word prediction and proof systems like Lean.

AI Is Replacing Programmers and Mathematicians—Fast Eric Schmidt predicts that within one year, the majority of programmers will be replaced by AI agents, and graduate-level mathematicians will be rivaled by models trained on mathematical reasoning. Why? Because both code and math follow structured languages, which AI can learn to manipulate with extreme precision using techniques like word prediction and proof systems like Lean.

Wes Roth

55,697 görüntüleme • 1 yıl önce

.Trenton Bricken explains how we know LLMs are actually generalizing - aka they're not just stochastic parrots: - Training models on code makes them better at reasoning in language. - Models fine tuned on math problems become better at entity detection. - We can just straightforwardly read the world-models developed by smaller NNs which are easier to interpret (Othello). Transfer learning shows models are developing a deeper understanding of their data. Full episode out Thursday.

.Trenton Bricken explains how we know LLMs are actually generalizing - aka they're not just stochastic parrots: - Training models on code makes them better at reasoning in language. - Models fine tuned on math problems become better at entity detection. - We can just straightforwardly read the world-models developed by smaller NNs which are easier to interpret (Othello). Transfer learning shows models are developing a deeper understanding of their data. Full episode out Thursday.

Dwarkesh Patel

237,715 görüntüleme • 2 yıl önce

How to harness foundation models for *generalization in the wild* in robot manipulation? Introducing VoxPoser: use LLM+VLM to label affordances and constraints directly in 3D perceptual space for zero-shot robot manipulation in the real world! 🌐 🧵👇

How to harness foundation models for generalization in the wild in robot manipulation? Introducing VoxPoser: use LLM+VLM to label affordances and constraints directly in 3D perceptual space for zero-shot robot manipulation in the real world! 🌐 🧵👇

Wenlong Huang

293,876 görüntüleme • 3 yıl önce

📈 NVIDIA tops AI leaderboards and benchmarks with open models driven by extreme co-design across compute, networking, memory, storage, and software. This includes models for biology, AI physics, agentic AI, physical AI, robotics, and autonomous vehicles. By being vertically integrated and horizontally open, NVIDIA enables its ecosystem to scale agentic workflows at every layer of the stack with the lowest cost per token.

📈 NVIDIA tops AI leaderboards and benchmarks with open models driven by extreme co-design across compute, networking, memory, storage, and software. This includes models for biology, AI physics, agentic AI, physical AI, robotics, and autonomous vehicles. By being vertically integrated and horizontally open, NVIDIA enables its ecosystem to scale agentic workflows at every layer of the stack with the lowest cost per token.

NVIDIA

85,293 görüntüleme • 2 ay önce

Today’s a big day! We have something really exciting to share with the open-source community. We’re launching two open-weight language models: gpt-oss-120b and gpt-oss-20b. They’re incredible models, built for developers, trained for reasoning, efficiency, and real-world use.🧵

Today’s a big day! We have something really exciting to share with the open-source community. We’re launching two open-weight language models: gpt-oss-120b and gpt-oss-20b. They’re incredible models, built for developers, trained for reasoning, efficiency, and real-world use.🧵

Romain Huet

34,189 görüntüleme • 11 ay önce

Learning to drive is an extreme sport

Learning to drive is an extreme sport

®️ E L E

273,159 görüntüleme • 2 ay önce

"One of the biggest misconceptions" Cerebras CFO Bob Komin pushes back on the small-models narrative. "We serve all models, and there is no limit to the size of the models that we can serve. Today, we're serving trillion parameter models. We're serving trillion parameter models that are internal for OpenAI today. We are currently running OpenAI 5.4 and 5.5 with them."

"One of the biggest misconceptions" Cerebras CFO Bob Komin pushes back on the small-models narrative. "We serve all models, and there is no limit to the size of the models that we can serve. Today, we're serving trillion parameter models. We're serving trillion parameter models that are internal for OpenAI today. We are currently running OpenAI 5.4 and 5.5 with them."

Deirdre Bosa

84,373 görüntüleme • 1 ay önce

Had enough centralization? Time to FLock OFF! Introducing FLock OFF - SN96 - Subnet (UID 96) on Bittensor - a permissionless federated learning network optimized for Small Language Models (SLMs) training on edge devices. Made by FLock.io, accelerated by Yuma

Had enough centralization? Time to FLock OFF! Introducing FLock OFF - SN96 - Subnet (UID 96) on Bittensor - a permissionless federated learning network optimized for Small Language Models (SLMs) training on edge devices. Made by FLock.io, accelerated by Yuma

FLock.io

97,851 görüntüleme • 1 yıl önce

Today we're introducing Truman. NOW LIVE. Understand the macroeconomy like never before. Not another chatbot. An AI agent to give you an edge. Built on Truflation's real-time data, proprietary forecasting models, and source-backed reasoning.

Today we're introducing Truman. NOW LIVE. Understand the macroeconomy like never before. Not another chatbot. An AI agent to give you an edge. Built on Truflation's real-time data, proprietary forecasting models, and source-backed reasoning.

Truflation

15,878 görüntüleme • 7 gün önce