正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

New paper on kernel generation: "AdaExplore: Failure-Driven Adaptation and Diversity-Preserving Search for Efficient Kernel Generation" Agents often waste a lot of time making the same mistakes and trying to fix unpromising solutions. AdaExplore accumulates reusable skills based on failures (Adapt) and searches with a mixture of large changes and... show more

Sean Welleck

7,660 subscribers

29,716 次观看 • 5 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

4 条评论

Sean Welleck 的头像

Sean Welleck5 个月前

Amazing work by Weihua Du (@StigLidu) and the team @JingmingZhuo @yi_xin_dong @Andre3035858461 @sunweiwei12 @regunivers Manupa Karunaratne, Ivan Fox @Tim_Dettmers @tqchenml Yiming Yang!

Kirito (e/acc) 🏴‍☠️ 的头像

Kirito (e/acc) 🏴‍☠️5 个月前

@sunweiwei12 @elliotarledge

$💥 \newline 的头像$

💥 \newline5 个月前

the adapt loop in this paper is basically what most people are missing when they try to run agents in production. if you aren't clustering failure patterns and updating the skill memory, your agent is just burning tokens to make the same logic error for the third time in ten minutes. clustering those patterns and injecting them as representative kernels is the real solve for reliability. without that feedback loop, an autonomous agent is just a very expensive way to generate stack traces.

Jakie PLA 的头像

Jakie PLA5 个月前

Love the Adapt-Explore pipeline. Teaching agents to catalog failure modes like I do with failed prints? That's exactly the iterative learning smart manufacturing needs.

相关视频

Timelapse #171 (14.5 hrs) - Contract work - continue testing D-Spark and the Laguna model drastically increased throughput of my kernel RL engine with D-Spark to hold up to high concurrencies and remove some bottlenecks with compilation time - working on kernel bench multi-GPU and kernel bench mini at the same time, where mini is used for benchmarking small local models on relatively isolated and easier problems that can be returned quickly for posts - Bunch of yapping on discord

Timelapse #171 (14.5 hrs) - Contract work - continue testing D-Spark and the Laguna model drastically increased throughput of my kernel RL engine with D-Spark to hold up to high concurrencies and remove some bottlenecks with compilation time - working on kernel bench multi-GPU and kernel bench mini at the same time, where mini is used for benchmarking small local models on relatively isolated and easier problems that can be returned quickly for posts - Bunch of yapping on discord

Elliot Arledge

27,408 次观看 • 2 个月前

The next generation of on-chain activity will be driven by autonomous AI agents on high-performance networks. Shido Network provides the infrastructure for AI agents to interact with smart contracts, move assets, and execute logic across a fast and scalable ecosystem.

The next generation of on-chain activity will be driven by autonomous AI agents on high-performance networks. Shido Network provides the infrastructure for AI agents to interact with smart contracts, move assets, and execute logic across a fast and scalable ecosystem.

Shido

151,624 次观看 • 7 个月前

Bill Gates outlines the future of immunization: a new generation of vaccines for diseases like TB, malaria, and HIV, alongside next-generation COVID vaccines. A key innovation highlighted is the shift from needle-based delivery to a small, discreet skin patch.

Bill Gates outlines the future of immunization: a new generation of vaccines for diseases like TB, malaria, and HIV, alongside next-generation COVID vaccines. A key innovation highlighted is the shift from needle-based delivery to a small, discreet skin patch.

Camus

1,057,138 次观看 • 1 年前

Make Pixels Dance: High-Dynamic Video Generation paper page: Creating high-dynamic videos such as motion-rich actions and sophisticated visual effects poses a significant challenge in the field of artificial intelligence. Unfortunately, current state-of-the-art video generation methods, primarily focusing on text-to-video generation, tend to produce video clips with minimal motions despite maintaining high fidelity. We argue that relying solely on text instructions is insufficient and suboptimal for video generation. In this paper, we introduce PixelDance, a novel approach based on diffusion models that incorporates image instructions for both the first and last frames in conjunction with text instructions for video generation. Comprehensive experimental results demonstrate that PixelDance trained with public data exhibits significantly better proficiency in synthesizing videos with complex scenes and intricate motions, setting a new standard for video generation.

Make Pixels Dance: High-Dynamic Video Generation paper page: Creating high-dynamic videos such as motion-rich actions and sophisticated visual effects poses a significant challenge in the field of artificial intelligence. Unfortunately, current state-of-the-art video generation methods, primarily focusing on text-to-video generation, tend to produce video clips with minimal motions despite maintaining high fidelity. We argue that relying solely on text instructions is insufficient and suboptimal for video generation. In this paper, we introduce PixelDance, a novel approach based on diffusion models that incorporates image instructions for both the first and last frames in conjunction with text instructions for video generation. Comprehensive experimental results demonstrate that PixelDance trained with public data exhibits significantly better proficiency in synthesizing videos with complex scenes and intricate motions, setting a new standard for video generation.

AK

102,782 次观看 • 2 年前

🚀 Introducing Emu3.5 — a large-scale multimodal world model that natively predicts the next vision-language state. 🔥 Trained on over 10T interleaved vision-language tokens and enhanced with reinforcement learning, Emu3.5 achieves powerful multimodal reasoning and generation. ⚡ Powered by our new Discrete Diffusion Adaptation (DiDA) for 20× faster inference. 🔥 Emu3.5 outperforms Nano Banana across image generation, editing, interleaved tasks and more. 🌍 Explore Emu3.5: Github: #Emu3 #MultimodalAI #WorldModel #NextTokenPrediction

🚀 Introducing Emu3.5 — a large-scale multimodal world model that natively predicts the next vision-language state. 🔥 Trained on over 10T interleaved vision-language tokens and enhanced with reinforcement learning, Emu3.5 achieves powerful multimodal reasoning and generation. ⚡ Powered by our new Discrete Diffusion Adaptation (DiDA) for 20× faster inference. 🔥 Emu3.5 outperforms Nano Banana across image generation, editing, interleaved tasks and more. 🌍 Explore Emu3.5: Github: #Emu3 #MultimodalAI #WorldModel #NextTokenPrediction

BAAI

51,880 次观看 • 11 个月前

Saturday Night logs working remotely for a US based agency - > Was working on a kernel-adapter base level architecture to map any agent at will for the multi-agentic system. > The above architecture was inspired from the HEXAGONAL architecture ( the ports and adapters architecture ). > Wired in some new agent swarms > Did some UI refinements > Worked on some memory retrieval bugs PS - New steps include converting the multi agentic setup into bots across discord, slack, telegram and much more. Lots of work to do, we go on a small break on Sunday !!

Saturday Night logs working remotely for a US based agency - > Was working on a kernel-adapter base level architecture to map any agent at will for the multi-agentic system. > The above architecture was inspired from the HEXAGONAL architecture ( the ports and adapters architecture ). > Wired in some new agent swarms > Did some UI refinements > Worked on some memory retrieval bugs PS - New steps include converting the multi agentic setup into bots across discord, slack, telegram and much more. Lots of work to do, we go on a small break on Sunday !!

Siddharth

52,731 次观看 • 3 个月前

Southwest Washington growers should be able to spend less time doing paperwork and more time on their tractors. It's already hard enough to keep a small, family farm afloat. My bills will take on foreign ownership of our farmland, secure loans for a new generation of farmers, and protect our freedom to fix ag equipment where, when, and how we want.

Southwest Washington growers should be able to spend less time doing paperwork and more time on their tractors. It's already hard enough to keep a small, family farm afloat. My bills will take on foreign ownership of our farmland, secure loans for a new generation of farmers, and protect our freedom to fix ag equipment where, when, and how we want.

Rep. Marie Gluesenkamp Perez

14,947 次观看 • 1 年前

New release from Meta FAIR — Meta Motivo is a first-of-its-kind behavioral foundation model for controlling virtual physics-based humanoid agents for a wide range of complex whole-body tasks. The model is capable of expressing human-like behaviors and achieves performance competitive with task-specific methods and outperforms state-of-the-art unsupervised RL and model-based baselines. Try the demo ➡️ Get the model and code ➡️ We’re excited about how this research could pave the way for fully embodied agents, leading to more lifelike NPCs, democratization of character animation and new types of immersive experiences.

New release from Meta FAIR — Meta Motivo is a first-of-its-kind behavioral foundation model for controlling virtual physics-based humanoid agents for a wide range of complex whole-body tasks. The model is capable of expressing human-like behaviors and achieves performance competitive with task-specific methods and outperforms state-of-the-art unsupervised RL and model-based baselines. Try the demo ➡️ Get the model and code ➡️ We’re excited about how this research could pave the way for fully embodied agents, leading to more lifelike NPCs, democratization of character animation and new types of immersive experiences.

AI at Meta

129,283 次观看 • 1 年前

Meta announces Movie Gen A Cast of Media Foundation Models We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user’s image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models

Meta announces Movie Gen A Cast of Media Foundation Models We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user’s image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models

AK

62,719 次观看 • 2 年前

Presto! Distilling Steps and Layers for Accelerating Music Generation Despite advances in diffusion-based text-to-music (TTM) methods, efficient, high-quality generation remains a challenge. We introduce Presto!, an approach to inference acceleration for score-based diffusion transformers via reducing both sampling steps and cost per step. To reduce steps, we develop a new score-based distribution matching distillation (DMD) method for the EDM-family of diffusion models, the first GAN-based distillation method for TTM. To reduce the cost per step, we develop a simple, but powerful improvement to a recent layer distillation method that improves learning via better preserving hidden state variance. Finally, we combine our step and layer distillation methods together for a dual-faceted approach. We evaluate our step and layer distillation methods independently and show each yield best-in-class performance. Our combined distillation method can generate high-quality outputs with improved diversity, accelerating our base model by 10-18x (230/435ms latency for 32 second mono/stereo 44.1kHz, 15x faster than comparable SOTA) -- the fastest high-quality TTM to our knowledge.

Presto! Distilling Steps and Layers for Accelerating Music Generation Despite advances in diffusion-based text-to-music (TTM) methods, efficient, high-quality generation remains a challenge. We introduce Presto!, an approach to inference acceleration for score-based diffusion transformers via reducing both sampling steps and cost per step. To reduce steps, we develop a new score-based distribution matching distillation (DMD) method for the EDM-family of diffusion models, the first GAN-based distillation method for TTM. To reduce the cost per step, we develop a simple, but powerful improvement to a recent layer distillation method that improves learning via better preserving hidden state variance. Finally, we combine our step and layer distillation methods together for a dual-faceted approach. We evaluate our step and layer distillation methods independently and show each yield best-in-class performance. Our combined distillation method can generate high-quality outputs with improved diversity, accelerating our base model by 10-18x (230/435ms latency for 32 second mono/stereo 44.1kHz, 15x faster than comparable SOTA) -- the fastest high-quality TTM to our knowledge.

AK

30,446 次观看 • 2 年前

We have an opportunity to turn the page on a decade of Donald Trump trying to keep us divided and afraid of each other. America, that is not who we are. It is time for a new generation of leadership.

We have an opportunity to turn the page on a decade of Donald Trump trying to keep us divided and afraid of each other. America, that is not who we are. It is time for a new generation of leadership.

Kamala Harris

351,988 次观看 • 1 年前

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.

AK

125,353 次观看 • 3 年前

Test-time scaling, reasoning, and generally search-like processes clearly drive significant gains in LLMs. Largely owed to the structure of language. One would think the same could apply to non-linguistic domains, like image generation, but that obviously depends on whether the structure of the domain's representation lends itself to search. 1D ordered tokens (e.g., image FlexTok, video FlexTok) seem like a natural fit since they enable a step-by-step coarse-to-fine generation. We investigated that and found they indeed enable search and scale far better with test-time compute than 2D grids. See the visuals on the webpage. Appearing in ICML Conference 2026. 🔗 📄

Test-time scaling, reasoning, and generally search-like processes clearly drive significant gains in LLMs. Largely owed to the structure of language. One would think the same could apply to non-linguistic domains, like image generation, but that obviously depends on whether the structure of the domain's representation lends itself to search. 1D ordered tokens (e.g., image FlexTok, video FlexTok) seem like a natural fit since they enable a step-by-step coarse-to-fine generation. We investigated that and found they indeed enable search and scale far better with test-time compute than 2D grids. See the visuals on the webpage. Appearing in ICML Conference 2026. 🔗 📄

Amir Zamir

15,735 次观看 • 4 个月前

When data agents fail, they often fail silently - giving confident-sounding answers that are wrong, and it can be hard to figure out what caused the failure. "Building and Evaluating Data Agents" is a new short course created with Snowflake and taught by Anupam Datta and Josh Reini that teaches you to build data agents with comprehensive evaluation built in. Skills you'll gain: - Build reliable LLM data agents using the Goal-Plan-Action framework and runtime evaluations that catch failures mid-execution - Use OpenTelemetry tracing and evaluation infrastructure to diagnose exactly where agents fail and systematically improve performance - Orchestrate multi-step workflows across web search, SQL, and document retrieval in LangGraph-based agents The result: visibility into every step of your agent's reasoning, so if something breaks, you have a systematic approach to fix it. Sign up to get started:

When data agents fail, they often fail silently - giving confident-sounding answers that are wrong, and it can be hard to figure out what caused the failure. "Building and Evaluating Data Agents" is a new short course created with Snowflake and taught by Anupam Datta and Josh Reini that teaches you to build data agents with comprehensive evaluation built in. Skills you'll gain: - Build reliable LLM data agents using the Goal-Plan-Action framework and runtime evaluations that catch failures mid-execution - Use OpenTelemetry tracing and evaluation infrastructure to diagnose exactly where agents fail and systematically improve performance - Orchestrate multi-step workflows across web search, SQL, and document retrieval in LangGraph-based agents The result: visibility into every step of your agent's reasoning, so if something breaks, you have a systematic approach to fix it. Sign up to get started:

Andrew Ng

102,472 次观看 • 1 年前

Charlie McAvoy on building up a new culture with the Bruins: "It is a new day and age. We're trying to usher in a new group of guys that are from a different generation."

Charlie McAvoy on building up a new culture with the Bruins: "It is a new day and age. We're trying to usher in a new group of guys that are from a different generation."

Conor Ryan

11,109 次观看 • 1 年前

Doug, a small fan with big dreams, and his best friend Scout explore the biggest show in sports through relatable stories that celebrate teams, spotlight superstars, and unlock the language of the game for a new generation of fans 👏 New episodes and content of “The Doug Out!” will roll out regularly on YouTube beginning April 7 with an episode highlighting the Seattle Mariners and their star catcher Cal Raleigh!

Doug, a small fan with big dreams, and his best friend Scout explore the biggest show in sports through relatable stories that celebrate teams, spotlight superstars, and unlock the language of the game for a new generation of fans 👏 New episodes and content of “The Doug Out!” will roll out regularly on YouTube beginning April 7 with an episode highlighting the Seattle Mariners and their star catcher Cal Raleigh!

MLB

101,491 次观看 • 6 个月前

To Adapt or Not to Adapt? Real-Time Adaptation for Semantic Segmentation paper page: The goal of Online Domain Adaptation for semantic segmentation is to handle unforeseeable domain changes that occur during deployment, like sudden weather events. However, the high computational costs associated with brute-force adaptation make this paradigm unfeasible for real-world applications. In this paper we propose HAMLET, a Hardware-Aware Modular Least Expensive Training framework for real-time domain adaptation. Our approach includes a hardware-aware back-propagation orchestration agent (HAMT) and a dedicated domain-shift detector that enables active control over when and how the model is adapted (LT). Thanks to these advancements, our approach is capable of performing semantic segmentation while simultaneously adapting at more than 29FPS on a single consumer-grade GPU. Our framework's encouraging accuracy and speed trade-off is demonstrated on OnDA and SHIFT benchmarks through experimental results.

To Adapt or Not to Adapt? Real-Time Adaptation for Semantic Segmentation paper page: The goal of Online Domain Adaptation for semantic segmentation is to handle unforeseeable domain changes that occur during deployment, like sudden weather events. However, the high computational costs associated with brute-force adaptation make this paradigm unfeasible for real-world applications. In this paper we propose HAMLET, a Hardware-Aware Modular Least Expensive Training framework for real-time domain adaptation. Our approach includes a hardware-aware back-propagation orchestration agent (HAMT) and a dedicated domain-shift detector that enables active control over when and how the model is adapted (LT). Thanks to these advancements, our approach is capable of performing semantic segmentation while simultaneously adapting at more than 29FPS on a single consumer-grade GPU. Our framework's encouraging accuracy and speed trade-off is demonstrated on OnDA and SHIFT benchmarks through experimental results.

AK

18,871 次观看 • 3 年前

Loopy Taming Audio-Driven Portrait Avatar with Long-Term Motion Dependency paper page: With the introduction of diffusion-based video generation techniques, audio-conditioned human video generation has recently achieved significant breakthroughs in both the naturalness of motion and the synthesis of portrait details. Due to the limited control of audio signals in driving human motion, existing methods often add auxiliary spatial signals to stabilize movements, which may compromise the naturalness and freedom of motion. In this paper, we propose an end-to-end audio-only conditioned video diffusion model named Loopy. Specifically, we designed an inter- and intra-clip temporal module and an audio-to-latents module, enabling the model to leverage long-term motion information from the data to learn natural motion patterns and improving audio-portrait movement correlation. This method removes the need for manually specified spatial motion templates used in existing methods to constrain motion during inference. Extensive experiments show that Loopy outperforms recent audio-driven portrait diffusion models, delivering more lifelike and high-quality results across various scenarios.

AK

128,912 次观看 • 2 年前