Abstract: Training a neural network and embedding a watermark in the network to prove ownership. The network includes a plurality of trainable parameters associated with network nodes in which the plurality of trainable parameters is split into a first set of trainable parameters and a second set of trainable parameters. A first set of training samples is input to the network and the network is trained by iterating the first set of samples through the network to update the first set of parameters and hindering the second set of parameters to be updated during iteration of the first set of samples. A second set of samples is input and the watermark is embedded by iterating the second set of samples through the network to update the second set of parameters and hindering the first set of parameters to be updated during iteration of the second set of samples.

Abstract: A method is provided for producing a warhead component including an inner shell. The method includes arranging a first casing of a thermoplastic to enclose the inner shell, arranging pre-formed projectiles between the first casing of thermoplastic and the inner shell, and heat-treating the first casing so that the first casing shrinks and fixes the pre-formed projectiles to the inner shell. A warhead, a projectile, and a bomb are also provided.

Abstract: A method is provided for producing a component for a warhead, wherein the method involves the steps: i.) an inner shell is arranged on a tool, ii.) preshaped projectiles are arranged on the inner shell, iii.) powder is arranged to enclose the preshaped projectiles, iv.) the powder is pressed such that the powder, the preshaped projectiles, and the inner shell are joined together, v.) the tool is removed from the component formed from the powder, the preshaped projectiles, and the inner shell. A warhead and a projectile are also provided.

Abstract: This document concerns a device for airflow-assisted feeding of granular material in an agricultural implement, comprising an air inlet (1611, 1612), a material inlet (167), and an outlet (162) for material-laden airflow, a roof (163), the horizontal extension of which defines a material pickup zone (160), wherein the material inlet (167), air inlet (1611, 1612) and outlet (162) are connected to the material pickup zone (160), wherein the device furthermore comprises an air inlet valve (164) containing a valve seat (1641) and a valve body (1642). The valve body is rotatable about a geometric axis which is substantially perpendicular to a flow direction at the air inlet. The document also concerns a method for airflow-assisted feeding of material in an agricultural implement, a meter housing, and an agricultural implement comprising such a meter housing. It also concerns a frame for a material pickup module and a method for forming a meter housing.

Abstract: A method for producing a warhead with directed blasting action involves the step of making an outer casing and arranging therein a charge with a cavity situated at the front end of the charge. An insert whose shape corresponds to the shape of the cavity is arranged on the surface of the charge. The method furthermore involves the step of depositing a material on the insert in an additive manufacturing process. A warhead manufactured according to the method includes an outer casing, a charge with a cavity situated at the front end of the charge, and an insert. A material is deposited on the insert.

Abstract: A method is provided for producing a warhead component including an inner shell. The method includes arranging a first casing of a thermoplastic to enclose the inner shell, arranging pre-formed projectiles between the first casing of thermoplastic and the inner shell, and heat-treating the first casing so that the first casing shrinks and fixes the pre-formed projectiles to the inner shell. A warhead, a projectile, and a bomb are also provided.

Abstract: A projectile is outfitted with a target finder and is intended to be fired from a launching device and includes one or several brake flaps where the brake flaps, after being extended, cause the projectile to be arranged in a controlled oscillating movement. A fuse is also provided.

Abstract: The present disclosure relates to a crash monitoring system for a vehicle, a vehicle comprising such a crash monitoring system, a method for monitoring a vehicle crash and a computer program element for such a crash monitoring system. The crash monitoring system comprises an environment sensor unit, a vehicle information supplying unit, a crash sensor unit and a control unit. The environment sensor unit is configured to generate field data of an impending crash event between at least a first vehicle and an obstacle. The vehicle information supplying unit is configured to provide operational information of at least the first vehicle to the control unit. The crash sensor unit is configured to generate crash data during the crash event of at least the first vehicle. The control unit is configured to execute a crash simulation based on the field data and the operational information and generate predictive attribute data.

Abstract: A procedure for producing a warhead with controlled fragmentation includes manufacturing an inner shell and filling it with an explosive substance. The procedure further includes providing a net around the outer surface of the inner shell as well as depositing a material on and around the net. A warhead is produced by the procedure and includes an inner shell. A material applied to the warhead comprises weak points for guided fragmentation of the deposited material upon detonation of the warhead.

Abstract: A neural network having one or more public parts and one or more confidential parts is trained to perform a primary task. A deployment instantiation of the neural network is trained based on optimal performance of the primary task, and based on sub-optimal performance of the primary task conditioned on the confidential parts of the deployment instantiation being inaccessible. An adversary instantiation of the neural network is trained based on optimal performance of the primary task conditioned on the public parts being identical for the deployment instantiation and for the adversary instantiation, and conditioned on the confidential parts of the deployment instantiation being inaccessible. The training of the deployment instantiation and the training of the adversary instantiation are based on a plurality of training data samples, and are performed iteratively by alternating between the training of the deployment instantiation and the training of the adversary instantiation.

Abstract: A method is provided for producing a component for a warhead, wherein the method involves the steps: i.) an inner shell is arranged on a tool, ii.) preshaped projectiles are arranged on the inner shell, iii.) powder is arranged to enclose the preshaped projectiles, iv.) the powder is pressed such that the powder, the preshaped projectiles, and the inner shell are joined together, v.) the tool is removed from the component formed from the powder, the preshaped projectiles, and the inner shell. A warhead and a projectile are also provided.

Abstract: A method for determining a reference control profile for a vehicle, including determining (A1) a plurality of control profiles for the vehicle (1) based on topographic information for one and the same future route segment and selecting (A2) one of the plurality of control profiles as a reference control profile, the selection (A2) being based on a tradeoff criterion considering resource consumption and drivability of the plurality of control profiles. The disclosure also relates to a control arrangement (2) for determining a reference control profile for a vehicle (1) and a vehicle comprising the control arrangement (2).

Abstract: Training a neural network and embedding a watermark in the network to prove ownership. The network includes a plurality of trainable parameters associated with network nodes in which the plurality of trainable parameters is split into a first set of trainable parameters and a second set of trainable parameters. A first set of training samples is input to the network and the network is trained by iterating the first set of samples through the network to update the first set of parameters and hindering the second set of parameters to be updated during iteration of the first set of samples. A second set of samples is input and the watermark is embedded by iterating the second set of samples through the network to update the second set of parameters and hindering the first set of parameters to be updated during iteration of the second set of samples.

Abstract: The present disclosure relates to a crash monitoring system for a vehicle, a vehicle comprising such a crash monitoring system, a method for monitoring a vehicle crash and a computer program element for such a crash monitoring system. The crash monitoring system comprises an environment sensor unit, a vehicle information supplying unit, a crash sensor unit and a control unit. The environment sensor unit is configured to generate field data of an impending crash event between at least a first vehicle and an obstacle. The vehicle information supplying unit is configured to provide operational information of at least the first vehicle to the control unit. The crash sensor unit is configured to generate crash data during the crash event of at least the first vehicle. The control unit is configured to execute a crash simulation based on the field data and the operational information and generate predictive attribute data.

Abstract: The present invention relates to a nozzle for an air trapping device configured to remove air from a fluid, the nozzle comprising a body having an input opening configured to receive the fluid and an output opening configured to distribute the fluid along an edge of the output opening, wherein the edge comprises a control element configured to reduce surface tension of the fluid. The present invention further relates to an air trapping device configured to remove air from a fluid and comprising the nozzle.

Abstract: A distributor device for use in a filling shoe for filling a mould cavity of a powder compression die, the distributor device having an inlet portion connectable to a powder supply; an outlet portion with an outlet opening; and a distributor portion arranged between the inlet portion and the outlet portion. The distributor portion includes one or more guide elements arranged to divide the distributor portion into a plurality of distributor channels. The distributor channels have an input with an input cross-sectional area at an upstream end of the distributor channel and an output with an output cross-sectional area at a downstream end of the distributor channel, wherein the input cross-sectional area differs from the output cross-sectional area for at least one of the distributor channels.

Abstract: The present invention relates to a nozzle for an air trapping device configured to remove air from a fluid, the nozzle comprising a body having an input opening configured to receive the fluid and an output opening configured to distribute the fluid along an edge of the output opening, wherein the edge comprises a control element configured to reduce surface tension of the fluid. The present invention further relates to an air trapping device 200 configured to remove air from a fluid and comprising the nozzle.

Abstract: A diapharagm pump for a bioprocess system comprises a pump head (1; 101; 201; 301) comprising: a common inlet (3); a common outlet (5); a plurality of pump cavities (7) each including at least one cooperating pair of one-way valves, the at least one pair of one-way valves including an inlet valve (9) and an outlet valve (11), wherein the respective inlet valves (9) are in fluid communication with the common inlet (3) and the respective outlet valves (11) are in fluid communication with the common outlet (5) and wherein a centre of the outlet valve (11) for each pump cavity is positionable above a centre of the inlet valve (9) for the same pump cavity when the diaphragm pump is oriented for use to inhibit trapped gas; and a plurality of moveable diaphragms (13) each respectively provided in a respective of said pump cavities (7) for varying a volume of the pump cavities.

Abstract: A computer-implemented machine learning method is disclosed for training of a neural network to perform a primary task. The method comprises determining the neural network to comprise one or more public parts and one or more confidential parts, training a deployment instantiation of the neural network based on optimal performance of the primary task, and based on sub-optimal performance of the primary task conditioned on the confidential parts of the deployment instantiation being inaccessible, and training an adversary instantiation of the neural network based on optimal performance of the primary task conditioned on the public parts being identical for the deployment instantiation and for the adversary instantiation, and conditioned on the confidential parts of the deployment instantiation being inaccessible.

Abstract: This document concerns a device for airflow-assisted feeding of granular material in an agricultural implement, comprising an air inlet (1611, 1612), a material inlet (167), and an outlet (162) for material-laden airflow, a roof (163), the horizontal extension of which defines a material pickup zone (160), wherein the material inlet (167), air inlet (1611, 1612) and outlet (162) are connected to the material pickup zone (160), wherein the device furthermore comprises an air inlet valve (164) containing a valve seat (1641) and a valve body (1642). The valve body is rotatable about a geometric axis which is substantially perpendicular to a flow direction at the air inlet. The document also concerns a method for airflow-assisted feeding of material in an agricultural implement, a meter housing, and an agricultural implement comprising such a meter housing. It also concerns a frame for a material pickup module and a method for forming a meter housing.