Dexterous magnetic manipulation of ferromagnetic objects is well established, with three to six degrees of freedom possible depending on object geometry. There are objects for which non-contact dexterous manipulation is desirable that do not contain an appreciable amount of ferromagnetic material but do contain electrically conductive material. Time-varying magnetic fields generate eddy currents in conductive materials, with resulting forces and torques due to the interaction of the eddy currents with the magnetic field. This phenomenon has previously been used to induce drag to reduce the motion of objects as they pass through a static field, or to apply force on an object in a single direction using a dynamic field, but has not been used to perform the type of dexterous manipulation of conductive objects that has been demonstrated with ferromagnetic objects. Here we show that manipulation, with six degrees of freedom, of conductive objects is possible by using multiple rotating magnetic dipole fields. Using dimensional analysis, combined with multiphysics numerical simulations and experimental verification, we characterize the forces and torques generated on a conductive sphere in a rotating magnetic dipole field. With the resulting model, we perform dexterous manipulation in simulations and physical experiments. <<<

Deep-learning models have become pervasive tools in science and engineering. However, their energy requirements now increasingly limit their scalability. Deep-learning accelerators aim to perform deep learning energy-efficiently, usually targeting the inference phase and often by exploiting physical substrates beyond conventional electronics. Approaches so far have been unable to apply the backpropagation algorithm to train unconventional novel hardware in situ. The advantages of backpropagation have made it the de facto training method for large-scale neural networks, so this deficiency constitutes a major impediment. Here we introduce a hybrid in situ-in silico algorithm, called physics-aware training, that applies backpropagation to train controllable physical systems. Just as deep learning realizes computations with deep neural networks made from layers of mathematical functions, our approach allows us to train deep physical neural networks made from layers of controllable physical systems, even when the physical layers lack any mathematical isomorphism to conventional artificial neural network layers. To demonstrate the universality of our approach, we train diverse physical neural networks based on optics, mechanics and electronics to experimentally perform audio and image classification tasks. Physics-aware training combines the scalability of backpropagation with the automatic mitigation of imperfections and noise achievable with in situ algorithms. Physical neural networks have the potential to perform machine learning faster and more energy-efficiently than conventional electronic processors and, more broadly, can endow physical systems with automatically designed physical functionalities, for example, for robotics, materials and smart sensors. <<<

Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech. <<<