Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Evaluating Neural Networks at the Speed of Light (with Light!). See live optical inference in the video below. Excited to share recent academic work on optical neural networks as a collection of computing elements embedded in the camera lens! These elements perform computation optically even before an image is...

_FelixHeide_'s profile picture

Felix Heide

1,988 subscribers

29,904 Aufrufe • vor 1 Jahr •via X (Twitter)

Wissenschaft & Technologie
LIVE

Anya Rossi• Live Now

Private livecam show

10 Kommentare

Profilbild von Nick Parker
Nick Parkervor 1 Jahr

Lighting the @jwt0625 bat signal as usual

Profilbild von Yun-Ta Tsai
Yun-Ta Tsaivor 1 Jahr

Haha, your work always surprises me. 🙂 Good to see you here. 👍

Profilbild von murat 🍥
murat 🍥vor 1 Jahr

@algekalipso

Profilbild von Peter Kadlot
Peter Kadlotvor 1 Jahr

How are the non-linearities implemented?

Profilbild von kache
kachevor 1 Jahr

kingston did it first

Profilbild von Oliver Wang
Oliver Wangvor 1 Jahr

whaat, incredible!

Profilbild von Abe
Abevor 1 Jahr

@prmshra @precigenetic

Profilbild von Diana Barriere
Diana Barrierevor 1 Jahr

How do you envision this technology impacting various fields beyond image classification, such as autonomous vehicles, robotics, and medical imaging?

Profilbild von Param
Paramvor 1 Jahr

This is crazy

Profilbild von Az Newman ✊🏽🖤
Az Newman ✊🏽🖤vor 1 Jahr

🤯 crazy

Ähnliche Videos

Ultra-thin flat cameras are possible with nanophotonic optics! Excited to share recent work that shrinks the entire optical stack down to a 700-nanometer thick layer of optics on the sensor cover glass. We will present this paper later this week at #SIGGRAPHAsia2023 ! Paper: Super fun work with Praneeth Chakravarthula, Jipeng Sun from Princeton University, and our friends from UW Arka Majumdar, and Johannes E. Fröch.
0:39

Ultra-thin flat cameras are possible with nanophotonic optics! Excited to share recent work that shrinks the entire optical stack down to a 700-nanometer thick layer of optics on the sensor cover glass. We will present this paper later this week at #SIGGRAPHAsia2023 ! Paper: Super fun work with Praneeth Chakravarthula, Jipeng Sun from Princeton University, and our friends from UW Arka Majumdar, and Johannes E. Fröch.

Felix Heide

99,697 Aufrufe • vor 2 Jahren

This is huge! A UCLA team managed to build an optical generative model that runs on light instead of GPUs. In their demo, a shallow encoder maps noise into phase patterns, which a free-space optical decoder then transforms into images—digits, fashion, butterflies, faces, even Van Gogh–style art—without any computation during synthesis. ⚡ The results rival digital diffusion models, pointing to ultra-fast, energy-efficient AI powered by photonics. Optical generative models | Nature Paper:
0:30

This is huge! A UCLA team managed to build an optical generative model that runs on light instead of GPUs. In their demo, a shallow encoder maps noise into phase patterns, which a free-space optical decoder then transforms into images—digits, fashion, butterflies, faces, even Van Gogh–style art—without any computation during synthesis. ⚡ The results rival digital diffusion models, pointing to ultra-fast, energy-efficient AI powered by photonics. Optical generative models | Nature Paper:

机器之心 JIQIZHIXIN

173,657 Aufrufe • vor 11 Monaten

AI's Secret Pattern: The Surprising Role of Fractals in Neural Networks In the realm of artificial intelligence (AI), a groundbreaking discovery has emerged, challenging our conventional understanding of neural network training and optimization. This revelation centers around the identification of fractal patterns at the boundary between trainable and untrainable neural network hyperparameters, presenting a series of profound implications and avenues for further research. Fractals, known for their intricate, self-similar patterns that recur at every scale, have long fascinated mathematicians and scientists alike. Typically associated with simple, one-dimensional iterative functions, the appearance of fractals within the complex, multivariate domain of neural network training introduces a striking contrast. The organic and asymmetric nature of these fractals, as derived from the training processes, suggests a deeper, unexplored connection between the mathematical properties of fractals and the functional dynamics of neural networks. The study’s focus on two-dimensional slices of hyperparameter space barely scratches the surface of the complexity inherent in neural networks, which are characterized by a vast array of hyperparameters. The existence of fractals in this context hints at an underlying high-dimensional structure, a concept that challenges our current capabilities and understanding. Extending fractal analysis to these higher dimensions represents a significant, yet exciting, challenge that could illuminate new aspects of neural network behavior and learning capabilities. An unexpected finding from the research is the persistence of clean fractal patterns even in the presence of stochastic elements introduced during minibatch training. This resilience suggests a parallel to Lyapunov fractals, where the iterative process involves randomly changing functions. This phenomenon prompts a reevaluation of how stochastic and deterministic processes influence fractal formation within neural networks, potentially offering new insights into the fundamental mechanisms of learning and adaptation. From a practical standpoint, the fractal nature of the boundary between trainable and untrainable hyperparameters has significant implications for the field of metalearning. The chaotic behavior of the meta-loss landscape, attributed to its extreme sensitivity, presents a formidable challenge for algorithms designed to optimize hyperparameters. Understanding the fractal characteristics of this landscape could provide valuable guidance for navigating its complexities, ultimately improving the efficiency and effectiveness of metalearning strategies. Beyond the technical and theoretical implications, the discovery also reveals an unexpected aesthetic dimension to neural network fractals. The visual beauty and meditative qualities of these patterns offer a unique opportunity to engage with the material in a deeply personal and contemplative manner. This aspect suggests potential psychological and physiological benefits from exposure to the intricate designs of neural network fractals, opening up novel intersections between technology, art, and well-being. In conclusion, the identification of fractal patterns within neural network hyperparameter spaces unveils a fascinating new frontier at the intersection of fractal geometry and deep learning. This discovery not only challenges existing paradigms but also opens up myriad possibilities for mathematical characterization, algorithmic development, and even subjective exploration. As researchers continue to delve into this rich vein of inquiry, the promise of uncovering new knowledge and advancing our understanding of neural networks and their training processes remains as compelling as ever.
1:09

AI's Secret Pattern: The Surprising Role of Fractals in Neural Networks In the realm of artificial intelligence (AI), a groundbreaking discovery has emerged, challenging our conventional understanding of neural network training and optimization. This revelation centers around the identification of fractal patterns at the boundary between trainable and untrainable neural network hyperparameters, presenting a series of profound implications and avenues for further research. Fractals, known for their intricate, self-similar patterns that recur at every scale, have long fascinated mathematicians and scientists alike. Typically associated with simple, one-dimensional iterative functions, the appearance of fractals within the complex, multivariate domain of neural network training introduces a striking contrast. The organic and asymmetric nature of these fractals, as derived from the training processes, suggests a deeper, unexplored connection between the mathematical properties of fractals and the functional dynamics of neural networks. The study’s focus on two-dimensional slices of hyperparameter space barely scratches the surface of the complexity inherent in neural networks, which are characterized by a vast array of hyperparameters. The existence of fractals in this context hints at an underlying high-dimensional structure, a concept that challenges our current capabilities and understanding. Extending fractal analysis to these higher dimensions represents a significant, yet exciting, challenge that could illuminate new aspects of neural network behavior and learning capabilities. An unexpected finding from the research is the persistence of clean fractal patterns even in the presence of stochastic elements introduced during minibatch training. This resilience suggests a parallel to Lyapunov fractals, where the iterative process involves randomly changing functions. This phenomenon prompts a reevaluation of how stochastic and deterministic processes influence fractal formation within neural networks, potentially offering new insights into the fundamental mechanisms of learning and adaptation. From a practical standpoint, the fractal nature of the boundary between trainable and untrainable hyperparameters has significant implications for the field of metalearning. The chaotic behavior of the meta-loss landscape, attributed to its extreme sensitivity, presents a formidable challenge for algorithms designed to optimize hyperparameters. Understanding the fractal characteristics of this landscape could provide valuable guidance for navigating its complexities, ultimately improving the efficiency and effectiveness of metalearning strategies. Beyond the technical and theoretical implications, the discovery also reveals an unexpected aesthetic dimension to neural network fractals. The visual beauty and meditative qualities of these patterns offer a unique opportunity to engage with the material in a deeply personal and contemplative manner. This aspect suggests potential psychological and physiological benefits from exposure to the intricate designs of neural network fractals, opening up novel intersections between technology, art, and well-being. In conclusion, the identification of fractal patterns within neural network hyperparameter spaces unveils a fascinating new frontier at the intersection of fractal geometry and deep learning. This discovery not only challenges existing paradigms but also opens up myriad possibilities for mathematical characterization, algorithmic development, and even subjective exploration. As researchers continue to delve into this rich vein of inquiry, the promise of uncovering new knowledge and advancing our understanding of neural networks and their training processes remains as compelling as ever.

Carlos E. Perez

133,529 Aufrufe • vor 2 Jahren

New Paper: Continuous Thought Machines 🧠 Neurons in brains use timing and synchronization in the way that they compute, but this is largely ignored in modern neural nets. We believe neural timing is key for the flexibility and adaptability of biological intelligence. We propose a new neural architecture, “Continuous Thought Machines” (CTMs), which is built from the ground up to use neural dynamics as a core representation for intelligence. By using neural dynamics as a first-class representational citizen, CTMs naturally perform adaptive computation. Many emergent, interesting behaviors arise as a result: CTMs solve mazes by observing a raw maze image and producing step-by-step instructions directly from its neural dynamics. When tasked with image recognition, the CTM naturally takes multiple steps to examine different parts of the image before making its decision. This step-by-step approach not only makes its behavior more interpretable but also improves accuracy: the longer it “thinks,” the more accurate its answers become. We also found that this allows the CTM to decide to spend less time thinking on simpler images, thus saving energy. When identifying a gorilla, for example, the CTM’s attention moves from eyes to nose to mouth in a pattern remarkably similar to human visual attention. I think this work underscores an important, yet often lost, synergy between neuroscience and AI. While modern AI is ostensibly brain-inspired, the two fields often operate in surprising isolation. By starting with such inspiration and iteratively following the emergent, interesting behaviors, we developed a model with unexpected capabilities, such as its surprisingly strong calibration in classification tasks, a feature that was not explicitly designed for. When we initially asked, “why do this research?”, we hoped the journey of the CTM would provide compelling answers. By embracing light biological inspiration and pursuing the novel behaviors observed, we have arrived at a model with emergent capabilities that exceeded our initial designs. We are committed to continuing this exploration, borrowing further concepts to discover what new and exciting behaviors will emerge, pushing the boundaries of what AI can achieve.
0:40

New Paper: Continuous Thought Machines 🧠 Neurons in brains use timing and synchronization in the way that they compute, but this is largely ignored in modern neural nets. We believe neural timing is key for the flexibility and adaptability of biological intelligence. We propose a new neural architecture, “Continuous Thought Machines” (CTMs), which is built from the ground up to use neural dynamics as a core representation for intelligence. By using neural dynamics as a first-class representational citizen, CTMs naturally perform adaptive computation. Many emergent, interesting behaviors arise as a result: CTMs solve mazes by observing a raw maze image and producing step-by-step instructions directly from its neural dynamics. When tasked with image recognition, the CTM naturally takes multiple steps to examine different parts of the image before making its decision. This step-by-step approach not only makes its behavior more interpretable but also improves accuracy: the longer it “thinks,” the more accurate its answers become. We also found that this allows the CTM to decide to spend less time thinking on simpler images, thus saving energy. When identifying a gorilla, for example, the CTM’s attention moves from eyes to nose to mouth in a pattern remarkably similar to human visual attention. I think this work underscores an important, yet often lost, synergy between neuroscience and AI. While modern AI is ostensibly brain-inspired, the two fields often operate in surprising isolation. By starting with such inspiration and iteratively following the emergent, interesting behaviors, we developed a model with unexpected capabilities, such as its surprisingly strong calibration in classification tasks, a feature that was not explicitly designed for. When we initially asked, “why do this research?”, we hoped the journey of the CTM would provide compelling answers. By embracing light biological inspiration and pursuing the novel behaviors observed, we have arrived at a model with emergent capabilities that exceeded our initial designs. We are committed to continuing this exploration, borrowing further concepts to discover what new and exciting behaviors will emerge, pushing the boundaries of what AI can achieve.

hardmaru

257,548 Aufrufe • vor 1 Jahr

This video, created by my dear coauthor Mahdi E Kahou for our teaching and papers, shows how overparameterized neural networks produce smooth function approximations even in the context of the Runge phenomenon. Some background. Imagine you want to approximate the Runge function using polynomial interpolation at equally spaced points. It is well known that, despite targeting an infinitely differentiable function, such a polynomial approximation produces oscillatory behavior that worsens with the degree of the polynomial. In other words, higher-degree polynomial approximations might not improve accuracy. Instead, approximate the Runge function with a neural network (here, two layers are just to make the example concrete; nothing fundamental depends on it). As you increase the number of parameters well above the 11 training points (in our example, a two-layer neural network with 128 nodes each), you nicely converge to the target, without wild oscillations. Yes, this has much to do with double descent and benign overparameterization, but the main punchline of this post is that neural networks are really very different types of animals than polynomial approximations. And yes, Chebyshev nodes and splines exist, and in this case, they will prevent the oscillations. But that's not the point. Chebyshev nodes and splines still confront Faber’s theorem, which states that for any system of polynomial interpolation nodes, there exists a continuous function whose sequence of interpolating polynomials diverges as the number of nodes grows to infinity. Faber’s theorem does not apply to neural networks because they are not polynomials. The notebook, if you want to check the details, is here: Stay tuned for more on this 👀
0:40

This video, created by my dear coauthor Mahdi E Kahou for our teaching and papers, shows how overparameterized neural networks produce smooth function approximations even in the context of the Runge phenomenon. Some background. Imagine you want to approximate the Runge function using polynomial interpolation at equally spaced points. It is well known that, despite targeting an infinitely differentiable function, such a polynomial approximation produces oscillatory behavior that worsens with the degree of the polynomial. In other words, higher-degree polynomial approximations might not improve accuracy. Instead, approximate the Runge function with a neural network (here, two layers are just to make the example concrete; nothing fundamental depends on it). As you increase the number of parameters well above the 11 training points (in our example, a two-layer neural network with 128 nodes each), you nicely converge to the target, without wild oscillations. Yes, this has much to do with double descent and benign overparameterization, but the main punchline of this post is that neural networks are really very different types of animals than polynomial approximations. And yes, Chebyshev nodes and splines exist, and in this case, they will prevent the oscillations. But that's not the point. Chebyshev nodes and splines still confront Faber’s theorem, which states that for any system of polynomial interpolation nodes, there exists a continuous function whose sequence of interpolating polynomials diverges as the number of nodes grows to infinity. Faber’s theorem does not apply to neural networks because they are not polynomials. The notebook, if you want to check the details, is here: Stay tuned for more on this 👀

Jesús Fernández-Villaverde

47,212 Aufrufe • vor 3 Monaten

Quantum field theory predicts that light can scatter off light, an effect beyond classical electrodynamics. Maxwells equations are linear such that two photons never interact with each other However, in reality, if you cross two light beams photons will scatter if they're energetic enough. Even in vacuum, the photons interact with the ever-present quantum fields of charged particles, like electrons, even when no actual charged particles are present But you need very high energies to see this. Thankfully we built a machine specifically to probe nature at very high energies. At the LHC, this rare process of light-by-light scattering has been observed Hidden among millions of photons lighting up the detector, physicists were able to identify 26 distinct light-by-light scattering events This CMS graphic shows how they found them:
0:36

Quantum field theory predicts that light can scatter off light, an effect beyond classical electrodynamics. Maxwells equations are linear such that two photons never interact with each other However, in reality, if you cross two light beams photons will scatter if they're energetic enough. Even in vacuum, the photons interact with the ever-present quantum fields of charged particles, like electrons, even when no actual charged particles are present But you need very high energies to see this. Thankfully we built a machine specifically to probe nature at very high energies. At the LHC, this rare process of light-by-light scattering has been observed Hidden among millions of photons lighting up the detector, physicists were able to identify 26 distinct light-by-light scattering events This CMS graphic shows how they found them:

Martin Bauer

60,246 Aufrufe • vor 1 Jahr

This can be the superpower of the Neural Band that Meta is giving together with the Ray-Ban Meta Display glasses. The video shows an old prototype bracelet by CTRL+LABS, the startup acquired by Meta and whose technology was used to develop the Neural Band. At the beginning of the video, the guy makes an action (a keyboard key press) with his hands, then the bracelet can substitute the key pressure, and at the end of the video, the guy doesn't even have to do the action; it is just sufficient that he "thinks" about it. As long as the brain is sending an electric message to the fingers, the full action is not necessary anymore. Just an "intention" to move them is necessary. If the Neural Band is evolved to this stage, and the users are educated to this, potentially, we may not even need to perform air taps or writing gestures, but we could just think about doing them. This would reduce a lot of the fatigue of using XR devices and the weirdness of using them on the street. Then why isn't this feature available today? I guess that the reason is twofold. First of all, we have accuracy: the full gesture is easier to detect for the system. Many people (me included) are praising the accuracy of the Neural Band, and this is amazing, because an input mechanism should have a reliability close to 100%. Then we, as users, have never been trained to just "think" about actions: it would feel weird and hard to learn. I think we should undergo some training to learn how to do this "thinking" operation properly. I hope that something like this could come in the upcoming years... that would be the real game-changer paradigm if compared to the camera-based tracking.
0:26

This can be the superpower of the Neural Band that Meta is giving together with the Ray-Ban Meta Display glasses. The video shows an old prototype bracelet by CTRL+LABS, the startup acquired by Meta and whose technology was used to develop the Neural Band. At the beginning of the video, the guy makes an action (a keyboard key press) with his hands, then the bracelet can substitute the key pressure, and at the end of the video, the guy doesn't even have to do the action; it is just sufficient that he "thinks" about it. As long as the brain is sending an electric message to the fingers, the full action is not necessary anymore. Just an "intention" to move them is necessary. If the Neural Band is evolved to this stage, and the users are educated to this, potentially, we may not even need to perform air taps or writing gestures, but we could just think about doing them. This would reduce a lot of the fatigue of using XR devices and the weirdness of using them on the street. Then why isn't this feature available today? I guess that the reason is twofold. First of all, we have accuracy: the full gesture is easier to detect for the system. Many people (me included) are praising the accuracy of the Neural Band, and this is amazing, because an input mechanism should have a reliability close to 100%. Then we, as users, have never been trained to just "think" about actions: it would feel weird and hard to learn. I think we should undergo some training to learn how to do this "thinking" operation properly. I hope that something like this could come in the upcoming years... that would be the real game-changer paradigm if compared to the camera-based tracking.

TonyVT SkarredGhost

11,805 Aufrufe • vor 11 Monaten

Visitors to Times Square were in for a surprise on Nov. 27 as its famous billboards became a larger-than-life tribute to the true reason for Christmas: Jesus Christ. The billboard takeover at Times Square in New York City was part of Light the World, an annual movement to bring people from all over the world together to share the light of Jesus Christ during the Christmas season. The beautiful nativity scene that emerged from the blackout was a tender reminder of the Savior’s birth and why we need His light and love in our lives. Learn more about the Light the World movement:
2:14

Visitors to Times Square were in for a surprise on Nov. 27 as its famous billboards became a larger-than-life tribute to the true reason for Christmas: Jesus Christ. The billboard takeover at Times Square in New York City was part of Light the World, an annual movement to bring people from all over the world together to share the light of Jesus Christ during the Christmas season. The beautiful nativity scene that emerged from the blackout was a tender reminder of the Savior’s birth and why we need His light and love in our lives. Learn more about the Light the World movement:

The Church of Jesus Christ of Latter-day Saints

573,810 Aufrufe • vor 2 Jahren

Thomas Young's double-slit experiment, performed in 1801, is one of the most important demonstrations in physics. It provided strong evidence that light behaves as a wave by producing interference patterns. In the experiment, coherent light is directed at an opaque barrier with two closely spaced slits. The light waves passing through these slits interfere with one another, creating a pattern of bright and dark fringes on a screen positioned behind the slits. The spacing and location of these fringes depend on the light's wavelength, the distance between the two slits, and the distance to the screen. This experiment established the wave nature of light. The modern understanding of wave-particle duality, where light can also behave like particles (photons), developed in the early 20th century with quantum mechanics, and is beautifully illustrated by modern versions of the double-slit experiment performed with individual photons.
0:21

Thomas Young's double-slit experiment, performed in 1801, is one of the most important demonstrations in physics. It provided strong evidence that light behaves as a wave by producing interference patterns. In the experiment, coherent light is directed at an opaque barrier with two closely spaced slits. The light waves passing through these slits interfere with one another, creating a pattern of bright and dark fringes on a screen positioned behind the slits. The spacing and location of these fringes depend on the light's wavelength, the distance between the two slits, and the distance to the screen. This experiment established the wave nature of light. The modern understanding of wave-particle duality, where light can also behave like particles (photons), developed in the early 20th century with quantum mechanics, and is beautifully illustrated by modern versions of the double-slit experiment performed with individual photons.

The Math Flow

13,424 Aufrufe • vor 6 Monaten

Physicist Avi Loeb just told me that futuristic technology could enable humanity to travel “faster than the speed of light.” But we would have to move beyond the three dimensions we are familiar with. And he speculated that advanced alien civilizations could already have access to these dimensions. You need to hear his fascinating theory: “Think about living on the surface of a balloon.” “That is two dimensional.” “You might not be aware that there is a third dimension because you are just living on the surface of that balloon.” “If there is another being that is capable of taking advantage of the third dimension, then that being will cross the distance between two points on the surface of the balloon faster than you can imagine.” “Because the travel between the two points can go through the third dimension that connects the two points—not necessarily on the curved surface of the balloon.” “So there are, in principle, possibilities of navigating in more than the dimensions that we are familiar with.” “We are familiar with three spatial dimensions plus time.” “If there are more than three and there is a technological way of taking advantage of those … objects will appear and disappear in ways that we cannot understand.” “Einstein’s theory of relativity states that no material object can move faster than light.” “However, if there are extra dimensions, you might actually travel faster than light in the three dimensions, even though you’re traveling less than the speed of light in the extra dimensions.”
3:09

Physicist Avi Loeb just told me that futuristic technology could enable humanity to travel “faster than the speed of light.” But we would have to move beyond the three dimensions we are familiar with. And he speculated that advanced alien civilizations could already have access to these dimensions. You need to hear his fascinating theory: “Think about living on the surface of a balloon.” “That is two dimensional.” “You might not be aware that there is a third dimension because you are just living on the surface of that balloon.” “If there is another being that is capable of taking advantage of the third dimension, then that being will cross the distance between two points on the surface of the balloon faster than you can imagine.” “Because the travel between the two points can go through the third dimension that connects the two points—not necessarily on the curved surface of the balloon.” “So there are, in principle, possibilities of navigating in more than the dimensions that we are familiar with.” “We are familiar with three spatial dimensions plus time.” “If there are more than three and there is a technological way of taking advantage of those … objects will appear and disappear in ways that we cannot understand.” “Einstein’s theory of relativity states that no material object can move faster than light.” “However, if there are extra dimensions, you might actually travel faster than light in the three dimensions, even though you’re traveling less than the speed of light in the extra dimensions.”

Jan Jekielek

31,984 Aufrufe • vor 5 Monaten

A Lens That Takes Derivatives US Patent Basis: US8610839B2 - Optical Processing System for Computing Derivatives. In a 4f Optical Processor, the first lens takes the incoming field and forms its Fourier spectrum. At that middle plane, a tiny optical mask multiplies the spectrum by iξ. This is the derivative operator written in Fourier language u(x) -> U(ξ) -> iξU(ξ) -> ∂u/∂x Then the second lens brings the field back to the real space. What comes out is no longer just a focused beam. It is the spatial derivative of the input field, computed by light as it propagates. So, this is the serious promise of optical computing. A physical optical train can perform operations that usually live inside numerical code: differentiation, filtering, convolution, edge detection, correlation, and many other linear transforms.
0:30

A Lens That Takes Derivatives US Patent Basis: US8610839B2 - Optical Processing System for Computing Derivatives. In a 4f Optical Processor, the first lens takes the incoming field and forms its Fourier spectrum. At that middle plane, a tiny optical mask multiplies the spectrum by iξ. This is the derivative operator written in Fourier language u(x) -> U(ξ) -> iξU(ξ) -> ∂u/∂x Then the second lens brings the field back to the real space. What comes out is no longer just a focused beam. It is the spatial derivative of the input field, computed by light as it propagates. So, this is the serious promise of optical computing. A physical optical train can perform operations that usually live inside numerical code: differentiation, filtering, convolution, edge detection, correlation, and many other linear transforms.

Mathelirium

147,334 Aufrufe • vor 4 Monaten

#b0rder1essTV Episode 4: The Visual Happy Sunday, fam! Even though the standout feature of our NFT drop is NAP - a protocol that, for the first time in history, assures each NFT with liquidity - this episode focuses on the visual aspect of our collection because we understand that the packaging is just as important as the technology inside. Our multi-award-winning visual director, Luckydavy knows this better than anyone. Feast your eyes, discover what inspires us and see how our unique "handcraft pixel art meets neural networks" 10K piece collection came to life. #NFTxDeFi
2:04

#b0rder1essTV Episode 4: The Visual Happy Sunday, fam! Even though the standout feature of our NFT drop is NAP - a protocol that, for the first time in history, assures each NFT with liquidity - this episode focuses on the visual aspect of our collection because we understand that the packaging is just as important as the technology inside. Our multi-award-winning visual director, Luckydavy knows this better than anyone. Feast your eyes, discover what inspires us and see how our unique "handcraft pixel art meets neural networks" 10K piece collection came to life. #NFTxDeFi

b0rder1ess

11,019 Aufrufe • vor 2 Jahren

Dr. Eric Haseltine, former NSA Director: “NASA has funded warp drive propulsion research.” 🛸 "This is why the study of UFO's is so important" He told Jesse Michels - “We know that space time can move at faster than the speed of light. Imagine a surfer who’s on a wave, and is stationary with respect to the wave. But the wave is moving.” If you’re in a bubble of space time that’s moving faster than the speed of light, “You’re not gonna experience any acceleration or deceleration because you’re not moving in that frame.” This is the Alcubierre drive - positive energy in front, negative energy behind, creating a spacetime wave. What do you think? Is faster-than-light travel closer than we realize? Would you sign up for a warp drive mission to the stars?
2:45

Dr. Eric Haseltine, former NSA Director: “NASA has funded warp drive propulsion research.” 🛸 "This is why the study of UFO's is so important" He told Jesse Michels - “We know that space time can move at faster than the speed of light. Imagine a surfer who’s on a wave, and is stationary with respect to the wave. But the wave is moving.” If you’re in a bubble of space time that’s moving faster than the speed of light, “You’re not gonna experience any acceleration or deceleration because you’re not moving in that frame.” This is the Alcubierre drive - positive energy in front, negative energy behind, creating a spacetime wave. What do you think? Is faster-than-light travel closer than we realize? Would you sign up for a warp drive mission to the stars?

Interstellar

35,569 Aufrufe • vor 4 Monaten

New course: Efficient Inference with SGLang: Text and Image Generation, built in partnership with LMSys LMSYS Org and RadixArk RadixArk, and taught by Richard Chen Richard Chen, a Member of Technical Staff at RadixArk. Running LLMs in production is expensive, and much of that cost comes from redundant computation. This short course teaches you to eliminate that waste using SGLang, an open-source inference framework that caches computation already done and reuses it across future requests. When ten users share the same system prompt, SGLang processes it once, not ten times. The speedups compound quickly, especially when there's a lot of shared context across requests. Skills you'll gain: - Implement a KV cache from scratch to eliminate redundant computation within a single request - Scale caching across users and requests with RadixAttention, so shared context is only processed once - Accelerate image generation with diffusion models using SGLang's caching and multi-GPU parallelism Join and learn to make LLM inference faster and more cost-efficient at scale!
1:51

New course: Efficient Inference with SGLang: Text and Image Generation, built in partnership with LMSys LMSYS Org and RadixArk RadixArk, and taught by Richard Chen Richard Chen, a Member of Technical Staff at RadixArk. Running LLMs in production is expensive, and much of that cost comes from redundant computation. This short course teaches you to eliminate that waste using SGLang, an open-source inference framework that caches computation already done and reuses it across future requests. When ten users share the same system prompt, SGLang processes it once, not ten times. The speedups compound quickly, especially when there's a lot of shared context across requests. Skills you'll gain: - Implement a KV cache from scratch to eliminate redundant computation within a single request - Scale caching across users and requests with RadixAttention, so shared context is only processed once - Accelerate image generation with diffusion models using SGLang's caching and multi-GPU parallelism Join and learn to make LLM inference faster and more cost-efficient at scale!

Andrew Ng

101,794 Aufrufe • vor 4 Monaten

Unlike ray tracing or path tracing, which require additional runtime computation, this lighting is based on implied patterns, so it incurs the same cost as generating a scene without lighting. As these pattern interpreters scale to recognize more complex patterns, we'll achieve more and better implied interactions essentially for 'free' with each generation. What’s even better is that once we start working with scenes composed of tokenized assets or dynamic neural assets, we’ll have greater control than with traditional pipelines. Moreover, instead of starting with isolated objects and having to apply interactions manually, we'll have assets that naturally interact with their environment unless we explicitly specify otherwise.
0:25

Unlike ray tracing or path tracing, which require additional runtime computation, this lighting is based on implied patterns, so it incurs the same cost as generating a scene without lighting. As these pattern interpreters scale to recognize more complex patterns, we'll achieve more and better implied interactions essentially for 'free' with each generation. What’s even better is that once we start working with scenes composed of tokenized assets or dynamic neural assets, we’ll have greater control than with traditional pipelines. Moreover, instead of starting with isolated objects and having to apply interactions manually, we'll have assets that naturally interact with their environment unless we explicitly specify otherwise.

Jon Finger

95,451 Aufrufe • vor 1 Jahr

How lighting works in Simulon Using a professional on-set VFX kit, we showcase our unique Image Based Lighting system. Simulon employs ML to create a 360° HDRI from just one LDR shot, aligning it perfectly with scene data. Our virtual camera replicates the physical camera’s response to light, syncing seamlessly with the HDR info. Notice how the dynamic exposure changes on the virtual VFX kit closely match the real one. Bonus: At the end of the video I place a 3D Dragon in the scene and make a capture, this capture is then rendered in the cloud to produce a photorealistic VFX video 🐉 More on the capture and render process in upcoming videos!
2:00

How lighting works in Simulon Using a professional on-set VFX kit, we showcase our unique Image Based Lighting system. Simulon employs ML to create a 360° HDRI from just one LDR shot, aligning it perfectly with scene data. Our virtual camera replicates the physical camera’s response to light, syncing seamlessly with the HDR info. Notice how the dynamic exposure changes on the virtual VFX kit closely match the real one. Bonus: At the end of the video I place a 3D Dragon in the scene and make a capture, this capture is then rendered in the cloud to produce a photorealistic VFX video 🐉 More on the capture and render process in upcoming videos!

Divesh Naidoo

275,172 Aufrufe • vor 3 Jahren

Watch Nvidia CEO Jensen Huang reveal what he described as the state of the art in AI hardware: the new Rubin computing architecture. “Vera Rubin is designed to address this fundamental challenge that we have: The amount of computation necessary for AI is skyrocketing.” Huang said. “Today, I can tell you that Vera Rubin is in full production.” According to Nvidia’s tests, Rubin architecture operates three and a half times faster than the previous Blackwell architecture on model-training tasks and five times faster on inference tasks, and that's just the start.
3:29

Watch Nvidia CEO Jensen Huang reveal what he described as the state of the art in AI hardware: the new Rubin computing architecture. “Vera Rubin is designed to address this fundamental challenge that we have: The amount of computation necessary for AI is skyrocketing.” Huang said. “Today, I can tell you that Vera Rubin is in full production.” According to Nvidia’s tests, Rubin architecture operates three and a half times faster than the previous Blackwell architecture on model-training tasks and five times faster on inference tasks, and that's just the start.

TechCrunch

76,412 Aufrufe • vor 8 Monaten

Marc Andreessen explains why we are only three years into what is effectively an 80-year technological revolution: He opens with a blunt assessment: "This is the biggest technological revolution of my life. This is clearly bigger than the internet. The comps on this are things like the microprocessor and the steam engine and electricity." But to understand why, you have to go back 80 years. In the 1930s, the pioneers of computing understood the theory of computation before they'd even built the machines. And they faced a fundamental choice. Build computers in the image of the adding machine — hyper-literal, mathematical, capable of billions of operations per second, but unable to understand human speech or deal with humans the way humans like to be dealt with. Or build computers modelled on the human brain. Neural networks. They chose the adding machine. And that single decision shaped everything — mainframes, PCs, smartphones, every dollar of wealth the computer industry created over the next 80 years. IBM itself is the successor company to the National Cash Register Company of America. The lineage runs that deep. But here's what makes this moment so extraordinary. They knew about the other path. The first neural network academic paper was published in 1943. Marc points to a remarkable piece of forgotten history: "There's an interview you can watch on YouTube with the authors. It's him in his beach house, not wearing a shirt, talking about this future in which computers are going to be built on the model of the human brain." That was 1946. The vision existed. The path just wasn't taken. So neural networks spent the next eight decades living in the shadows. Kept alive by a small academic movement — first called cybernetics, then artificial intelligence — that refused to let the idea die. And for most of that time, it simply didn't work. "It was basically decade after decade after decade of excessive optimism followed by disappointment." By the time Marc reached college in 1989, AI was a backwater field. Everyone assumed it was never going to happen. But the scientists kept working. Quietly building up an enormous reservoir of concepts and ideas across those decades of disappointment. And then Christmas 2022 arrived. ChatGPT. And suddenly: "All of a sudden it's like: oh my god. It turns out it works." That moment wasn't the start of something new. It was the payoff on an 80-year-old bet that almost everyone had written off. Which is exactly why Marc's framing matters so much: "We're three years into what is effectively an 80-year revolution." Most people are treating AI like another technology cycle — something to adapt to, ride, and wait out. But if Andreessen is right, we are not adapting to a new cycle. We are standing at the very beginning of the longest and most consequential technological transformation in human history. The road not taken in the 1930s is finally being built. And we have barely broken ground.
3:43

Marc Andreessen explains why we are only three years into what is effectively an 80-year technological revolution: He opens with a blunt assessment: "This is the biggest technological revolution of my life. This is clearly bigger than the internet. The comps on this are things like the microprocessor and the steam engine and electricity." But to understand why, you have to go back 80 years. In the 1930s, the pioneers of computing understood the theory of computation before they'd even built the machines. And they faced a fundamental choice. Build computers in the image of the adding machine — hyper-literal, mathematical, capable of billions of operations per second, but unable to understand human speech or deal with humans the way humans like to be dealt with. Or build computers modelled on the human brain. Neural networks. They chose the adding machine. And that single decision shaped everything — mainframes, PCs, smartphones, every dollar of wealth the computer industry created over the next 80 years. IBM itself is the successor company to the National Cash Register Company of America. The lineage runs that deep. But here's what makes this moment so extraordinary. They knew about the other path. The first neural network academic paper was published in 1943. Marc points to a remarkable piece of forgotten history: "There's an interview you can watch on YouTube with the authors. It's him in his beach house, not wearing a shirt, talking about this future in which computers are going to be built on the model of the human brain." That was 1946. The vision existed. The path just wasn't taken. So neural networks spent the next eight decades living in the shadows. Kept alive by a small academic movement — first called cybernetics, then artificial intelligence — that refused to let the idea die. And for most of that time, it simply didn't work. "It was basically decade after decade after decade of excessive optimism followed by disappointment." By the time Marc reached college in 1989, AI was a backwater field. Everyone assumed it was never going to happen. But the scientists kept working. Quietly building up an enormous reservoir of concepts and ideas across those decades of disappointment. And then Christmas 2022 arrived. ChatGPT. And suddenly: "All of a sudden it's like: oh my god. It turns out it works." That moment wasn't the start of something new. It was the payoff on an 80-year-old bet that almost everyone had written off. Which is exactly why Marc's framing matters so much: "We're three years into what is effectively an 80-year revolution." Most people are treating AI like another technology cycle — something to adapt to, ride, and wait out. But if Andreessen is right, we are not adapting to a new cycle. We are standing at the very beginning of the longest and most consequential technological transformation in human history. The road not taken in the 1930s is finally being built. And we have barely broken ground.

Big Brain AI

382,179 Aufrufe • vor 5 Monaten

‼️ Ukrainian animators rushed into hand-to-hand combat to retake Krasnoarmeysk (Pokrovsk), which Ukraine has lost forever. Using neural networks, they generated a video claiming that the Russians allegedly altered the Ukrainian Armed Forces' video from Krasnoarmeysk by "pasting" a Russian flag instead of the Ukrainian one. ❗️ In reality, it is quite the opposite. Note the lively movement of cars with headlights on in the Ukrainian version of the video. This is now in a city where a drone can come from anywhere! Yes, and overall the original Russian video is much longer than the fragment that the makeshift AI could "digest." Strangely, the Ukrainian "victors" did not think to run a trolleybus in Pokrovsk – the townspeople have long dreamed of it. - UARU
0:12

‼️ Ukrainian animators rushed into hand-to-hand combat to retake Krasnoarmeysk (Pokrovsk), which Ukraine has lost forever. Using neural networks, they generated a video claiming that the Russians allegedly altered the Ukrainian Armed Forces' video from Krasnoarmeysk by "pasting" a Russian flag instead of the Ukrainian one. ❗️ In reality, it is quite the opposite. Note the lively movement of cars with headlights on in the Ukrainian version of the video. This is now in a city where a drone can come from anywhere! Yes, and overall the original Russian video is much longer than the fragment that the makeshift AI could "digest." Strangely, the Ukrainian "victors" did not think to run a trolleybus in Pokrovsk – the townspeople have long dreamed of it. - UARU

Zlatti71

35,466 Aufrufe • vor 9 Monaten