Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Φ-SO : Physical Symbolic Optimization - Learning Physics from Data 🧠 The Physical Symbolic Optimization package uses deep reinforcement learning to discover physical laws from data. Here is Φ-SO discovering the analytical expression of a damped harmonic oscillator.

Jousef Murad

19,926 subscribers

433,240 views • 2 years ago •via X (Twitter)

Science & Technology Education

Anya Rossi• Live Now

Private livecam show

9 Comments

Jousef Murad2 years ago

🌎 Repo: 👉 Paper:

Luke Barnes2 years ago

This isn't physics. It's curve fitting. If you want to do physics, start by having the code discover the underlying differential equation. Handy, though.

Rasmus Nordström2 years ago

ELI5 please, is this just brute forcing different formulas to see what fits that data or is it more magic to it?

Moritz Zaiss2 years ago

Nice, I love these approaches! A differentiable physicist that doesn't need coffee. Check out the work of @GMartius in this context!

Jiwoong Lee2 years ago

Would it be more straightforward to optimize in frequency domain? Curious.

Luis Tobon2 years ago

Try Sindy.

Wassim Tenachi2 years ago

Thanks for sharing my work, glad to see people excited about this ! 😊

Unser Kampf : notre combat2 years ago

How many points are tested ? 4k+1 ?

Unser Kampf : notre combat2 years ago

Il like very much. Your works completes mine.

Related Videos

Reinforcement learning is used to speed the production of behavior for the Boston Dynamics Atlas humanoid robot. At the heart of the learning process is a physics-based simulator that generates training data for a variety of maneuvers.

Reinforcement learning is used to speed the production of behavior for the Boston Dynamics Atlas humanoid robot. At the heart of the learning process is a physics-based simulator that generates training data for a variety of maneuvers.

RAI Institute

76,668 views • 1 year ago

Tesla is continuously improving the deep learning neural networks that power our real-world AI. Data labeling is vital to our autonomous systems. We're seeking passionate, detail-oriented minds to train the future of physical AI. Join us.

Tesla is continuously improving the deep learning neural networks that power our real-world AI. Data labeling is vital to our autonomous systems. We're seeking passionate, detail-oriented minds to train the future of physical AI. Join us.

Ryan Donnelly

161,564 views • 1 year ago

My team is hiring in multimodal representation learning. We are working on tactile representations to equip robots with a physical understanding of the world. Bored of scraping disembodied internet data and looking for a new challenge? Apply here:

My team is hiring in multimodal representation learning. We are working on tactile representations to equip robots with a physical understanding of the world. Bored of scraping disembodied internet data and looking for a new challenge? Apply here:

Mustafa Mukadam

56,704 views • 2 years ago

Models have run out of internet to learn from. The next frontier is the physical world. Every week we turn 21,000+ hours of real world activity into labeled, sensor-enriched data, continuous observation of how the physical world actually moves.

Models have run out of internet to learn from. The next frontier is the physical world. Every week we turn 21,000+ hours of real world activity into labeled, sensor-enriched data, continuous observation of how the physical world actually moves.

375ai

10,774 views • 2 months ago

Noble Machines' Moby robot hauling a 50 lb crate. The whole-body control (WBC) is the physical foundation of the learning stack. It absorbs payload variation, contact response, and center-of-mass shifts in real time at the control layer and enabled cleaner data-focused learning

Noble Machines' Moby robot hauling a 50 lb crate. The whole-body control (WBC) is the physical foundation of the learning stack. It absorbs payload variation, contact response, and center-of-mass shifts in real time at the control layer and enabled cleaner data-focused learning

The Humanoid Hub

17,482 views • 3 months ago

☄️ Newton’s Room is a Mixed Reality physics education experience. ⚛️ Users have to apply Newton's laws of physics to solve puzzles that adapt and respond to their physical environment. 🆓 Download the app today for Free on Meta Quest! 🥽 Enjoy learning physics in your room.

☄️ Newton’s Room is a Mixed Reality physics education experience. ⚛️ Users have to apply Newton's laws of physics to solve puzzles that adapt and respond to their physical environment. 🆓 Download the app today for Free on Meta Quest! 🥽 Enjoy learning physics in your room.

Treeview

2,190,001 views • 2 years ago

KL divergence is a fundamental concept used in machine learning, from optimization a neural net to RL training of LLMs. but here is what it actually means:

KL divergence is a fundamental concept used in machine learning, from optimization a neural net to RL training of LLMs. but here is what it actually means:

ℏεsam

20,381 views • 7 months ago

New video series: Physics Informed Machine Learning! Physics may be embedded into AI/ML in 5 stages: 1 choose what to Model 2 curate training Data 3 design an Architecture 4 craft a Loss Function, and 5 implement Optimization Algorithm to train the model

New video series: Physics Informed Machine Learning! Physics may be embedded into AI/ML in 5 stages: 1 choose what to Model 2 curate training Data 3 design an Architecture 4 craft a Loss Function, and 5 implement Optimization Algorithm to train the model

Steven Brunton

78,531 views • 2 years ago

1/7 Physical AI has a data problem. More compute alone won’t teach a robot how to move, grasp, react, or recover from mistakes. LLMs had the entire internet to learn from. Robots need something different: huge amounts of real action data. That’s where Axis Robotics gets interesting. $AXIS is building a data engine for Physical AI, turning robot training into something people can contribute to directly through a browser. The loop is straightforward: Generate → Collect → Process → Deploy → Correct → Repeat. What looks simple on the surface could become something much bigger underneath: a scalable way to produce the training data robots need to become genuinely useful in the physical world. The real unlock for Physical AI might not be another bigger model. It might be better data.

1/7 Physical AI has a data problem. More compute alone won’t teach a robot how to move, grasp, react, or recover from mistakes. LLMs had the entire internet to learn from. Robots need something different: huge amounts of real action data. That’s where Axis Robotics gets interesting. $AXIS is building a data engine for Physical AI, turning robot training into something people can contribute to directly through a browser. The loop is straightforward: Generate → Collect → Process → Deploy → Correct → Repeat. What looks simple on the surface could become something much bigger underneath: a scalable way to produce the training data robots need to become genuinely useful in the physical world. The real unlock for Physical AI might not be another bigger model. It might be better data.

_victorad

18,903 views • 13 days ago

Jensen Huang says agentic AI is not a chatbot. It is a "digital robot" -- and the next bottleneck is not the model. It is first-person data. "Agentic AI is just a digital robot. It understands, it reasons, it plans, and it acts and uses tools." Then he makes the bigger jump: "The world's video data is from a third person, not first person. And so agentic systems, robotic systems, physical AI -- the data is the hardest problem." That is the hidden constraint in robotics. LLMs got the internet. Physical AI needs demonstrations, synthetic worlds, reinforcement learning, and first-person experience. This is why Jensen keeps calling it a stack: robots are not waiting for one magical model. They need the data engine, the simulator, the reward system, the runtime, and the computer. The next wave of AI is not just software that talks. It is software that sees, plans, acts, and eventually moves through the world.

Jensen Huang says agentic AI is not a chatbot. It is a "digital robot" -- and the next bottleneck is not the model. It is first-person data. "Agentic AI is just a digital robot. It understands, it reasons, it plans, and it acts and uses tools." Then he makes the bigger jump: "The world's video data is from a third person, not first person. And so agentic systems, robotic systems, physical AI -- the data is the hardest problem." That is the hidden constraint in robotics. LLMs got the internet. Physical AI needs demonstrations, synthetic worlds, reinforcement learning, and first-person experience. This is why Jensen keeps calling it a stack: robots are not waiting for one magical model. They need the data engine, the simulator, the reward system, the runtime, and the computer. The next wave of AI is not just software that talks. It is software that sees, plans, acts, and eventually moves through the world.

Karl Mehta

14,140 views • 2 months ago

Robot policies must be both reliable and highly capable to be useful; the best way to achieve this level of performance is with reinforcement learning. However, for reinforcement learning you are usually stuck between two difficult options: reinforcement in the real world is often risky and expensive, while reinforcement learning in a traditional simulator takes a lot of engineering work and has a persistent sim-to-real gap. What if instead you could train your robot purely in a world model? RISE by Jiazhi Yang et al. uses a compositional world model to predict the future and evaluate progress. This allows for a self-improving pipeline, which learns a world model from real data and then learns how the robot should perform different tasks. This pipeline results in a data-driven way to improve policy performance from real data but without real-world reinforcement learning. Watch Episode #86 of RoboPapers, with Chris Paxton and Jiafei Duan, to learn more!

Robot policies must be both reliable and highly capable to be useful; the best way to achieve this level of performance is with reinforcement learning. However, for reinforcement learning you are usually stuck between two difficult options: reinforcement in the real world is often risky and expensive, while reinforcement learning in a traditional simulator takes a lot of engineering work and has a persistent sim-to-real gap. What if instead you could train your robot purely in a world model? RISE by Jiazhi Yang et al. uses a compositional world model to predict the future and evaluate progress. This allows for a self-improving pipeline, which learns a world model from real data and then learns how the robot should perform different tasks. This pipeline results in a data-driven way to improve policy performance from real data but without real-world reinforcement learning. Watch Episode #86 of RoboPapers, with Chris Paxton and Jiafei Duan, to learn more!

RoboPapers

38,772 views • 2 months ago

Using reinforcement learning we have expanded the range of techniques the Ultra Mobile Vehicle (UMV) uses to handle terrain and obstacles, including hops, out-of-plane balance, and level-ground flips. Millions of physics-based simulations provide training data to support zero-shot transfers.

Using reinforcement learning we have expanded the range of techniques the Ultra Mobile Vehicle (UMV) uses to handle terrain and obstacles, including hops, out-of-plane balance, and level-ground flips. Millions of physics-based simulations provide training data to support zero-shot transfers.

RAI Institute

75,199 views • 1 year ago

This is 🤯 ! Exclusive GPS data from Ross Herridge & Cricket Australia The physical requirements of being a bowler in international cricket are WILD. 7Cricket | #AUSvWI

This is 🤯 ! Exclusive GPS data from Ross Herridge & Cricket Australia The physical requirements of being a bowler in international cricket are WILD. 7Cricket | #AUSvWI

Trent Copeland

70,384 views • 2 years ago

Using our brain simulator, we’ve trained a reinforcement learning agent to maximize bits per second. Here is the RL policy converting brain data to cursor control in simulation:

Using our brain simulator, we’ve trained a reinforcement learning agent to maximize bits per second. Here is the RL policy converting brain data to cursor control in simulation:

Neuralink

18,410 views • 1 year ago

Our course recommendation of the day is “Post-training of LLMs, ” where you’ll learn how to customize pre-trained language models using Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning (RL). You'll learn when to use each method, how to curate training data, and implement them in code to shape model behavior effectively. Enroll here:

Our course recommendation of the day is “Post-training of LLMs, ” where you’ll learn how to customize pre-trained language models using Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning (RL). You'll learn when to use each method, how to curate training data, and implement them in code to shape model behavior effectively. Enroll here:

DeepLearning.AI

29,369 views • 11 months ago

From the latest episode with Ivan Poupyrev from Archetype AI The idea is simple but revolutionary: -- using sensor data, similar to how humans observe the physical world through our biological sensors -- to let machines 'talk' to the physical world. In the case of packages, this means not just knowing where it is, but what's happening to it in real-time.

From the latest episode with Ivan Poupyrev from Archetype AI The idea is simple but revolutionary: -- using sensor data, similar to how humans observe the physical world through our biological sensors -- to let machines 'talk' to the physical world. In the case of packages, this means not just knowing where it is, but what's happening to it in real-time.

This Week in Startups

131,725 views • 2 years ago

The Machine That Learns The Law Behind The Data A very very interesting US Patent US10963540B2 - Physics Informed Learning Machine describes a learning system that does not begin with data alone. It begins with a physical model, usually written as a differential equation (or PDE) dx/dt = f(x,t) A normal Machine Learning model sees scattered data and tries to fit it. A physics-informed learning machine starts with a law. Then it treats the data as evidence that updates what the model believes about the physical system. For this application, I use the patent idea on NASA C-MAPSS Turbofan engine data. The machine watches multivariate telemetry from a degrading engine and infers a hidden health state that is not measured directly. From that posterior belief, it estimates the engine’s remaining useful life. In the main 3D scene, the engine lifetime is turned into a tunnel. The spiral ribbons are real sensor channels evolving over cycle-time. The glowing core is the inferred health state. The surrounding cloud is uncertainty. The orange wall ahead is the predicted failure horizon. So the big picture is: sensor evidence comes in, posterior belief tightens, and the machine moves from uncertainty toward a concrete failure prediction. The inset posteriors make that explicit. The health posterior shows where the model believes the hidden engine condition sits at the current moment, and how sharply it believes it. The RUL posterior shows the same idea for remaining life... early on it is broad, later it shifts left and narrows as the machine becomes more certain about how close failure is. This idea is not limited to engines. The same idea can apply to data centers, CPUs, GPUs, cooling systems, power grids, robotics, batteries, and any machine that produces telemetry while obeying physical constraints. In an age where machine learning runs on massive hardware infrastructure, this kind of model matters: it can turn noisy sensor streams into early warnings before expensive systems fail.

The Machine That Learns The Law Behind The Data A very very interesting US Patent US10963540B2 - Physics Informed Learning Machine describes a learning system that does not begin with data alone. It begins with a physical model, usually written as a differential equation (or PDE) dx/dt = f(x,t) A normal Machine Learning model sees scattered data and tries to fit it. A physics-informed learning machine starts with a law. Then it treats the data as evidence that updates what the model believes about the physical system. For this application, I use the patent idea on NASA C-MAPSS Turbofan engine data. The machine watches multivariate telemetry from a degrading engine and infers a hidden health state that is not measured directly. From that posterior belief, it estimates the engine’s remaining useful life. In the main 3D scene, the engine lifetime is turned into a tunnel. The spiral ribbons are real sensor channels evolving over cycle-time. The glowing core is the inferred health state. The surrounding cloud is uncertainty. The orange wall ahead is the predicted failure horizon. So the big picture is: sensor evidence comes in, posterior belief tightens, and the machine moves from uncertainty toward a concrete failure prediction. The inset posteriors make that explicit. The health posterior shows where the model believes the hidden engine condition sits at the current moment, and how sharply it believes it. The RUL posterior shows the same idea for remaining life... early on it is broad, later it shifts left and narrows as the machine becomes more certain about how close failure is. This idea is not limited to engines. The same idea can apply to data centers, CPUs, GPUs, cooling systems, power grids, robotics, batteries, and any machine that produces telemetry while obeying physical constraints. In an age where machine learning runs on massive hardware infrastructure, this kind of model matters: it can turn noisy sensor streams into early warnings before expensive systems fail.

Mathelirium

17,843 views • 4 months ago

🔥🚨BREAKING NEWS: Hackers may have stolen every American Social Security number and physical address from National Public Data, around 2.9 BILLION records. This is one of the largest data breeches of all time.

🔥🚨BREAKING NEWS: Hackers may have stolen every American Social Security number and physical address from National Public Data, around 2.9 BILLION records. This is one of the largest data breeches of all time.

Dom Lucre | Breaker of Narratives

3,507,297 views • 2 years ago

What actually happens when an OptimAI Agent prepares a tweet? • Style fingerprinting: learning the patterns behind your writing • Voice alignment: adapting tone across contexts and audiences • Engagement signal analysis: decoding what the network responds to • Parameter optimization: refining outputs through reinforcement loops Each post becomes a new data point. Each signal strengthens the agent. Not just generation. Continuous learning in motion. Try now:

What actually happens when an OptimAI Agent prepares a tweet? • Style fingerprinting: learning the patterns behind your writing • Voice alignment: adapting tone across contexts and audiences • Engagement signal analysis: decoding what the network responds to • Parameter optimization: refining outputs through reinforcement loops Each post becomes a new data point. Each signal strengthens the agent. Not just generation. Continuous learning in motion. Try now:

OptimAI Network

11,291 views • 6 months ago