Abstract: The specification discloses robotic apparatus for extraction of a worn treatment panel module from a treatment deck of ore material treatment apparatus without the need for work personnel entering the treatment apparatus, the robotic apparatus including a robotic support structure carried on transport means whereby the robotic support structure follows a course on the treatment deck, the robotic apparatus further including a treatment panel module handling mechanism connected to the robotic support structure for movement along a defined path of movement and first drive means for selectably moving the treatment panel module handling mechanism along the defined path of movement, the treatment panel module handling mechanism still further including tool means adapted to mount tool means cooperable at least with a treatment panel module intended to be removed from a selected treatment deck, the tool means being able to engage a panel module to lift it out of the treatment deck.

Abstract: The present specification discloses a mobile scanning device (30) for scanning wear conditions of treatment panel modules (13) of a treatment deck (14) in vibratory treatment apparatus (10), the mobile scanning device (30) including a support structure (11), steerable transport means (12) carrying the support structure (11) along a predefined course or a selectable steered course on or over the treatment deck (14), the support structure (11) carrying power means (16), drive means (17) powered by the power means (16) to drive the steerable transport means (12) scanning means (18) to scan a lower zone (42) below the support structure (11) to establish scanned information, the scanned information being reflective of at least wear levels occurring on the upper surfaces (31) of the treatment panel modules (13), and transmission means (23) arranged to transmit the scanned information to remote control means (20, 21) external of the treatment deck (14).

Abstract: A variational autoencoder (VAE) neural network system, comprising an encoder neural network to encode an input data item to define a posterior distribution for a set of latent variables, and a decoder neural network to generate an output data item representing values of a set of latent variables sampled from the posterior distribution. The system is configured for training with an objective function including a term dependent on a difference between the posterior distribution and a prior distribution. The prior and posterior distributions are arranged so that they cannot be matched to one another. The VAE system may be used for compressing and decompressing data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. One of the methods includes performing, using a plurality of training examples, a training step to obtain respective gradients of a loss function with respect to each of the parameters in the parameter tensors; obtaining a validation loss for a plurality of validation examples that are different from the plurality of training examples generating an optimizer input from at least the respective gradients and the validation loss; processing the optimizer input using an optimizer neural network to generate an output defining a respective update for each of the parameters in the parameter tensors of the neural network; and for each of the parameters in the parameter tensors, applying the respective update to a current value of the parameter to generate an updated value for the parameter.

Abstract: A variational autoencoder (VAE) neural network system, comprising an encoder neural network to encode an input data item to define a posterior distribution for a set of latent variables, and a decoder neural network to generate an output data item representing values of a set of latent variables sampled from the posterior distribution. The system is configured for training with an objective function including a term dependent on a difference between the posterior distribution and a prior distribution. The prior and posterior distributions are arranged so that they cannot be matched to one another. The VAE system may be used for compressing and decompressing data.

Abstract: A variational autoencoder (VAE) neural network system, comprising an encoder neural network to encode an input data item to define a posterior distribution for a set of latent variables, and a decoder neural network to generate an output data item representing values of a set of latent variables sampled from the posterior distribution. The system is configured for training with an objective function including a term dependent on a difference between the posterior distribution and a prior distribution. The prior and posterior distributions are arranged so that they cannot be matched to one another. The VAE system may be used for compressing and decompressing data.

Abstract: A variational autoencoder (VAE) neural network system, comprising an encoder neural network to encode an input data item to define a posterior distribution for a set of latent variables, and a decoder neural network to generate an output data item representing values of a set of latent variables sampled from the posterior distribution. The system is configured for training with an objective function including a term dependent on a difference between the posterior distribution and a prior distribution. The prior and posterior distributions are arranged so that they cannot be matched to one another. The VAE system may be used for compressing and decompressing data.