Abstract: Described herein are embodiments of a neural network-based task planner (TaskNet) for autonomous vehicle. Given a high-level task, the TaskNet planner decomposes it into a sequence of sub-tasks, each of which is further decomposed into task primitives with specifications. TaskNet comprises a first model for predicating the global sequence of working area to cover large terrain, and a second model for determining local operation order and specifications for each operation. The neural models may include convolutional layers for extracting features from grid map-based environment representation, and fully connected layers to combine extracted features with past sequences and predict the next sub-task or task primitive. Embodiments of the TaskNet are trained using an excavation trace generator and evaluate its performance using a 3D physically-based terrain and excavator simulator.

Abstract: The invention relates to environment management, particularly to methods for purifying a wastewater in order to climate a toxic impact of heavy and radioactive metals. A method for purifying fresh, combined and saline wastewater from radioactive and heavy metals using an electrolysis and a special active substance (sorbent), wherein the wastewater is fed to an electrolyzer with a chamber that is separated by a special membrane that is permeable for ions of metals separately of water, then changes of the pH occur in order to form complex compounds, which comprise ions of radioactive and heavy metals. Afterwards, the adsorption of the obtained compounds by the special active substance (sorbent) and filtering-off on a precoat filter that retains ions of heavy and radioactive metals are performed. The obtained filtrate is cemented without drying and evaporation in order to perform final deposition of the radioactive compounds.

Abstract: A multipurpose accessory and storage system for an electronic device includes a housing member attaching to the electronic device and a charging station structured on the housing member to charge, arrest and dispense an accessory item. The accessory item includes a protective membrane housing a logic board, a power component, an image sensor, a transmitter configured to transfer a signal or data to the electronic device or the housing member and a software program stored in the logic board. The transmitter establishes a pairing or electronic connection between the electronic device or the housing member and the accessory item. The pairing or electronic connection enables content captured by the image sensor to be relayed to the electronic device to permit a user to view the content on the display of the electronic device and/or to control components and features of the accessory item via the electronic device.

Abstract: A system and method to distribute pesticides, fertilizers, water, and other materials on a farm with accuracy and precision is disclosed in order to combat the problems imposed on the environment due to over-fertilization and over use of pesticides. This system and method is a social networking control system in which multiple farms have independent grids of sensors capable of detecting the presence of pesticides, fertilizers, water, and other materials in the air, in the top-soil, and in the groundwater. These grids of sensors detect the location and concentration of these materials and reports them back to a social control system for analysis. The control system regulates the deposition of further chemicals through computer control of the chemical dispersal systems.

Abstract: A pumping system is configured to be deployed in a well that has a vertical portion and a lateral portion. The pumping system includes a pump positioned in the vertical portion, a velocity tube assembly that extends from the vertical portion into the lateral portion and a multiphase diverter connected between the pump and the velocity tube assembly. The multiphase diverter includes a housing and a plurality of ejection ports that extend through the housing at a downward angle.

Abstract: In at least some implementations, an assembly includes a reservoir, a primary fuel pump having an inlet in communication with the reservoir's internal volume, an outlet, a motor and a pumping element driven by the motor and a secondary fuel pump with a body having first and second inlets and an outlet. The first inlet receives fuel from the primary fuel pump and a nozzle is communicated with the first inlet and fuel flows out of the nozzle into the body via the first inlet. The second inlet is in communication with the reservoir inlet, and the outlet is in communication with the internal volume. The flow of fuel through the nozzle draws fuel from the fuel source through the second fuel inlet and that fuel is combined with the flow of fuel from the nozzle and discharged into the reservoir.

Abstract: Methods and systems for training and/or using a language selection model for use in determining a particular language of a spoken utterance captured in audio data. Features of the audio data can be processed using the trained language selection model to generate a predicted probability for each of N different languages, and a particular language selected based on the generated probabilities. Speech recognition results for the particular language can be utilized responsive to selecting the particular language of the spoken utterance. Many implementations are directed to training the language selection model utilizing tuple losses in lieu of traditional cross-entropy losses. Training the language selection model utilizing the tuple losses can result in more efficient training and/or can result in a more accurate and/or robust model—thereby mitigating erroneous language selections for spoken utterances.

Abstract: Technique for initialization of encoders and decoders. In some cases, the decoder receives a slice and identifies if the slice is either a forward predicted B-slice or a backward predicted B-slice, and not both a forward and backward predicted B-slice, and based upon this identification initializes, using a P-slice technique, a context associated with the slice.

Abstract: A backup manager for providing backup services to clients includes storage for storing a priority list. The backup manager also includes a processor programmed to obtain a request to generate a backup for a client of the clients; make a determination that the client is a high restoration priority client using the priority list; in response to the determination: obtain a raw disk backup for the client; and store the raw disk backup in backup storage in raw disk format.

Abstract: A controller device for user interactivity with a server of a cloud gaming system is provided, including the following: at least one input device that generates inputs for the server; wherein the controller device communicates directly to an access device for connection to a network that connects the controller device to the server without connecting to a client device; wherein the server receives and processes the inputs to render gameplay video that is transmitted over the network for rendering to a display device that is local to the controller device.

Abstract: A method or system for predicting failures in an earth-boring tool. Communication between one or more nodes in the earth-boring tool may be monitored. Metadata from the communication may be stored in a storage device. The metadata may be compared to historical communication metadata. The metadata may be fed into models built from historic metadata. Predictions from the models may be aggregated in to a recommendation. A failure prediction for each of the one or more nodes may be generated from the comparison.

Abstract: The present disclosure relates to laminate composite panels comprising a first layer of fiberglass reinforced polyester (“FRP”) and a second layer of a polyurethane material. The polyurethane layer can comprise a reaction product between an aromatic polyester polyether polyol and a diisocyanate. The polyurethane is cured-in-place on a surface of the FRP, thus creating a laminate composite material wherein the layers are adhered together without application of a separate adhesive layer. The panels exhibit improved water resistance and strength properties. The laminate composite can be adhered to a second material such as wood, polystyrene, polyurethane, or non-wovens. Such panels are particularly suitable for use in recreational vehicles, motor homes, trailers, and as building materials. Methods of making such laminate composite materials are also provided.

Abstract: Non-volatile memory structures are presented for a content addressable memory (CAM) that can perform in-memory search operations for both ternary and binary valued key values. Each ternary or binary valued key bit is stored in a pair of memory cells along a bit line of a NAND memory array, with the stored keys searched by applying each ternary or binary valued bit of an input key as voltage levels on a pair of word lines. The system is highly scalable. The system can also be used to perform nearest neighbor searches between stored vectors and an input vector to find stored vectors withing a specified Hamming distance of the input vector.

Abstract: Aspects of the subject disclosure may include, for example, identifying from a user device, authentication methods associated with services, obtaining network security information and identifying device security information for the authentication methods. Further embodiments include, in response to analyzing the network security information and the device security information: determining a first security risk for a first authentication method; and identifying a second authentication method with a second security risk lower than the first security risk. Additional embodiments include transmitting a notification to the user device indicating not to utilize the first authentication method based on the first security risk and to utilize the second authentication method, and transmitting a notification to network devices indicating the first security risk associated with the first authentication method.