Abstract: There is disclosed a method and system for identifying relevant training data points for a prediction by a machine learning algorithm (MLA). The method comprises determining indicators of influence for training data points used to train the MLA. A normalizing function is applied to the indicators of influence to determine normalized indicators of influence. Training data points having the highest and lowest normalized indicators of influence are output for display.

Abstract: A flutter reducer for protecting at least partially a tarpaulin defining a tarpaulin rear end edge from flutter induced damages when the tarpaulin is secured laterally across a container moved at speed, the container defining a container aperture delimited by an aperture peripheral edge. The flutter reducer is operatively coupled to the tarpaulin so that when the tarpaulin is extended across the container aperture configuration and the container is moved by the vehicle, the flutter reducer limits movements of the tarpaulin at its tarpaulin rear end edge to reduce flutter due to relative wind caused by movement through air of the container.

Abstract: A crank assembly for operating a tarpaulin manipulator used for extending and retracting a tarpaulin respectively across and from a top aperture of a vehicle container. The crank assembly includes a crank and a crank assembly shaft for transmitting rotations of the crank to the tarpaulin manipulator through a transmission assembly. The crank can be moved between a mounted configuration in which the crank protrudes axially from the crank assembly shaft and a stowed configuration in which the crank is folded over the crank assembly shaft.

Abstract: A crank assembly for operating a tarpaulin manipulator used for extending and retracting a tarpaulin respectively across and from a top aperture of a vehicle container. The crank assembly includes a crank and a crank assembly shaft for transmitting rotations of the crank to the tarpaulin manipulator through a transmission assembly. The crank can be moved between a mounted configuration in which the crank protrudes axially from the crank assembly shaft and a stowed configuration in which the crank is folded over the crank assembly shaft.

Abstract: There is disclosed a method and system for identifying relevant training data points for a prediction by a machine learning algorithm (MLA). The method comprises determining indicators of influence for training data points used to train the MLA. A normalizing function is applied to the indicators of influence to determine normalized indicators of influence. Training data points having the highest and lowest normalized indicators of influence are output for display.

Abstract: A method for preparing a food that is less allergenic from a particulate solid food matrix that is more allergenic and contains allergenic proteins, in particular a peanut and/or tree-nut matrix. The method comprises the step of processing dispersion of the solid food matrix in an aqueous medium by high-pressure homogenization to obtain a homogenate consisting of a dispersion of allergenic proteins initially contained in the matrix mixed with solid particles of the matrix. The homogenization step can be preceded by a step of grinding the matrix, and followed by a step of processing the homogenate with a protein-degrading agent, in particular with a probiotic microorganism.