Загрузка видео...

Не удалось загрузить видео

На главную

✨ICLR 2024 Spotlight Still hand-labeling phases or tuning GANs with your trajectories? We propose FLD, a self-supervised representation method that extracts spatial-temporal relationships in high-dimensional trajectories. RL policies informed by FLD show extended generality.🧵

breadli428's profile picture

Chenhao Li

7,620 subscribers

81,979 просмотров • 2 лет назад •via X (Twitter)

ИскусствоНаука и технологииОбразование
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 11

Фото профиля Chenhao Li
Chenhao Li2 лет назад

The availability of reference trajectories has significantly propelled the advancement of motion learning techniques. How do we efficiently represent these high-dimensional, long-horizon, highly nonlinear signals such that downstream learning tasks can generalize better?

Фото профиля Chenhao Li
Chenhao Li2 лет назад

Fourier Latent Dynamics (FLD) achieves so by featuring dynamics in a continuously parameterized latent space. The latent encodings consist of a varying component (phi) that indicates local time indices and an invariant component (theta) that represents global motion features.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

Each motion (or any other quasi-periodic signal) induces a ring in the latent manifold, where the radius denotes the invariant high-level features, and the angle denotes the current frame of that motion.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

This way, with a few parameters, FLD is able to represent diverse, theoretically infinitely long motion trajectories or any other quasi-periodic signal. In the table, d denotes the number of state dimensions, tau trajectory length, H window size, and c number of latent channels.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

The open-loop reconstruction of FLD by consecutively propagating the latent dynamics and decoding to the state space demonstrates high fidelity on test data unseen during training, compared with less structured models, including FF networks and Periodic Autoencoders (PAE).

Фото профиля Chenhao Li
Chenhao Li2 лет назад

With such an efficient latent representation, we train a downstream tracking policy with RL. A motion is sampled at each episode. Its latent phase propagates through time while the high-level features stay constant. The tracking reward is computed each step with reconstruction.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

This structure enables real-time motion tracking. During inference, FLD encodes real-time motion input as tracking targets, irrespective of their periodic or quasi-periodic nature.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

FLD provides a fallback mechanism with anomaly detection from its autoencoder nature. When a potentially risky target is proposed, FLD rejects the input and resorts to a safe motion by propagating the latent dynamics with the previous parameters.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

We provide an ablation video illustrating the importance of the fallback mechanism as a safety filter in online motion tracking.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

This work has been published at ICLR 2024 @iclr_conf by the Biomimetic Robotics Lab @MITMechE @MIT. It will be presented as a spotlight paper in May in Vienna, Austria.

Фото профиля Chenhao Li
Chenhao Li2 лет назад

Authors Chenhao Li @breadli428, Elijah Stanger-Jones, Steve Heim, Sangbae Kim Paper Project Code Videos @MITMechE @MITEECS @MIT_CSAIL @MIT

Похожие видео

Achieve SOTA on multimodal, complex tasks with minimal supervision! Introducing In-Context Abstraction Learning (ICAL): a method that extracts insights from visual demos and human feedback to continually teach multimodal LLM agents 🚀🤖🧠 ICAL learns from sub-optimal demonstrations and human feedback by transforming noisy trajectories into optimized, annotated examples that improve multimodal decision-making in new domains. ICAL is the first to show state-of-the-art performance across three challenging domains: we obtain a 34% relative improvement on TEACh household robot instruction following from dialogue, a 59% relative improvement on Visual Web Arena for autonomous visual web agents, and beat few-shot GPT-4V and supervised baselines on Ego4D action forecasting.
0:30

Achieve SOTA on multimodal, complex tasks with minimal supervision! Introducing In-Context Abstraction Learning (ICAL): a method that extracts insights from visual demos and human feedback to continually teach multimodal LLM agents 🚀🤖🧠 ICAL learns from sub-optimal demonstrations and human feedback by transforming noisy trajectories into optimized, annotated examples that improve multimodal decision-making in new domains. ICAL is the first to show state-of-the-art performance across three challenging domains: we obtain a 34% relative improvement on TEACh household robot instruction following from dialogue, a 59% relative improvement on Visual Web Arena for autonomous visual web agents, and beat few-shot GPT-4V and supervised baselines on Ego4D action forecasting.

Gabriel Sarch

19,340 просмотров • 2 лет назад

Introducing Tether 🪢, a fun little idea to scale data by having our robot “play” in the real world for over 24 hours, throughout the day and overnight—improving policies from zero to mastery with minimal supervision! But play is messy, with out-of-distribution scenarios that are hard to anticipate. To perform autonomous functional play in the real world, from just a handful of demos, we propose a highly robust few-shot imitation method that warps demo trajectories using visual correspondences. Then, continuously running it within a multi-task VLM-guided cycle, we generate a data stream that produces 1000+ expert-level demos. This generated data is finally funneled downstream to train imitation learning policies, which improve from zero to near-perfect success rates. We’ll be presenting Tether at #ICLR2026 in just a few weeks! But before that, deep dive with me… 🧵
0:30

Introducing Tether 🪢, a fun little idea to scale data by having our robot “play” in the real world for over 24 hours, throughout the day and overnight—improving policies from zero to mastery with minimal supervision! But play is messy, with out-of-distribution scenarios that are hard to anticipate. To perform autonomous functional play in the real world, from just a handful of demos, we propose a highly robust few-shot imitation method that warps demo trajectories using visual correspondences. Then, continuously running it within a multi-task VLM-guided cycle, we generate a data stream that produces 1000+ expert-level demos. This generated data is finally funneled downstream to train imitation learning policies, which improve from zero to near-perfect success rates. We’ll be presenting Tether at #ICLR2026 in just a few weeks! But before that, deep dive with me… 🧵

Will Liang

46,341 просмотров • 6 месяцев назад

Tracking Everything Everywhere All at Once paper page: present a new test-time optimization method for estimating dense and long-range motion from a video sequence. Prior optical flow or particle video tracking algorithms typically operate within limited temporal windows, struggling to track through occlusions and maintain global consistency of estimated motion trajectories. We propose a complete and globally consistent motion representation, dubbed OmniMotion, that allows for accurate, full-length motion estimation of every pixel in a video. OmniMotion represents a video using a quasi-3D canonical volume and performs pixel-wise tracking via bijections between local and canonical space. This representation allows us to ensure global consistency, track through occlusions, and model any combination of camera and object motion. Extensive evaluations on the TAP-Vid benchmark and real-world footage show that our approach outperforms prior state-of-the-art methods by a large margin both quantitatively and qualitatively.
5:18

Tracking Everything Everywhere All at Once paper page: present a new test-time optimization method for estimating dense and long-range motion from a video sequence. Prior optical flow or particle video tracking algorithms typically operate within limited temporal windows, struggling to track through occlusions and maintain global consistency of estimated motion trajectories. We propose a complete and globally consistent motion representation, dubbed OmniMotion, that allows for accurate, full-length motion estimation of every pixel in a video. OmniMotion represents a video using a quasi-3D canonical volume and performs pixel-wise tracking via bijections between local and canonical space. This representation allows us to ensure global consistency, track through occlusions, and model any combination of camera and object motion. Extensive evaluations on the TAP-Vid benchmark and real-world footage show that our approach outperforms prior state-of-the-art methods by a large margin both quantitatively and qualitatively.

AK

280,570 просмотров • 3 лет назад

Introducing FoundationMotion. A large-scale, video-derived motion annotation dataset & auto-labeling pipeline + advanced models for motion understanding. Fully open-source: code, datasets, and models, free to use and build on. Understanding motion is core to physical reasoning, yet today’s leading models still struggle with simple spatial actions like “turn right” or “move up” or “flip the toast” - mainly due to the lack of large, fine-grained motion datasets. We present FoundationMotion, a fully automated pipeline that: • detects & tracks objects in videos • extracts trajectories • uses LLMs + frames to generate rich motion captions & QA pairs → creating large-scale, high-quality motion datasets at scale. After fine-tuning the open-source models Qwen and NVILA on our annotations, these models now outperform the closed-source Gemini-3-Flash and GPT-5.1 on spatial understanding tasks across autonomous driving, robotics, and everyday scenarios. 📜Paper: 🌐Webpage: 💻 Code: 🕸️Model: 📊 Dataset: 👉 Interactive Demo: Let’s move research forward together. FoundationMotion is also referred to as Wolf V2 🐺, the second chapter in the Wolf series:
1:25

Introducing FoundationMotion. A large-scale, video-derived motion annotation dataset & auto-labeling pipeline + advanced models for motion understanding. Fully open-source: code, datasets, and models, free to use and build on. Understanding motion is core to physical reasoning, yet today’s leading models still struggle with simple spatial actions like “turn right” or “move up” or “flip the toast” - mainly due to the lack of large, fine-grained motion datasets. We present FoundationMotion, a fully automated pipeline that: • detects & tracks objects in videos • extracts trajectories • uses LLMs + frames to generate rich motion captions & QA pairs → creating large-scale, high-quality motion datasets at scale. After fine-tuning the open-source models Qwen and NVILA on our annotations, these models now outperform the closed-source Gemini-3-Flash and GPT-5.1 on spatial understanding tasks across autonomous driving, robotics, and everyday scenarios. 📜Paper: 🌐Webpage: 💻 Code: 🕸️Model: 📊 Dataset: 👉 Interactive Demo: Let’s move research forward together. FoundationMotion is also referred to as Wolf V2 🐺, the second chapter in the Wolf series:

Boyi Li

67,569 просмотров • 9 месяцев назад

TokenFlow: Consistent Diffusion Features for Consistent Video Editing paper page: The generative AI revolution has recently expanded to videos. Nevertheless, current state-of-the-art video models are still lagging behind image models in terms of visual quality and user control over the generated content. In this work, we present a framework that harnesses the power of a text-to-image diffusion model for the task of text-driven video editing. Specifically, given a source video and a target text-prompt, our method generates a high-quality video that adheres to the target text, while preserving the spatial layout and motion of the input video. Our method is based on a key observation that consistency in the edited video can be obtained by enforcing consistency in the diffusion feature space. We achieve this by explicitly propagating diffusion features based on inter-frame correspondences, readily available in the model. Thus, our framework does not require any training or fine-tuning, and can work in conjunction with any off-the-shelf text-to-image editing method. We demonstrate state-of-the-art editing results on a variety of real-world videos.
0:36

TokenFlow: Consistent Diffusion Features for Consistent Video Editing paper page: The generative AI revolution has recently expanded to videos. Nevertheless, current state-of-the-art video models are still lagging behind image models in terms of visual quality and user control over the generated content. In this work, we present a framework that harnesses the power of a text-to-image diffusion model for the task of text-driven video editing. Specifically, given a source video and a target text-prompt, our method generates a high-quality video that adheres to the target text, while preserving the spatial layout and motion of the input video. Our method is based on a key observation that consistency in the edited video can be obtained by enforcing consistency in the diffusion feature space. We achieve this by explicitly propagating diffusion features based on inter-frame correspondences, readily available in the model. Thus, our framework does not require any training or fine-tuning, and can work in conjunction with any off-the-shelf text-to-image editing method. We demonstrate state-of-the-art editing results on a variety of real-world videos.

AK

143,469 просмотров • 3 лет назад

Nvidia announces GAvatar: Animatable 3D Gaussian Avatars with Implicit Mesh Learning paper page: Gaussian splatting has emerged as a powerful 3D representation that harnesses the advantages of both explicit (mesh) and implicit (NeRF) 3D representations. In this paper, we seek to leverage Gaussian splatting to generate realistic animatable avatars from textual descriptions, addressing the limitations (e.g., flexibility and efficiency) imposed by mesh or NeRF-based representations. However, a naive application of Gaussian splatting cannot generate high-quality animatable avatars and suffers from learning instability; it also cannot capture fine avatar geometries and often leads to degenerate body parts. To tackle these problems, we first propose a primitive-based 3D Gaussian representation where Gaussians are defined inside pose-driven primitives to facilitate animation. Second, to stabilize and amortize the learning of millions of Gaussians, we propose to use neural implicit fields to predict the Gaussian attributes (e.g., colors). Finally, to capture fine avatar geometries and extract detailed meshes, we propose a novel SDF-based implicit mesh learning approach for 3D Gaussians that regularizes the underlying geometries and extracts highly detailed textured meshes. Our proposed method, GAvatar, enables the large-scale generation of diverse animatable avatars using only text prompts. GAvatar significantly surpasses existing methods in terms of both appearance and geometry quality, and achieves extremely fast rendering (100 fps) at 1K resolution.
5:36

Nvidia announces GAvatar: Animatable 3D Gaussian Avatars with Implicit Mesh Learning paper page: Gaussian splatting has emerged as a powerful 3D representation that harnesses the advantages of both explicit (mesh) and implicit (NeRF) 3D representations. In this paper, we seek to leverage Gaussian splatting to generate realistic animatable avatars from textual descriptions, addressing the limitations (e.g., flexibility and efficiency) imposed by mesh or NeRF-based representations. However, a naive application of Gaussian splatting cannot generate high-quality animatable avatars and suffers from learning instability; it also cannot capture fine avatar geometries and often leads to degenerate body parts. To tackle these problems, we first propose a primitive-based 3D Gaussian representation where Gaussians are defined inside pose-driven primitives to facilitate animation. Second, to stabilize and amortize the learning of millions of Gaussians, we propose to use neural implicit fields to predict the Gaussian attributes (e.g., colors). Finally, to capture fine avatar geometries and extract detailed meshes, we propose a novel SDF-based implicit mesh learning approach for 3D Gaussians that regularizes the underlying geometries and extracts highly detailed textured meshes. Our proposed method, GAvatar, enables the large-scale generation of diverse animatable avatars using only text prompts. GAvatar significantly surpasses existing methods in terms of both appearance and geometry quality, and achieves extremely fast rendering (100 fps) at 1K resolution.

AK

141,058 просмотров • 2 лет назад

Can an inexpensive, off-the-shelf IMU be the only sensor to estimate the full state (position, velocity, orientation) of a quadrotor flying through a track at high speed and even be on-pair with vision-based localization? The answer is yes, within certain limitations! In this #RAL2023 paper, we propose a learning-based odometry algorithm that couples a model-based filter driven by the inertial measurements with a learning-based module with access to the control commands. Our system outperforms by a large margin the state-of-the-art visual-inertial odometry (#VIO) algorithms and the state-of-the-art learned-inertial odometry algorithm, #TLIO, for the task of drone racing. Additionally, we show that our system is as accurate as a VIO algorithm that uses a camera to localize to a known map of the racing track. The main limitation of our approach is that it cannot generalize to trajectories that have not been seen at training time. However, in drone racing competitions, the track is known beforehand. Human pilots spend hours or even days of practice on the race track before the competition. Similarly, our system can be trained with the data collected during practice time and deployed during the competition. Future work will investigate how to generalize to trajectories not seen at training time. The code is released! Paper: Video: Code: Kudos to Giovanni Cioffi Leonard Bauersfeld Elia Kaufmann European Research Council (ERC) University of Zurich UZH Science UZH Space Hub NCCR Robotics Aerial Core #RAL2023 #IROS2023 #SLAM
0:31

Can an inexpensive, off-the-shelf IMU be the only sensor to estimate the full state (position, velocity, orientation) of a quadrotor flying through a track at high speed and even be on-pair with vision-based localization? The answer is yes, within certain limitations! In this #RAL2023 paper, we propose a learning-based odometry algorithm that couples a model-based filter driven by the inertial measurements with a learning-based module with access to the control commands. Our system outperforms by a large margin the state-of-the-art visual-inertial odometry (#VIO) algorithms and the state-of-the-art learned-inertial odometry algorithm, #TLIO, for the task of drone racing. Additionally, we show that our system is as accurate as a VIO algorithm that uses a camera to localize to a known map of the racing track. The main limitation of our approach is that it cannot generalize to trajectories that have not been seen at training time. However, in drone racing competitions, the track is known beforehand. Human pilots spend hours or even days of practice on the race track before the competition. Similarly, our system can be trained with the data collected during practice time and deployed during the competition. Future work will investigate how to generalize to trajectories not seen at training time. The code is released! Paper: Video: Code: Kudos to Giovanni Cioffi Leonard Bauersfeld Elia Kaufmann European Research Council (ERC) University of Zurich UZH Science UZH Space Hub NCCR Robotics Aerial Core #RAL2023 #IROS2023 #SLAM

Davide Scaramuzza

37,061 просмотров • 3 лет назад

The most important tool in Probability and Statistics - Markov Chain Monte Carlo (MCMC) Method Fresh out of undergraduate Probability and Stats courses, it’s easy to feel invincible. You’ve tamed Gaussians, gammas, betas, all those neat closed-form toy distributions. Then research hits and you meet the harsher truth. Real posteriors and energy landscapes are jagged, asymmetric, multimodal, and too high-dimensional to integrate or sample from directly. You can’t compute the normalising constant. You can’t do the integrals by hand. And i.i.d. samples are basically science fiction. Markov Chain Monte Carlo is the hack we invented to survive that reality. Instead of drawing perfect samples, you send a carefully designed random walk wandering through the landscape, then use its long-run positions as your window into the target distribution. Here’s the problem. Standard trace plots and diagnostics can still cheerfully lie to you. High-dimensional geometry can make a chain that looks healthy while it’s effectively frozen. Multimodal targets, bad tuning, and hidden correlations can quietly wreck your posterior summaries. This series is about those blind spots. We’ll use visuals like this one to show how MCMC actually moves, where the guarantees get slippery, and how to think clearly about convergence and diagnostics in serious Bayesian, physics, and ML work. #BayesianInference #MCMC #MonteCarloMethods #ProbabilityLandscape #StatisticsEducation #ComputationalScience
0:59

The most important tool in Probability and Statistics - Markov Chain Monte Carlo (MCMC) Method Fresh out of undergraduate Probability and Stats courses, it’s easy to feel invincible. You’ve tamed Gaussians, gammas, betas, all those neat closed-form toy distributions. Then research hits and you meet the harsher truth. Real posteriors and energy landscapes are jagged, asymmetric, multimodal, and too high-dimensional to integrate or sample from directly. You can’t compute the normalising constant. You can’t do the integrals by hand. And i.i.d. samples are basically science fiction. Markov Chain Monte Carlo is the hack we invented to survive that reality. Instead of drawing perfect samples, you send a carefully designed random walk wandering through the landscape, then use its long-run positions as your window into the target distribution. Here’s the problem. Standard trace plots and diagnostics can still cheerfully lie to you. High-dimensional geometry can make a chain that looks healthy while it’s effectively frozen. Multimodal targets, bad tuning, and hidden correlations can quietly wreck your posterior summaries. This series is about those blind spots. We’ll use visuals like this one to show how MCMC actually moves, where the guarantees get slippery, and how to think clearly about convergence and diagnostics in serious Bayesian, physics, and ML work. #BayesianInference #MCMC #MonteCarloMethods #ProbabilityLandscape #StatisticsEducation #ComputationalScience

Mathelirium

32,296 просмотров • 7 месяцев назад

New framework: Kick down your robot, it will get back up every time 🥋 Chinese startup RoboParty is a Beijing startup founded April 2025 by Huang Yi, originally shipping ROBOTO ORIGIN, the world's first full-stack open-source bipedal humanoid. They released UFO: Unsupervised Reinforcement Learning Framework for Humanoid Control. DEFINITIONS -> what differs is where the learning signal comes from: - SUPERVISED: humans supply the right answers (labels), the model imitates them. - UNSUPERVISED: no answer key, the model finds structure in raw data on its own. - REINFORCEMENT LEARNING: no answer key either, the model tries things and a reward scores each attempt. → UNSUPERVISED RL: trial and error where the agent invents its own rewards, instead of engineers hand-writing one per task. REPRESENTATION LEARNING: compress raw states into a useful internal map. TEMPORAL DISTANCE: distance on that map is "how many steps from A to B." CONTRASTIVE: trained by pulling together what's close in time, pushing apart what isn't. -> CONTRASTIVE TEMPORAL-DISTANCE REPRESENTATION LEARNING: the model builds an internal map of body states where distance means how many steps it takes to get from one to another. It is trained by contrast: states that occur close together in a movement get pulled together in the map, randomly paired states get pushed apart. UFO is an open-source training framework that teaches humanoid robots skills, like getting up, walking, goal-reaching, teleoperation, without reference motions -> no motion-capture or human-video demonstrations to imitate. Its core is TeCH, a contrastive temporal-distance representation-learning algorithm: the robot explores, builds pseudo-goals by temporal rolling, and learns goal-conditioned policies from a single unified progress reward. One framework trains five different robots (Unitree G1/H1, RoboParty RP0/RP1, AgiBot X2) with automatic config conversion in ~2–3 hours per robot! The real novelty here "no demonstrations at all". No data-collection arms race,the dominant humanoid-locomotion recipe is tracking: imitate mocap/retargeted-human reference trajectories. The robot self-generates goals from its own exploration and learns from a progress reward, needing zero reference motion data. Everybody else is fighting over data acquisition, while this team just teleports out of the race entirely (inb4 "competition is for losers 💀 ). This strategy reminds me of the DeepSeek playbook applied to robots: open-source the whole stack to become the global default and commoditize everyone else. RoboParty is giving away hardware and now control software (UFO) to be the Android of humanoids. Yet another reason for the US to ban Chinese open models perhaps 🥶 ? What I also really like about this approach is the cross-embodiment infrastructure, one framework trains Unitree G1/H1, RoboParty RP0/RP1, and AgiBot X2 with automatic configuration conversion. Just like Physical Intelligence, RoboParty seems to place itself as a neutral hardware agnostic middle man. Also woth mentioning: their ability ot perform stable skill injection, e.g. adding a cartwheel without forgetting how to walk. A common failure of RL humanoid policies is that teaching a new agile skill destabilizes the existing ones (catastrophic forgetting). UFO claims you can inject rare motions (cartwheel) without collapsing learned behavior. If it holds, that's a significant incremental/continual skill-learning! But again, I have to underline it: no arXiv, no external validation, no success-rate numbers. -> robotics badely needs an independent unbiased evaluator imho. Still, look at that cool demo: robot is getting kicked and pushed around (serious disturbance) during teleoperation (controlled the person at the back wearing the VR headset), and still managed to always get back up. This is some serious demonstration of stability and robustness!
0:49

New framework: Kick down your robot, it will get back up every time 🥋 Chinese startup RoboParty is a Beijing startup founded April 2025 by Huang Yi, originally shipping ROBOTO ORIGIN, the world's first full-stack open-source bipedal humanoid. They released UFO: Unsupervised Reinforcement Learning Framework for Humanoid Control. DEFINITIONS -> what differs is where the learning signal comes from: - SUPERVISED: humans supply the right answers (labels), the model imitates them. - UNSUPERVISED: no answer key, the model finds structure in raw data on its own. - REINFORCEMENT LEARNING: no answer key either, the model tries things and a reward scores each attempt. → UNSUPERVISED RL: trial and error where the agent invents its own rewards, instead of engineers hand-writing one per task. REPRESENTATION LEARNING: compress raw states into a useful internal map. TEMPORAL DISTANCE: distance on that map is "how many steps from A to B." CONTRASTIVE: trained by pulling together what's close in time, pushing apart what isn't. -> CONTRASTIVE TEMPORAL-DISTANCE REPRESENTATION LEARNING: the model builds an internal map of body states where distance means how many steps it takes to get from one to another. It is trained by contrast: states that occur close together in a movement get pulled together in the map, randomly paired states get pushed apart. UFO is an open-source training framework that teaches humanoid robots skills, like getting up, walking, goal-reaching, teleoperation, without reference motions -> no motion-capture or human-video demonstrations to imitate. Its core is TeCH, a contrastive temporal-distance representation-learning algorithm: the robot explores, builds pseudo-goals by temporal rolling, and learns goal-conditioned policies from a single unified progress reward. One framework trains five different robots (Unitree G1/H1, RoboParty RP0/RP1, AgiBot X2) with automatic config conversion in ~2–3 hours per robot! The real novelty here "no demonstrations at all". No data-collection arms race,the dominant humanoid-locomotion recipe is tracking: imitate mocap/retargeted-human reference trajectories. The robot self-generates goals from its own exploration and learns from a progress reward, needing zero reference motion data. Everybody else is fighting over data acquisition, while this team just teleports out of the race entirely (inb4 "competition is for losers 💀 ). This strategy reminds me of the DeepSeek playbook applied to robots: open-source the whole stack to become the global default and commoditize everyone else. RoboParty is giving away hardware and now control software (UFO) to be the Android of humanoids. Yet another reason for the US to ban Chinese open models perhaps 🥶 ? What I also really like about this approach is the cross-embodiment infrastructure, one framework trains Unitree G1/H1, RoboParty RP0/RP1, and AgiBot X2 with automatic configuration conversion. Just like Physical Intelligence, RoboParty seems to place itself as a neutral hardware agnostic middle man. Also woth mentioning: their ability ot perform stable skill injection, e.g. adding a cartwheel without forgetting how to walk. A common failure of RL humanoid policies is that teaching a new agile skill destabilizes the existing ones (catastrophic forgetting). UFO claims you can inject rare motions (cartwheel) without collapsing learned behavior. If it holds, that's a significant incremental/continual skill-learning! But again, I have to underline it: no arXiv, no external validation, no success-rate numbers. -> robotics badely needs an independent unbiased evaluator imho. Still, look at that cool demo: robot is getting kicked and pushed around (serious disturbance) during teleoperation (controlled the person at the back wearing the VR headset), and still managed to always get back up. This is some serious demonstration of stability and robustness!

Léo

35,855 просмотров • 1 месяц назад

This kinetic sculpture, titled Nested Loops, hangs inside The National Museum of Mathematics (MoMath) in New York. It demonstrates exactly how higher-dimensional forms can emerge from the coordinated motion of simple elements moving through time. The circles are pathways of motion…timing mechanisms that guide anchor points along curved trajectories. As the mechanism rotates, strings stretched between these moving anchor points remain under constant tension. Because a taut string always forms the shortest path between two points, each string becomes a straight line, even though the points guiding it are traveling along curves! This is the key insight…circular motion is generating linear geometry. The loops create movement, the strings reveal structure, and as these relationships shift in coordinated time, the eye begins to perceive cubic forms like edges, diagonals, and vertices, all emerging from what is fundamentally a flat plane with moving parts. A cube does not need to be physically present to be perceived. It can be implied through the alignment of edges, the intersection of planes, and the convergence of perspective. What you’re witnessing is a dimensional emergence…a lower-dimensional system producing the visual signature of a higher-dimensional object through synchronized motion. Change the speed of one circle relative to another, and an entirely different geometric form appears. The same strings, the same frame…but a new structure emerges from the new rhythm. This is why the sculpture feels almost alive. It’s revealing that form is relational, temporal, and emergent, so what we perceive as solid objects are often just stable harmonies of moving relationships, frozen in a moment of coherence. This piece asks you to consider: What aspects of your reality are you perceiving as solid, when they’re actually just coherent patterns of motion temporarily aligned? Did this shift how you see structure and emergence? ✨🙌🏾💫 Artist © Chuck Hoberman (MoMath, New York)
0:25

This kinetic sculpture, titled Nested Loops, hangs inside The National Museum of Mathematics (MoMath) in New York. It demonstrates exactly how higher-dimensional forms can emerge from the coordinated motion of simple elements moving through time. The circles are pathways of motion…timing mechanisms that guide anchor points along curved trajectories. As the mechanism rotates, strings stretched between these moving anchor points remain under constant tension. Because a taut string always forms the shortest path between two points, each string becomes a straight line, even though the points guiding it are traveling along curves! This is the key insight…circular motion is generating linear geometry. The loops create movement, the strings reveal structure, and as these relationships shift in coordinated time, the eye begins to perceive cubic forms like edges, diagonals, and vertices, all emerging from what is fundamentally a flat plane with moving parts. A cube does not need to be physically present to be perceived. It can be implied through the alignment of edges, the intersection of planes, and the convergence of perspective. What you’re witnessing is a dimensional emergence…a lower-dimensional system producing the visual signature of a higher-dimensional object through synchronized motion. Change the speed of one circle relative to another, and an entirely different geometric form appears. The same strings, the same frame…but a new structure emerges from the new rhythm. This is why the sculpture feels almost alive. It’s revealing that form is relational, temporal, and emergent, so what we perceive as solid objects are often just stable harmonies of moving relationships, frozen in a moment of coherence. This piece asks you to consider: What aspects of your reality are you perceiving as solid, when they’re actually just coherent patterns of motion temporarily aligned? Did this shift how you see structure and emergence? ✨🙌🏾💫 Artist © Chuck Hoberman (MoMath, New York)

🧬Maxpein🧬

48,005 просмотров • 3 месяцев назад

Check out our #RSS2024 paper "#MPCC++: Model Predictive Contouring Control for Time-Optimal Flight with Safety Constraints." Model Predictive Contouring Control (MPCC) has shown promising results for agile robotics applications, including car and drone racing. Existing approaches struggle to introduce safety considerations, often resulting in crashes. What does it take to drive or fly fast and safe? We enhance our former MPCC by incorporating spatial constraints that reliably prevent obstacle collisions, allowing planning the fastest trajectory within these safety limits. To improve performance, we leverage real-world data to refine the dynamic model. Our approach is the first to achieve a 100% success rate in real-world experiments. This safety benefit comes without compromising performance, as our method achieves lap times comparable to the best-performing state-based #ReinforcementLearning (RL) policies. Reference M. Krinner, A. Romero, L. Bauersfeld, M. Zeilinger, A. Carron, D. Scaramuzza, "MPCC++: Model Predictive Contouring Control for Time-Optimal Flight with Safety Constraints" Robotics, Science and Systems, 2024 PDF: Video: Kudos to Maria Krinner, Angel Romero Aguilar, Leonard Bauersfeld, Melanie Zeilinger, Andrea Carron! Ángel Romero Leonard Bauersfeld University of Zurich UZH Science UZH Space Hub AUTOASSESS European Research Council (ERC) #MPC #ModelPredictiveControl
0:47

Check out our #RSS2024 paper "#MPCC++: Model Predictive Contouring Control for Time-Optimal Flight with Safety Constraints." Model Predictive Contouring Control (MPCC) has shown promising results for agile robotics applications, including car and drone racing. Existing approaches struggle to introduce safety considerations, often resulting in crashes. What does it take to drive or fly fast and safe? We enhance our former MPCC by incorporating spatial constraints that reliably prevent obstacle collisions, allowing planning the fastest trajectory within these safety limits. To improve performance, we leverage real-world data to refine the dynamic model. Our approach is the first to achieve a 100% success rate in real-world experiments. This safety benefit comes without compromising performance, as our method achieves lap times comparable to the best-performing state-based #ReinforcementLearning (RL) policies. Reference M. Krinner, A. Romero, L. Bauersfeld, M. Zeilinger, A. Carron, D. Scaramuzza, "MPCC++: Model Predictive Contouring Control for Time-Optimal Flight with Safety Constraints" Robotics, Science and Systems, 2024 PDF: Video: Kudos to Maria Krinner, Angel Romero Aguilar, Leonard Bauersfeld, Melanie Zeilinger, Andrea Carron! Ángel Romero Leonard Bauersfeld University of Zurich UZH Science UZH Space Hub AUTOASSESS European Research Council (ERC) #MPC #ModelPredictiveControl

Davide Scaramuzza

17,903 просмотров • 2 лет назад

We believe we’re the first robotics company to demonstrate a robot peeling an apple with dual dexterous human-like hands. This breakthrough closes a key gap in robotics, achieving bimanual, contact-rich manipulation and moving far beyond the limits of simple grippers. 🧵↓ Today’s AI models (VLMs) are excellent at perception but struggle with action. Controlling high-degree-of-freedom hands for tasks like this is incredibly complex, and precise finger-level teleoperation is nearly impossible for humans. Our first step was a shared-autonomy system: rather than controlling every finger, the operator triggers pre-learned skills like a “rotate apple or tennis ball” primitive via a keyboard press or pedal. This makes scalable data collection and RL training possible. How does the AI manage this? We created "MoDE-VLA" (Mixture of Dexterous Experts). It fuses vision, language, force, and touch data by using a team of specialist "experts," making control in high-dimensional spaces stable and effective. The combination of these two innovations allows for seamless, contact-rich manipulation. The human provides high-level guidance, and the robot executes the complex in-hand coordination required. This work paves the way for robots that can safely handle delicate tasks in human environments. Want the full technical details? 📄 Read the full research paper: Visit us at NVIDIA GTC Booth #1838, Hall 3 to learn more! #Robotics #AI #DexterousManipulation #VLA #NVIDIAGTC Nancy Villicaña NVIDIA GTC
2:51

We believe we’re the first robotics company to demonstrate a robot peeling an apple with dual dexterous human-like hands. This breakthrough closes a key gap in robotics, achieving bimanual, contact-rich manipulation and moving far beyond the limits of simple grippers. 🧵↓ Today’s AI models (VLMs) are excellent at perception but struggle with action. Controlling high-degree-of-freedom hands for tasks like this is incredibly complex, and precise finger-level teleoperation is nearly impossible for humans. Our first step was a shared-autonomy system: rather than controlling every finger, the operator triggers pre-learned skills like a “rotate apple or tennis ball” primitive via a keyboard press or pedal. This makes scalable data collection and RL training possible. How does the AI manage this? We created "MoDE-VLA" (Mixture of Dexterous Experts). It fuses vision, language, force, and touch data by using a team of specialist "experts," making control in high-dimensional spaces stable and effective. The combination of these two innovations allows for seamless, contact-rich manipulation. The human provides high-level guidance, and the robot executes the complex in-hand coordination required. This work paves the way for robots that can safely handle delicate tasks in human environments. Want the full technical details? 📄 Read the full research paper: Visit us at NVIDIA GTC Booth #1838, Hall 3 to learn more! #Robotics #AI #DexterousManipulation #VLA #NVIDIAGTC Nancy Villicaña NVIDIA GTC

Sharpa

20,429 просмотров • 6 месяцев назад

We are thrilled to share our breakthrough research on "Agile Flight from Pixels without State Estimation," to be presented and live-demonstrated at #RSS2024 next week! You heard well: no state estimation means no explicit visual localization, no SLAM, no VIO, and no IMU! Paper: Video (Narrated): Last year, we demonstrated that #ReinforcementLearning (RL) policies could outperform world-champion drone-racing pilots using the same quadrotor hardware; however, unlike human pilots, these policies continuously estimated an explicit state from known gate positions, the camera feed, and inertial measurements (IMU). In this new work, we tackle the challenge of learning vision-based drone racing using an end-to-end reinforcement learning approach that eliminates the need for IMU data or explicit state estimation. Like professional pilots, we go directly from images to control commands. The training is facilitated by an asymmetric actor-critic with access to privileged information. To overcome the computational complexity during image-based RL training, we use an appropriate sensor representation, which can be efficiently simulated during training without rendering images. We achieve agile flight at speeds up to 40 km/h with accelerations up to 2 g's. Although our demonstration focuses on drone racing, we believe that our method has an impact beyond drone racing and can serve as a foundation for future research into real-world applications in structured environments. Besides the paper presentation, we will also give a live demo next Tuesday and Wednesday between and hrs at TU Delft: Reference: Ismail Geles*, Leonard Bauersfeld*, Angel Romero, Jiaxu Xing, Davide Scaramuzza "Demonstrating Agile Flight from Pixels without State Estimation" Robotics: Science and Systems (RSS), 2024. Kudos to Ismail Geles Leonard Bauersfeld Ángel Romero Jiaxu Xing! University of Zurich UZH Science UZH Space Hub Aerial Core AUTOASSESS European Research Council (ERC)
1:24

We are thrilled to share our breakthrough research on "Agile Flight from Pixels without State Estimation," to be presented and live-demonstrated at #RSS2024 next week! You heard well: no state estimation means no explicit visual localization, no SLAM, no VIO, and no IMU! Paper: Video (Narrated): Last year, we demonstrated that #ReinforcementLearning (RL) policies could outperform world-champion drone-racing pilots using the same quadrotor hardware; however, unlike human pilots, these policies continuously estimated an explicit state from known gate positions, the camera feed, and inertial measurements (IMU). In this new work, we tackle the challenge of learning vision-based drone racing using an end-to-end reinforcement learning approach that eliminates the need for IMU data or explicit state estimation. Like professional pilots, we go directly from images to control commands. The training is facilitated by an asymmetric actor-critic with access to privileged information. To overcome the computational complexity during image-based RL training, we use an appropriate sensor representation, which can be efficiently simulated during training without rendering images. We achieve agile flight at speeds up to 40 km/h with accelerations up to 2 g's. Although our demonstration focuses on drone racing, we believe that our method has an impact beyond drone racing and can serve as a foundation for future research into real-world applications in structured environments. Besides the paper presentation, we will also give a live demo next Tuesday and Wednesday between and hrs at TU Delft: Reference: Ismail Geles*, Leonard Bauersfeld*, Angel Romero, Jiaxu Xing, Davide Scaramuzza "Demonstrating Agile Flight from Pixels without State Estimation" Robotics: Science and Systems (RSS), 2024. Kudos to Ismail Geles Leonard Bauersfeld Ángel Romero Jiaxu Xing! University of Zurich UZH Science UZH Space Hub Aerial Core AUTOASSESS European Research Council (ERC)

Davide Scaramuzza

27,959 просмотров • 2 лет назад

Today, we're joined by Nikita Rudin, co-founder and CEO of Flexion to discuss the gap between current robotic capabilities and what’s required to deploy fully autonomous robots in the real world. Nikita explains how reinforcement learning and simulation have driven rapid progress in robot locomotion—and why locomotion is still far from “solved.” We dig into the sim2real gap, and how adding visual inputs introduces noise and significantly complicates sim-to-real transfer. We also explore the debate between end-to-end models and modular approaches, and why separating locomotion, planning, and semantics remains a pragmatic approach today. Nikita also introduces the concept of "real-to-sim", which uses real-world data to refine simulation parameters for higher fidelity training, discusses how reinforcement learning, imitation learning, and teleoperation data are combined to train robust policies for both quadruped and humanoid robots, and introduces Flexion's hierarchical approach that utilizes pre-trained Vision-Language Models (VLMs) for high-level task orchestration with Vision-Language-Action (VLA) models and low-level whole-body trackers. Finally, Nikita shares the behind-the-scenes in humanoid robot demos, his take on reinforcement learning in simulation versus the real world, the nuances of reward tuning, and offers practical advice for researchers and practitioners looking to get started in robotics today. 🗒️ For the full list of resources for this episode, visit the show notes page: 📖 CHAPTERS =============================== 00:00 - Introduction 04:07 - Is robot locomotion solved? 06:04 - Sim-to-real gap 08:58 - Adding semantics to policies 09:42 - Modular vs end-to-end architectures 10:29 - Planner model 12:21 - Adapting RL techniques from quadrupeds to humanoids 15:39 - Behind robot demos 18:09 - Humanoid robots in home environments 22:03 - Training approach 23:56 - VLA models 27:59 - Closing the sim-to-real gap 32:55 - Task orchestration using VLMs 36:38 - Tool use 38:10 - Model hierarchy 43:37 - Simulator versus simulation environment 44:57 - Combining imitation learning and reinforcement learning 46:42 - RL in real world versus RL in simulation 52:58 - Reward tuning and value functions in robotics 56:38 - Predictions 1:00:10 - Humanoids, quadropeds, and wheeled platforms 1:02:45 - Advice, recommended robot kits, and community pla
1:06:07

Today, we're joined by Nikita Rudin, co-founder and CEO of Flexion to discuss the gap between current robotic capabilities and what’s required to deploy fully autonomous robots in the real world. Nikita explains how reinforcement learning and simulation have driven rapid progress in robot locomotion—and why locomotion is still far from “solved.” We dig into the sim2real gap, and how adding visual inputs introduces noise and significantly complicates sim-to-real transfer. We also explore the debate between end-to-end models and modular approaches, and why separating locomotion, planning, and semantics remains a pragmatic approach today. Nikita also introduces the concept of "real-to-sim", which uses real-world data to refine simulation parameters for higher fidelity training, discusses how reinforcement learning, imitation learning, and teleoperation data are combined to train robust policies for both quadruped and humanoid robots, and introduces Flexion's hierarchical approach that utilizes pre-trained Vision-Language Models (VLMs) for high-level task orchestration with Vision-Language-Action (VLA) models and low-level whole-body trackers. Finally, Nikita shares the behind-the-scenes in humanoid robot demos, his take on reinforcement learning in simulation versus the real world, the nuances of reward tuning, and offers practical advice for researchers and practitioners looking to get started in robotics today. 🗒️ For the full list of resources for this episode, visit the show notes page: 📖 CHAPTERS =============================== 00:00 - Introduction 04:07 - Is robot locomotion solved? 06:04 - Sim-to-real gap 08:58 - Adding semantics to policies 09:42 - Modular vs end-to-end architectures 10:29 - Planner model 12:21 - Adapting RL techniques from quadrupeds to humanoids 15:39 - Behind robot demos 18:09 - Humanoid robots in home environments 22:03 - Training approach 23:56 - VLA models 27:59 - Closing the sim-to-real gap 32:55 - Task orchestration using VLMs 36:38 - Tool use 38:10 - Model hierarchy 43:37 - Simulator versus simulation environment 44:57 - Combining imitation learning and reinforcement learning 46:42 - RL in real world versus RL in simulation 52:58 - Reward tuning and value functions in robotics 56:38 - Predictions 1:00:10 - Humanoids, quadropeds, and wheeled platforms 1:02:45 - Advice, recommended robot kits, and community pla

The TWIML AI Podcast

22,592 просмотров • 8 месяцев назад

Reinforcement Learning from Human Feedback (RLHF) is gaining traction. This field aims to make AI more responsible by including human values and preferences. In this video, Nathan Lambert, a research scientist and RLHF team lead at Hugging Face explores its inner workings, applications and industry impact. RLHF has gained the spotlight in recent years. The growth of language models like Anthropic’s Claude and OpenAI's ChatGPT have increased interest in human-feedback integration. "There are some rumors that Open AI had two teams; one was doing RLHF and the other instruction fine-tuning. And the RLHF team kept getting more and more performance." Understanding RLHF The RLHF process has three main steps: Pre-training: Much like with GPT models, the journey starts with pre-training on a large corpus of data. This can range from text data, web scrapes, to specialized datasets. Reward Modeling: This is the RLHF counterpart of supervised fine-tuning in large language models. This stage involves creating a reward model that resonates with human values and preferences. RL Optimization: This stage parallels reward modeling and reinforcement learning in traditional AI models. The AI system fine-tunes itself based on the reward model, employing reinforcement learning algorithms for that extra layer of optimization. The Data Challenge Data collection and curation in RLHF closely resemble the challenges you'd encounter in large language model training. Datasets from organizations like OpenAI can serve as a useful foundation. However, the need for high-quality, task-specific data cannot be overstated. Implementing RLHF: A Practical Guide If you’re someone who loves getting hands-on with AI libraries like Hugging Face, implementing RLHF is right way to do. It’s essential to understand its limitations. Think about model stability, over-optimization, and exploration strategies, much like you would when prompt engineering. Ongoing Research and Next Steps While he suggests that some basics figured out, there are layers of complexity that still need to be unraveled: 1. New Benchmarks: How do we measure the effectiveness of RLHF? 2. Preference Modeling: How can the model be made to understand human preferences better? 3. Interpreting RLHF: Much like explainability in traditional models, how do we make RLHF more interpretable? 4. System-Wide Evaluation: Going beyond individual performance, how does RLHF affect an entire system? The Transformative Power of RLHF Whether you're an AI developer, a business analyst, or a marketer, RLHF promises to revolutionize your domain. Imagine customer service chatbots that understand human emotions better, or content generators that align more closely with human values. RLHF is an emerging field that focuses on enhancing machine learning models through human feedback. While it tackles important issues like bias and ethics, its broader goal is to improve system performance across various applications. Whether you're deeply invested in the ethics of AI or simply curious about advancements in machine learning, RLHF offers valuable insights. If you're interested in the next wave of AI development, this area is definitely one to watch.
2:00

Reinforcement Learning from Human Feedback (RLHF) is gaining traction. This field aims to make AI more responsible by including human values and preferences. In this video, Nathan Lambert, a research scientist and RLHF team lead at Hugging Face explores its inner workings, applications and industry impact. RLHF has gained the spotlight in recent years. The growth of language models like Anthropic’s Claude and OpenAI's ChatGPT have increased interest in human-feedback integration. "There are some rumors that Open AI had two teams; one was doing RLHF and the other instruction fine-tuning. And the RLHF team kept getting more and more performance." Understanding RLHF The RLHF process has three main steps: Pre-training: Much like with GPT models, the journey starts with pre-training on a large corpus of data. This can range from text data, web scrapes, to specialized datasets. Reward Modeling: This is the RLHF counterpart of supervised fine-tuning in large language models. This stage involves creating a reward model that resonates with human values and preferences. RL Optimization: This stage parallels reward modeling and reinforcement learning in traditional AI models. The AI system fine-tunes itself based on the reward model, employing reinforcement learning algorithms for that extra layer of optimization. The Data Challenge Data collection and curation in RLHF closely resemble the challenges you'd encounter in large language model training. Datasets from organizations like OpenAI can serve as a useful foundation. However, the need for high-quality, task-specific data cannot be overstated. Implementing RLHF: A Practical Guide If you’re someone who loves getting hands-on with AI libraries like Hugging Face, implementing RLHF is right way to do. It’s essential to understand its limitations. Think about model stability, over-optimization, and exploration strategies, much like you would when prompt engineering. Ongoing Research and Next Steps While he suggests that some basics figured out, there are layers of complexity that still need to be unraveled: 1. New Benchmarks: How do we measure the effectiveness of RLHF? 2. Preference Modeling: How can the model be made to understand human preferences better? 3. Interpreting RLHF: Much like explainability in traditional models, how do we make RLHF more interpretable? 4. System-Wide Evaluation: Going beyond individual performance, how does RLHF affect an entire system? The Transformative Power of RLHF Whether you're an AI developer, a business analyst, or a marketer, RLHF promises to revolutionize your domain. Imagine customer service chatbots that understand human emotions better, or content generators that align more closely with human values. RLHF is an emerging field that focuses on enhancing machine learning models through human feedback. While it tackles important issues like bias and ethics, its broader goal is to improve system performance across various applications. Whether you're deeply invested in the ethics of AI or simply curious about advancements in machine learning, RLHF offers valuable insights. If you're interested in the next wave of AI development, this area is definitely one to watch.

Muratcan Koylan

27,168 просмотров • 3 лет назад

.Donald J. Trump Mr. Trump, I'm about to dismantle every single one of your falsehoods from the first 2024 debate with .President Donald J. Trump in this thread 🧵. This may take a while due to the volume of lies, but rest assured, I won’t stop until every lie is exposed for all to see. First, when the current administration took office in 2021, the pandemic was in shambles under your watch, with thousands of Americans dying daily. Your neglect and lack of coherent strategy created an immense challenge to address. The so-called 'great economy' you boast about was merely riding on Obama's success, which you then tanked when a real crisis hit. Unemployment soared, millions of jobs disappeared, and the national debt ballooned like never before under your administration. Your pandemic response was a catastrophic failure: from underplaying the virus, hosting super-spreader rallies, to botching the rollout of essential supplies. Your claim of "no wars" is misleading. We remained entangled in ongoing conflicts, and you significantly ramped up drone strikes in some regions. The U.S. was still involved in ongoing conflicts in Afghanistan, Iraq, and Syria. Regarding the stock market recovery, you conveniently ignore the massive role played by Federal Reserve policies and Congressional stimulus and ridding Obama's recovery. It wasn't solely your administration's doing. Your nonsense about job creation under Biden is flat-out wrong. Millions of new jobs have been added across various sectors, not for undocumented immigrants or COVID recoveries. Lastly, your claim that inflation is "killing our country" is an exaggeration. Inflation has moderated significantly from its peak, and the economy has shown remarkable resilience. The U.S. had the fastest recovery thanks to President Biden's efforts to clean up your mess. The American people deserve a factual, balanced assessment, not overblown nonsense and misleading claims. To be continued...🧵 #DebatePresidencial2024 #DebateNight
0:58

.Donald J. Trump Mr. Trump, I'm about to dismantle every single one of your falsehoods from the first 2024 debate with .President Donald J. Trump in this thread 🧵. This may take a while due to the volume of lies, but rest assured, I won’t stop until every lie is exposed for all to see. First, when the current administration took office in 2021, the pandemic was in shambles under your watch, with thousands of Americans dying daily. Your neglect and lack of coherent strategy created an immense challenge to address. The so-called 'great economy' you boast about was merely riding on Obama's success, which you then tanked when a real crisis hit. Unemployment soared, millions of jobs disappeared, and the national debt ballooned like never before under your administration. Your pandemic response was a catastrophic failure: from underplaying the virus, hosting super-spreader rallies, to botching the rollout of essential supplies. Your claim of "no wars" is misleading. We remained entangled in ongoing conflicts, and you significantly ramped up drone strikes in some regions. The U.S. was still involved in ongoing conflicts in Afghanistan, Iraq, and Syria. Regarding the stock market recovery, you conveniently ignore the massive role played by Federal Reserve policies and Congressional stimulus and ridding Obama's recovery. It wasn't solely your administration's doing. Your nonsense about job creation under Biden is flat-out wrong. Millions of new jobs have been added across various sectors, not for undocumented immigrants or COVID recoveries. Lastly, your claim that inflation is "killing our country" is an exaggeration. Inflation has moderated significantly from its peak, and the economy has shown remarkable resilience. The U.S. had the fastest recovery thanks to President Biden's efforts to clean up your mess. The American people deserve a factual, balanced assessment, not overblown nonsense and misleading claims. To be continued...🧵 #DebatePresidencial2024 #DebateNight

Human☮🇺🇸🇺🇦🇺🇸🌊

209,510 просмотров • 2 лет назад

🚨 #SuperApp Updates 🚨 We are getting there and it will be worth the wait! 🚀 Our #SuperApp combines the best features of TrustWallet, Revolut and UniSwap to create an all-in-one platform that will revolutionize the way you manage your finances and crypto assets. Implementing such cutting-edge features comes with its own set of challenges but, rest assured, we're working diligently to overcome these obstacles and deliver a seamless experience to all our users and community🌍 While we are currently finalising the distribution process for all our testers, let me share a sneak peak of the #SuperApp and of our current points of focus👇 1️⃣ Scaling: With such a huge endeavor, we're rigorously testing our infrastructure to handle the influx of users. Rest assured, we won't let anything slow us down as we grow🛣️ 2️⃣ Security: As we are targeting high finances, both traditional and crypto, protecting your data and assets is our top priority. We're implementing robust security measures to ensure everyone’s peace of mind while using the #SuperApp🔐 3️⃣ Banking & Cards Integration: You will say goodbye to juggling multiple apps! Our #SuperApp will offer seamless integration with banking services and cards, making financial management a breeze! We are talking about multi-currency and multi-Iban accounts, easy access to freeze, unlock, limit or unlimit your cards and payments💥However, the testing of this feature is limited by the sandbox environment as the only testing method of cards is with actual real ones 🌕 But let's dive deeper into one extraordinary feature that sets us apart and which will be at the centre of testing phase 1: 4️⃣ Crypto Wallets: We're introducing something unique - the ability to view the total value of all your crypto assets across different wallets, multiple blockchains, in one place and we are levelling it up with instant top-up! An all in one non-custodial crypto wallet! No more switching between wallets to check your balances. It's all there at your fingertips, empowering you to make informed decisions, fast and easy🌊 We know you're excited, and we are too!✨ Together, we'll redefine the way we interact with all our finances!🚀 $ZPAY ZoidPay has become Another
5:38

🚨 #SuperApp Updates 🚨 We are getting there and it will be worth the wait! 🚀 Our #SuperApp combines the best features of TrustWallet, Revolut and UniSwap to create an all-in-one platform that will revolutionize the way you manage your finances and crypto assets. Implementing such cutting-edge features comes with its own set of challenges but, rest assured, we're working diligently to overcome these obstacles and deliver a seamless experience to all our users and community🌍 While we are currently finalising the distribution process for all our testers, let me share a sneak peak of the #SuperApp and of our current points of focus👇 1️⃣ Scaling: With such a huge endeavor, we're rigorously testing our infrastructure to handle the influx of users. Rest assured, we won't let anything slow us down as we grow🛣️ 2️⃣ Security: As we are targeting high finances, both traditional and crypto, protecting your data and assets is our top priority. We're implementing robust security measures to ensure everyone’s peace of mind while using the #SuperApp🔐 3️⃣ Banking & Cards Integration: You will say goodbye to juggling multiple apps! Our #SuperApp will offer seamless integration with banking services and cards, making financial management a breeze! We are talking about multi-currency and multi-Iban accounts, easy access to freeze, unlock, limit or unlimit your cards and payments💥However, the testing of this feature is limited by the sandbox environment as the only testing method of cards is with actual real ones 🌕 But let's dive deeper into one extraordinary feature that sets us apart and which will be at the centre of testing phase 1: 4️⃣ Crypto Wallets: We're introducing something unique - the ability to view the total value of all your crypto assets across different wallets, multiple blockchains, in one place and we are levelling it up with instant top-up! An all in one non-custodial crypto wallet! No more switching between wallets to check your balances. It's all there at your fingertips, empowering you to make informed decisions, fast and easy🌊 We know you're excited, and we are too!✨ Together, we'll redefine the way we interact with all our finances!🚀 $ZPAY ZoidPay has become Another

Eduard Oneci 🌊🌊🌊

13,923 просмотров • 3 лет назад

Hello everyone! 👋 I'm the Assistant of Chaliz! 🛠️ The second episode of recreating your OCs with LizMeta is here! 🎉 To those whose OCs weren't selected this time, it might be for various reasons, for example: We thought your OC required texture modifications. The specific features involved are still under development. We simply ran out of time and could only do this many customizations! etc... Please note that this doesn't mean your OC can't be created at all! Everyone's OCs are just too adorable! 🥰 I'm a bit scared I didn't do them justice, boo-hoo... 😭 (Haha, actually, I tried my best! I looked at your reference pictures, avatars, and even checked your profiles to guess your favorite styles... 👀) In fact, you can probably tell that the model has a huge potential range. If the OC I created looks different from your style, ahhhhh, please don't doubt the model!! 😱 It might just be a problem with my customization skills: ( A huge thank you to everyone who sent in pictures for our testing! Don't be discouraged if you weren't chosen this time! There is one last chance! For the final round, I will show the entire customization process, and the number of participants might not be very high this time, so I hope you can help by liking, following, and sharing! Thank you so much!! A Small Note Regarding Gradients As you can see, the hair texture of our model is excellent and looks just like hand-drawn art by an artist. All the highlights are separate layers, ensuring they move naturally with the hair and preventing any separation issues. The rigging of the entire model is also very natural and dynamic. ✨ However, precisely because of this high quality, the Live2D production software is struggling to handle the load and is almost at its limit. Therefore, we regret to inform you that we cannot add gradients to the model's base mesh. 🙅‍♂️ You will only be able to achieve a gradient effect by modifying the textures. To help with this, we will provide a gradient hair texture for easy modification. If you have more specific needs, you will need to modify the textures yourself or ask another artist to do it for you. Currently, we are working on adding easy-to-adjust gradients to the "Head-Only" model, but it is still in the testing phase and might be included in the second wave of updates. Ahhhhh, we really hope the Live2D team optimizes this stuff ASAP!!! 😫 We really want to add everything into the model, but the truth is, we're just out of space... so sad. When Live2D gets optimized, we can finally add gradients directly into the model itself! #Live2D #vtuber #LizMetaCustomizableVtuber
1:04

Hello everyone! 👋 I'm the Assistant of Chaliz! 🛠️ The second episode of recreating your OCs with LizMeta is here! 🎉 To those whose OCs weren't selected this time, it might be for various reasons, for example: We thought your OC required texture modifications. The specific features involved are still under development. We simply ran out of time and could only do this many customizations! etc... Please note that this doesn't mean your OC can't be created at all! Everyone's OCs are just too adorable! 🥰 I'm a bit scared I didn't do them justice, boo-hoo... 😭 (Haha, actually, I tried my best! I looked at your reference pictures, avatars, and even checked your profiles to guess your favorite styles... 👀) In fact, you can probably tell that the model has a huge potential range. If the OC I created looks different from your style, ahhhhh, please don't doubt the model!! 😱 It might just be a problem with my customization skills: ( A huge thank you to everyone who sent in pictures for our testing! Don't be discouraged if you weren't chosen this time! There is one last chance! For the final round, I will show the entire customization process, and the number of participants might not be very high this time, so I hope you can help by liking, following, and sharing! Thank you so much!! A Small Note Regarding Gradients As you can see, the hair texture of our model is excellent and looks just like hand-drawn art by an artist. All the highlights are separate layers, ensuring they move naturally with the hair and preventing any separation issues. The rigging of the entire model is also very natural and dynamic. ✨ However, precisely because of this high quality, the Live2D production software is struggling to handle the load and is almost at its limit. Therefore, we regret to inform you that we cannot add gradients to the model's base mesh. 🙅‍♂️ You will only be able to achieve a gradient effect by modifying the textures. To help with this, we will provide a gradient hair texture for easy modification. If you have more specific needs, you will need to modify the textures yourself or ask another artist to do it for you. Currently, we are working on adding easy-to-adjust gradients to the "Head-Only" model, but it is still in the testing phase and might be included in the second wave of updates. Ahhhhh, we really hope the Live2D team optimizes this stuff ASAP!!! 😫 We really want to add everything into the model, but the truth is, we're just out of space... so sad. When Live2D gets optimized, we can finally add gradients directly into the model itself! #Live2D #vtuber #LizMetaCustomizableVtuber

茶粒子kokoda

47,518 просмотров • 8 месяцев назад

🚨 SIGGRAPH Asia 2025 Paper Alert 🚨 ➡️Paper Title: WorldExplorer: Towards Generating Fully Navigable 3D Scenes 🌟Few pointers from the paper 🎯Generating 3D worlds from text is a highly anticipated goal in computer vision. Existing works are limited by the degree of exploration they allow inside of a scene, i.e., produce stretched-out and noisy artifacts when moving beyond central or panoramic perspectives. 🎯 To this end, authors of this paper proposed “WorldExplorer”, a novel method based on autoregressive video trajectory generation, which builds fully navigable 3D scenes with consistent visual quality across a wide range of viewpoints. 🎯They initialize their scenes by creating multi-view consistent images corresponding to a 360 degree panorama. 🎯Then, they expanded it by leveraging video diffusion models in an iterative scene generation pipeline. 🎯Concretely, they generated multiple videos along short, pre-defined trajectories, that explore the scene in depth, including motion around objects. 🎯Their novel scene memory conditions each video on the most relevant prior views, while a collision-detection mechanism prevents degenerate results, like moving into objects. 🎯Finally,they fuse all generated views into a unified 3D representation via 3D Gaussian Splatting optimization. 🎯Compared to prior approaches, WorldExplorer produces high-quality scenes that remain stable under large camera motion, enabling for the first time realistic and unrestricted exploration. 🎯They believe this marks a significant step toward generating immersive and truly explorable virtual 3D environments. 🏢Organization: TU München 🧙Paper Authors: Manuel-Andreas Schneider, Lukas Höllein , Matthias Niessner 📝 Read the Full Paper here: 🗂️ Project Page: 🧑‍💻 Code: 🎥 Be sure to watch the attached Technical Summary Video - Sound on 🔊🔊 Find this Valuable 💎 ? ♻️QT and teach your network something new Follow me 👣, naveen manwani , for the latest updates on Tech and AI-related news, insightful research papers, and exciting announcements. #SIGGRAPHAsia2025
3:36

🚨 SIGGRAPH Asia 2025 Paper Alert 🚨 ➡️Paper Title: WorldExplorer: Towards Generating Fully Navigable 3D Scenes 🌟Few pointers from the paper 🎯Generating 3D worlds from text is a highly anticipated goal in computer vision. Existing works are limited by the degree of exploration they allow inside of a scene, i.e., produce stretched-out and noisy artifacts when moving beyond central or panoramic perspectives. 🎯 To this end, authors of this paper proposed “WorldExplorer”, a novel method based on autoregressive video trajectory generation, which builds fully navigable 3D scenes with consistent visual quality across a wide range of viewpoints. 🎯They initialize their scenes by creating multi-view consistent images corresponding to a 360 degree panorama. 🎯Then, they expanded it by leveraging video diffusion models in an iterative scene generation pipeline. 🎯Concretely, they generated multiple videos along short, pre-defined trajectories, that explore the scene in depth, including motion around objects. 🎯Their novel scene memory conditions each video on the most relevant prior views, while a collision-detection mechanism prevents degenerate results, like moving into objects. 🎯Finally,they fuse all generated views into a unified 3D representation via 3D Gaussian Splatting optimization. 🎯Compared to prior approaches, WorldExplorer produces high-quality scenes that remain stable under large camera motion, enabling for the first time realistic and unrestricted exploration. 🎯They believe this marks a significant step toward generating immersive and truly explorable virtual 3D environments. 🏢Organization: TU München 🧙Paper Authors: Manuel-Andreas Schneider, Lukas Höllein , Matthias Niessner 📝 Read the Full Paper here: 🗂️ Project Page: 🧑‍💻 Code: 🎥 Be sure to watch the attached Technical Summary Video - Sound on 🔊🔊 Find this Valuable 💎 ? ♻️QT and teach your network something new Follow me 👣, naveen manwani , for the latest updates on Tech and AI-related news, insightful research papers, and exciting announcements. #SIGGRAPHAsia2025

naveen manwani

10,578 просмотров • 11 месяцев назад

We live in an increasingly fractured, polarised world where principles and values we once took for granted in Australia are being undermined. Today the Albanese Labor government gets away with doing things that would simply not have been acceptable in Australia a generation or two ago. Labor may have won the 2025 election with only 35% of the primary vote, but that hasn’t made the problems Labor created go away. We still have a cost-of-living crisis. We still have a housing crisis. We still have record immigration, increasing debt, huge budget deficits and declining economic productivity. The Prime Minister’s so-called productivity roundtable of his sycophants won’t change things at all unless it includes people who understand productivity and business: people like Gina Rhinehart, Solomon Lew, Gerry Harvey and Dick Smith. Taxpayers are still forced to foot the bill for subsidising renewable projects in the pursuit of net zero, coupled with rises in our electricity bills of more than 300% over the past 20 years or so. And more recently, Labor has gambled with our national security. For decades, Australia’s defence has mainly been funded by American taxpayers. America’s nuclear deterrence has prevented global war between major powers, America supplies us with our military platforms and our advanced weapons, and under AUKUS America will hand over to Australia the most closely-guarded military technology on the planet: nuclear propulsion. Yet our current Labor government is publicly appearing less than grateful for this generous support from the ‘arsenal of democracy’. Labor is ghosting our most important ally – the United States – and to all appearances is sucking up to communist China, Australia’s most potent adversary. China routinely commits acts of aggression against Australia as if it’s entitled to do so, and is not hiding its intentions to militarily and economically dominate our region. America simply wants the countries with which it shares close alliances and democratic values to step up and do some of the heavy lifting by increasing their defence spending. It’s called growing a back bone and taking responsibility to defend yourself instead of relying on Americans to fund your battles with their money and lives. The Prime Minister is gambling that he won’t have to, and that is a big mistake. It’s a purely political mistake based on arrogance. It’s not principled. Labor hates everything about Trump and Republicans. Labor has failed to appreciate the importance of America to our security, and that our relationship with the world’s most powerful nation must endure beyond the current occupants of the Lodge and the White House. It’s impossible to imagine Bob Hawke – Labor’s most successful and longest-serving Prime Minister – supporting this. He was a noted champion of the US alliance despite the very conservative Reagan and Bush administrations being in the White House. Hawke was also a vocal critic of communist China, especially after the Tiananmen Square massacre in 1989. Hawke understood Australian principles. Prime Minister Albanese never got the memo, and now our relations with America are declining when we can least afford it. This is not in our best interests. Australia needs a statement of binding principles to govern how we conduct ourselves on the world stage, and to ensure petty politics does not risk important international relationships that must last well beyond every three-year Parliamentary term. This statement needs to say who we are, what Australia stands for, and who we stand with: · democracies like the US and Israel, not communist dictatorships or terrorist regimes; · we stand for the rule of law, self-determination and freedom; and · international peace and order, reinforced by a strong military to deter those who – like China – would disrupt that peace and order. These are One Nation’s principles. They should be Australia’s principles. They used to be. They were principles that didn’t even need stating a generation ago, but Australia is heading down a path under Labor where we cannot be certain of them any longer.
4:33

We live in an increasingly fractured, polarised world where principles and values we once took for granted in Australia are being undermined. Today the Albanese Labor government gets away with doing things that would simply not have been acceptable in Australia a generation or two ago. Labor may have won the 2025 election with only 35% of the primary vote, but that hasn’t made the problems Labor created go away. We still have a cost-of-living crisis. We still have a housing crisis. We still have record immigration, increasing debt, huge budget deficits and declining economic productivity. The Prime Minister’s so-called productivity roundtable of his sycophants won’t change things at all unless it includes people who understand productivity and business: people like Gina Rhinehart, Solomon Lew, Gerry Harvey and Dick Smith. Taxpayers are still forced to foot the bill for subsidising renewable projects in the pursuit of net zero, coupled with rises in our electricity bills of more than 300% over the past 20 years or so. And more recently, Labor has gambled with our national security. For decades, Australia’s defence has mainly been funded by American taxpayers. America’s nuclear deterrence has prevented global war between major powers, America supplies us with our military platforms and our advanced weapons, and under AUKUS America will hand over to Australia the most closely-guarded military technology on the planet: nuclear propulsion. Yet our current Labor government is publicly appearing less than grateful for this generous support from the ‘arsenal of democracy’. Labor is ghosting our most important ally – the United States – and to all appearances is sucking up to communist China, Australia’s most potent adversary. China routinely commits acts of aggression against Australia as if it’s entitled to do so, and is not hiding its intentions to militarily and economically dominate our region. America simply wants the countries with which it shares close alliances and democratic values to step up and do some of the heavy lifting by increasing their defence spending. It’s called growing a back bone and taking responsibility to defend yourself instead of relying on Americans to fund your battles with their money and lives. The Prime Minister is gambling that he won’t have to, and that is a big mistake. It’s a purely political mistake based on arrogance. It’s not principled. Labor hates everything about Trump and Republicans. Labor has failed to appreciate the importance of America to our security, and that our relationship with the world’s most powerful nation must endure beyond the current occupants of the Lodge and the White House. It’s impossible to imagine Bob Hawke – Labor’s most successful and longest-serving Prime Minister – supporting this. He was a noted champion of the US alliance despite the very conservative Reagan and Bush administrations being in the White House. Hawke was also a vocal critic of communist China, especially after the Tiananmen Square massacre in 1989. Hawke understood Australian principles. Prime Minister Albanese never got the memo, and now our relations with America are declining when we can least afford it. This is not in our best interests. Australia needs a statement of binding principles to govern how we conduct ourselves on the world stage, and to ensure petty politics does not risk important international relationships that must last well beyond every three-year Parliamentary term. This statement needs to say who we are, what Australia stands for, and who we stand with: · democracies like the US and Israel, not communist dictatorships or terrorist regimes; · we stand for the rule of law, self-determination and freedom; and · international peace and order, reinforced by a strong military to deter those who – like China – would disrupt that peace and order. These are One Nation’s principles. They should be Australia’s principles. They used to be. They were principles that didn’t even need stating a generation ago, but Australia is heading down a path under Labor where we cannot be certain of them any longer.

Pauline Hanson 🇦🇺

52,109 просмотров • 1 год назад