OpenAI CEO Sam Altman believes the next major developments in the AI sector will be more disruptive than most people expect.
Category: robotics/AI – Page 470
Training all-mechanical neural networks for task learning through in situ backpropagation
Another well-known method for physical learning is Equilibrium Propagation (EP), sharing similar procedure with coupled learning and being able to define the arbitrary differentiable loss function32. This method has been demonstrated in various physical systems, numerically in nonlinear resistor networks33 and coupled phase oscillators34, experimentally on Ising machines35.
So far, the MNNs based on the physical learning have been developed using the platform of origami structures28,36 and disordered networks29,37 to demonstrate machine learning through simulations. The experimental proposals involve using directed springs with variable stiffness38 and manually adjusting the rest length of springs31.
Here, we present a highly-efficient training protocol for MNNs through mechanical analogue of in situ backpropagation, derived from the adjoint variable method, in which theoretically the exact gradient can be obtained from only the local information. By using 3D-printed MNNs, we demonstrate the feasibility of obtaining the gradient of the loss function experimentally solely from the bond elongation of MNNs in only two steps, using local rules, with high accuracy. Besides, leveraging the obtained gradient, we showcase the successful training in simulations of a mechanical network for behaviors learning and various machine learning tasks, achieving high accuracy in both regression and Iris flower classification tasks. The trained MNNs are then validated both numerically and experimentally. In addition, we illustrate the retrainability of MNNs after switching tasks and damage, a feature that may inspire further inquiry into more robust and resilient design of MNNs.
DeepMind AI predicts weather more accurately than existing forecasts
The latest weather forecasting AI model from Google DeepMind can beat the leading providers more than 97 per cent of the time, and it is quicker and cheaper to run.
AI Supercharging Crop Breeding to Protect Farmers from Climate
Avalo, a crop development company based in North Carolina, is using machine learning models to accelerate the creation of new and resilient crop varieties.
The traditional way to select for favorable traits in crops is to identify individual plants that exhibit the trait – such as drought resistance – and use those plants to pollinate others, before planting those seeds in fields to see how they perform. But that process requires growing a plant through its entire life cycle to see the result, which can take many years.
Avalo uses an algorithm to identify the genetic basis of complex traits like drought, or pest resistance in hundreds of crop varieties. Plants are cross-pollinated in the conventional way, but the algorithm can predict the performance of a seed without needing to grow it – speeding up the process by as much as 70%, according to Avalo chief technology officer Mariano Alvarez.
Can AI become demon-possessed?
That’s our question.
When we ask whether AI can become possessed by a spirit sent by Satan, we might beflummox our minds with an unnecessary detour. That detour is puzzlement over the relationship between disembodied spirits and machine intelligence. Such mental machinations are just as “creepy” as personifying a chatbot or robot. And they are oblique. They fail to provide a path toward understanding the presence of evil.
Google Quantum AI: New Quantum Chip Outperforms Classical Computers and Breaks Error Correction Threshold
Google Quantum AI announced that it is moving past the Sycamore era and taking another leap down its roadmap with the introduction of the 105-qubit Willow, a new quantum chip that has achieved a milestone in computational power and error correction, marking a major step toward large-scale, commercially viable quantum computing.
The team, which published their findings in Nature, is also eyeing a quantum device that overcomes the limitations of errors and offers real-world solutions to tough problems, the ultimate destination as they progress along their roadmap.
“The mission of the Google quantum AI team is to build quantum computing for otherwise unsolvable problems,” said Hartmut Neven, a vice president of engineering at Google and founder and manager of the Quantum Artificial Intelligence lab, at a recent roundtable about the new milestone.” So what problems do we have in mind? The first applications will be modeling and understanding systems where quantum effects are important. So that’s the case for common drug discovery, understanding and designing nuclear fusion reactors, bringing down the enormous energy costs of fertilizer production. But it then extends to multiple other areas, such as quantum machine learning.”
