Abstract: The present invention is directed to a liquid, storage stable isocyanate composition (Q) comprising monomeric diphenyl methane diisocyanate (MDI) as component (q1) and an amide with a molecular weight of less than 200 g/mol as component (q2) in an amount of 0.1 to 5 wt.-% based on the sum of the weight of components (q1) and (q2) which adds up to 100 wt.-%, wherein the NCO-content is in the range of from 29.5 to 40.0 wt.-% based on the total weight of the respective isocyanate composition (Q). The present invention is also directed to a process for the manufacture of said liquid, storage stable isocyanate composition (Q) as well as the use of an amide with a molecular weight of less than 200 g/mol for the liquification of at room temperature solid isocyanates. Further, the present invention is directed to the use of an isocyanate composition (Q) of the invention for the synthesis of polyurethane foams, compact polyurethane elastomers or thermoplastic polyurethanes.

Abstract: A main winch system of a rotary drilling rig includes a main action loop, a pilot control loop, and a feedback control loop. The main action loop includes an engine, a hydraulic pump, a main control valve, a balance valve, a main winch motor, and an oil tank. The pilot control loop includes a pilot handle, a solenoid valve I, a pressure relay, and the main control valve in the main action loop. The feedback control loop includes a solenoid valve II and the hydraulic pump and the main control valve in the main action loop. In response to a lifting action of the pilot handle for the main winch, under the control of a button, the solenoid valve II is energized continuously, and the solenoid valve I is energized and de-energized according to a given rule.

Abstract: A method for pre-training a language model includes: constructing a pre-training language data set, in which the pre-training language data set comprises unsupervised language data and supervised language data; generating a hierarchical multi-template and multi-task language data set based on the pre-training language data set; and pre-training the language model based on the hierarchical multi-template and multi-task language data set.

Abstract: Methods for determining a prediction value, an encoder, and a decoder are provided. Reconstructed values of neighboring samples of a current block are acquired, and then filtered to obtain a reference value set of the current block. When a size of the current block is smaller than a preset threshold value, a first constant value is calculated according to a bit depth value of a luma component of a sample in the current block. A difference between the first constant value and a first reference value in the reference value set is determined as a first prediction input value in a prediction input value set. Other prediction input values in the prediction input value set other than the first prediction input value are determined according to the reference value set. Prediction values of samples at specific positions in the current block is calculated and then filtered.

Abstract: A multiple platform, multiple creative testing system and method implement and utilize devices and processes for managing and implementing testing of multiple creatives across multiple independent online platforms. The system and method enable a provider of creatives to perform testing of multiple creatives, such as A/B testing, across multiple online platforms with an A test targeting a control group of targeted individuals and an independent B test concurrently targeting a disjoint test group of targeted individuals across multiple online platforms. The system and method accomplish testing through concurrently using two independent targeted tests utilizing separate creatives to disjoint groups of targeted individuals across multiple online platforms.

Abstract: The present disclosure discloses a monitoring management and control system based on panoramic big data comprising an imaging device, a credential of the imaging device, a memory, a networking device, and a processing unit. The imaging device is configured to detect a first object and determine an approximate location of the first object and a reliability value of the approximate location of the first object based on the credential of the imaging device, the memory, the networking device, and the processing unit.

Abstract: A method for determining a prompt vector of a pre-trained model, includes: obtaining a first one of prompt vectors and a first vector corresponding to sample data; obtaining N pruned models by N different pruning processing on the pre-trained model, where N is any integer greater than 1; obtaining a first score corresponding to the first one of the prompt vectors by fusing the first vector and the first one of the prompt vectors and inputting the fused first vector and first one of the prompt vectors into the N pruned models respectively; determining a second one of the prompt vectors by modifying, based on the first score, the first one of the prompt vectors; and based on the second one of the prompt vectors, returning to obtaining the first score until determining a target prompt vector corresponding to the sample data.

Abstract: A knowledge distillation system and method trains neural networks utilizing a non-conventional relay buffer and augmented data tuples. In at least one embodiment, the knowledge distillation system and method pretrain a teacher model that implements a contextual bandit algorithm. A lightweight student model determines online contextual bandit data tuples as to context x, arm/action a, and reward/payoff r. The data tuples are stored in a replay buffer. The teacher model randomly samples data tuples from the replay buffer and augments the sampled data tuples. Augmented data tuples are stored in the replay buffer. The student model batch processes augmented data tuples to update parameters of contextual bandit data tuples.

Abstract: A model training system includes at least one first cluster and a second cluster communicating with the at least first cluster. The at least one first cluster is configured to acquire a sample data set, generate training data according to the sample data set, and send the training data to the second cluster; and the second cluster is configured to train a pre-trained model according to the training data sent by the at least one first cluster.

Abstract: A method, an apparatus, a device and a storage medium for learning a knowledge representation are provided. The method can include: sampling a sub-graph of a knowledge graph from a knowledge base; serializing the sub-graph of the knowledge graph to obtain a serialized text; and reading using a pre-trained language model the serialized text in an order in the sub-graph of the knowledge graph, to perform learning to obtain a knowledge representation of each word in the serialized text. The knowledge representation learning in this embodiment is performed for entity and relationship representation learning in the knowledge base.

Abstract: The present disclosure relates to the technical field of video playback, and provides a multimedia information playback method and apparatus, an electronic device, and a computer readable storage medium. The method includes: during a process of an information flow being browsed, in response to that display time of a target information entry in the information flow on a display interface of the information flow reaches a first time threshold, displaying a preset prompt message on an upper layer of the target information entry, so as to prompt a user to input a corresponding gesture; and in response to that a preset gesture operation to the target information entry inputted by the user is detected, playing multimedia information corresponding to the target information entry.

Abstract: A control method and apparatus, which are applied to a vehicle with an electric power steering system. The electric power steering system comprises a first control subsystem and a second control subsystem. The method comprises: acquiring a front steering angle, a steering torque, a yaw rate and a vehicle speed of a vehicle; according to the front steering angle, the steering torque, the yaw rate and the vehicle speed, determining a vehicle yaw rate associated with a first control subsystem; determining, on the basis of the vehicle yaw rate, an expected deviation value associated with a second control subsystem; and determining a control parameter on the basis of the expected deviation value, such that the vehicle adjusts, according to the control parameter, the steering torque to run.

Abstract: A surface inspection system for foil article is disclosed. The surface inspection system comprises a box having a top long narrow opening and a bottom long narrow opening, a bridge interface, a first light source, a second light source, a first modular camera device having a first camera, and a second modular camera device having a second camera. In which, the first light source, the second light source, the first modular camera device, and the second modular camera device all accommodated in the box, and are coupled to a control box through the bridge interface. Particularly, this surface inspection system is allowed to be integrated in an automatic production line of a foil article like electro-forming aluminum foil (also called electronic aluminum foil), so as to achieve an in-line inspection of the surface morphology of the electro-forming aluminum foil.

Abstract: The disclosure provides a text recognition method, an electronic device, and a storage medium. The method includes: obtaining N segments of a sample text; inputting each of the N segments into a preset initial language model in sequence, to obtain first text vector information corresponding to the N segments; inputting each of the N segments into the initial language model in sequence again, to obtain second text vector information corresponding to a currently input segment; in response to determining that the currently input segment has the mask, predicting the mask according to the second text vector information and the first text vector information to obtain a predicted word at a target position corresponding to the mask; training the initial language model according to an original word and the predicted word to generate a long text language model; and recognizing an input text through the long text language model.

Abstract: Example methods and devices are disclosed. One example method includes receiving, by a first device, configuration information from a second device, where the configuration information is used by the first device to report information about an association relationship between a zero power-reference signal resource and a non-zero power-reference signal resource. The first device can then report the information about the association relationship based on the configuration information.

Abstract: A method is provided. The method includes converting a search request of a user into a first request semantic vector. The method further includes searching a search resource database for at least one first data semantic vector matched with the first request semantic vector, wherein the search resource database is constructed as a semantic vector space in which different types of data are converted into corresponding data semantic vectors, and the different types of data include at least texts, pictures and videos. The method further includes generating, based on the at least one first data semantic vector, a search result.

Abstract: This invention provides preparations and applications of a highly strong and tough photo-crosslinked hydrogel material. The hydrogel materials are made from the o-nitrobenzyl phototriggers modified photosensitive polymer derivative, double bond groups modified polymer derivative, and photoinitiator. This invention significantly improved the properties of hydrogels by controlling the average molecular weight (arm length) of photosensitive poly (ethylene glycol) (with o-nitrobenzyl phototriggers modification) and its derivatives, and the molar ratio of o-nitrobenzyl phototriggers to poly (ethylene glycol) and its derivatives. Compared with prior technologies, the strength and/or toughness of the hydrogels in this invention are increased by one order of magnitude or even higher. These advances solve the problems of weak mechanical properties in current hydrogel materials and widen the applications of hydrogel materials.

Abstract: Methods for determining a prediction value, an encoder, and a decoder are provided. Reconstructed values of neighboring samples of a current block are acquired, and then filtered to obtain a reference value set of the current block. When a size of the current block is smaller than a preset threshold value, a first constant value is calculated according to a bit depth value of a luma component of a sample in the current block. A difference between the first constant value and a first reference value in the reference value set is determined as a first prediction input value in a prediction input value set. Other prediction input values in the prediction input value set other than the first prediction input value are determined according to the reference value set. Prediction values of samples at specific positions in the current block is calculated and then filtered.

Abstract: Methods for determining a prediction value, an encoder, and a decoder are provided. Reconstructed values of neighboring samples of a current block are acquired, and then filtered to obtain a reference value set of the current block. When a size of the current block is smaller than a preset threshold value, a first constant value is calculated according to a bit depth value of a luma component of a sample in the current block. A difference between the first constant value and a first reference value in the reference value set is determined as a first prediction input value in a prediction input value set. Other prediction input values in the prediction input value set other than the first prediction input value are determined according to the reference value set. Prediction values of samples at specific positions in the current block is calculated and then filtered.

Abstract: Methods for determining a prediction value, an encoder, and a decoder are provided. Reconstructed values of neighboring samples of a current block are acquired, and then filtered to obtain a reference value set of the current block. When a size of the current block is smaller than a preset threshold value, a first constant value is calculated according to a bit depth value of a luma component of a sample in the current block. A difference between the first constant value and a first reference value in the reference value set is determined as a first prediction input value in a prediction input value set. Other prediction input values in the prediction input value set other than the first prediction input value are determined according to the reference value set. Prediction values of samples at specific positions in the current block is calculated and then filtered.