Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an environment representation neural network of a reinforcement learning system controls an agent to perform a given task. In one aspect, the method includes: receiving a current observation input and a future observation input; generating, from the future observation input, a future latent representation of the future state of the environment; processing, using the environment representation neural network, to generate a current internal representation of the current state of the environment; generating, from the current internal representation, a predicted future latent representation; evaluating an objective function measuring a difference between the future latent representation and the predicted future latent representation; and determining, based on a determined gradient of the objective function, an update to the current values of the environment representation parameters.

Abstract: A Method for generating a biomarker includes acquiring an image using an MRI system; processing the MRI image to generate a three-dimensional image of the lung; generating a first function corresponding to the distribution of the different signal intensity values; automatically calculating a filtering threshold of the first function from a second signal intensity value distribution function; segmenting a lung volume comprising: a main volume; a filtered volume of a volume of voxels quantified by the first function and filtered by at least the calculated filtering threshold, normalizing the values of the three-dimensional image of the lung volume; generating a biomarker indicating a normalized segmented volume ratio.

Abstract: The present disclosure relates to utilizing laser ablation to selectively remove regions of active material from lithium-ion battery electrodes during production, forming channels. These channels may improve the radius of curvature of the lithium-ion battery, form patterns on the surface of the electrode to store excess electrolyte, provide improved access to external sources of lithium during pre-lithiation or re-lithiation, and/or lower the tortuosity of electrodes for enhanced ion transfer between electrodes during cycling in prismatic cells and cylindrical cells.

Abstract: An MRI method is disclosed for the geometrical characterisation of pulmonary airways on the basis of a first tridimensional image of at least one bronchus of a bronchial tree and includes: acquiring a first MRI image, an MRI sequence being synchronised with a respiratory frequency; filtering the first image; segmenting a portion of the filtered image including the contours of the bronchial tree and its inner volumetric portion; estimating at least one plane of cut of a bronchus; generating at least one image slice of a bronchus; estimating, for each image slice, an area contained within the bronchial wall and/or a bronchial wall thickness.

Abstract: A computer-implemented method of training a neural network. The method comprises processing a first transformed view of a training data item, e.g. an image, with a target neural network to generate a target output, processing a second transformed view of the training data item, e.g. image, with an online neural network to generate a prediction of the target output, updating parameters of the online neural network to minimize an error between the prediction of the target output and the target output, and updating parameters of the target neural network based on the parameters of the online neural network. The method can effectively train an encoder neural network without using labelled training data items, and without using a contrastive loss, i.e. without needing “negative examples” which comprise transformed views of different data items.

Abstract: An MRI method is disclosed for the geometrical characterisation of pulmonary airways on the basis of a first tridimensional image of at least one bronchus of a bronchial tree and includes: acquiring a first MRI image, an MRI sequence being synchronised with a respiratory frequency; filtering the first image; segmenting a portion of the filtered image including the contours of the bronchial tree and its inner volumetric portion; estimating at least one plane of cut of a bronchus; generating at least one image slice of a bronchus; estimating, for each image slice, an area contained within the bronchial wall and/or a bronchial wall thickness.

Abstract: A modular section of water pipe (114) includes: a deformable membrane (130) able to encompass, in an operational state of the section, a tubular space (132) defining an axial direction (AA?) to conduct water, and a series (135) of rings (120, (140) extending along the axial direction (AA?) in the tubular space (132), and including: two end rings (120), each being at a separate end (116, 118) of the section (114) along the axial direction (AA?), the membrane (130) being fastened to the end rings (120), at least one central ring (140), arranged between the two end rings (120), and cables (150, 160) connecting each ring (120, 140) to the closest ring (120, 140) along the axial direction (AA?).

Abstract: A method of fabricating a martensitic stainless steel including: 1) heating steel to a temperature higher than austenizing temperature of the steel, then quenching the steel until a hottest portion of the steel is at a temperature less than or equal to a maximum temperature, and greater than or equal to a minimum temperature, a cooling rate being sufficiently fast for austenite not to transform into a ferrito-perlitic structure; 2) performing a first anneal followed by cooling until the hottest portion of the steel is at a temperature less than or equal to the maximum temperature and greater than or equal to the minimum temperature; 3) performing a second anneal followed by cooling to ambient temperature; and at the end of each of 1) and 2), performing: ?) as soon as temperature of the hottest portion of the steel reaches the maximum temperature, immediately heating the steel once more.

Abstract: The embodiments described herein present methods and apparatuses for on-path CAC operations in a MPLS-VPN environment. In an example embodiment, an ingress PE device receives a quality of service (QoS) resource reservation request; constructs an outgoing message that includes information allowing an egress PE device to identify the virtual private network routing and forwarding table (VRF) associated with a resource reservation resulting from the QoS resource reservation request; and transmits the outgoing message to the egress PE device, where the information allowing VRF identification is echoed back by the egress PE device and used by the ingress PE device to identify the VRF associated with the resource reservation resulting from the QoS resource reservation request. Other embodiments are described.

Abstract: The invention relates to peptides and, in particular, heptapeptides derived from milk casein ?s1 and having anxiolytic activity. The invention also relates to pharmaceutical and food compositions containing said peptides and to the preparation methods thereof.

Abstract: A device for polishing centrifugal impellers for a turbomachine including a vat configured to be filled with a polishing agent, and an impeller support configured to make the impeller rotate around its axis and move it along its axis such that all of points of the impeller have a helical movement whereof the pitch is close to that of the helix from which the general shape of the airflow channels of the impeller comes, delimited by the blades of the impeller.

Abstract: The embodiments described herein present methods and apparatuses for on-path CAC operations in a MPLS-VPN environment. In an example embodiment, an ingress PE device receives a quality of service (QoS) resource reservation request; constructs an outgoing message that includes information allowing an egress PE device to identify the virtual private network routing and forwarding table (VRF) associated with a resource reservation resulting from the QoS resource reservation request; and transmits the outgoing message to the egress PE device, where the information allowing VRF identification is echoed back by the egress PE device and used by the ingress PE device to identify the VRF associated with the resource reservation resulting from the QoS resource reservation request. Other embodiments are described.

Abstract: The embodiments described herein present methods and apparatuses for on-path CAC operations in a MPLS-VPN environment. In an example embodiment, an ingress PE device receives a quality of service (QoS) resource reservation request; constructs an outgoing message that includes information allowing an egress PE device to identify the virtual private network routing and forwarding table (VRF) associated with a resource reservation resulting from the QoS resource reservation request; and transmits the outgoing message to the egress PE device, where the information allowing VRF identification is echoed back by the egress PE device and used by the ingress PE device to identify the VRF associated with the resource reservation resulting from the QoS resource reservation request. Other embodiments are described.

Abstract: The invention relates to pharmaceutical compositions and food products containing peptides derived from the ?s1-casein in milk and having benzodiazepine-type activity and particularly, anxiolytic activity.

Abstract: A method of fabricating a martensitic stainless steel including: 1) heating steel to a temperature higher than austenizing temperature of the steel, then quenching the steel until a hottest portion of the steel is at a temperature less than or equal to a maximum temperature, and greater than or equal to a minimum temperature, a cooling rate being sufficiently fast for austenite not to transform into a ferrito-perlitic structure; 2) performing a first anneal followed by cooling until the hottest portion of the steel is at a temperature less than or equal to the maximum temperature and greater than or equal to the minimum temperature; 3) performing a second anneal followed by cooling to ambient temperature; and at the end of each of 1) and 2), performing: ?) as soon as temperature of the hottest portion of the steel reaches the maximum temperature, immediately heating the steel once more.

Abstract: The invention relates to pharmaceutical compositions and food products containing peptides derived from the ?s1-casein in milk and having benzodiazepine-type activity and particularly, anxiolytic activity.

Abstract: A device for polishing centrifugal impellers for a turbomachine including a vat configured to be filled with a polishing agent, and an impeller support configured to make the impeller rotate around its axis and move it along its axis such that all of points of the impeller have a helical movement whereof the pitch is close to that of the helix from which the general shape of the airflow channels of the impeller comes, delimited by the blades of the impeller.

Abstract: Each network node having at least one destination-oriented link toward a directed acyclic graph (DAG) destination can receive a corresponding set of path performance metrics via the destination-oriented link. The set of path performance metrics, initiated by the DAG destination outputting initial link metrics on each of its source-connecting links, identifies aggregate link metrics for a corresponding path to the DAG destination via the corresponding destination-oriented link. The network node outputs a corresponding updated set of path performance metrics on each of its source-connecting links based on the received set of path performance metrics and the corresponding link metric for the corresponding source-connecting link. Hence, each network node in the DAG can assess the performance of each connected path to the DAG destination, and forward a data packet via a selected destination-oriented link based on the corresponding path performance metrics and forwarding policies for the forwarded data packet.

Abstract: The present invention relates to novel compositions having (a) at least one pregelatinized starch and (b) at least one cook-up starch or flour, and which develop upon mixing with water a low cold viscosity, develop no or a low viscosity during the pasteurisation step and develop a significantly higher viscosity at a further heating step.