Abstract: A method, computer program, and computer system is provided for identifying a speaker in at text based work. Labeled and unlabeled instances corresponding to one or more speakers are extracted. Pseudo-labels are inferred for the extracted unlabeled instances based on the labeled instances. One or more of the unlabeled instances are labeled based on the inferred pseudo-labels.

Abstract: A method and apparatus comprising computer code configured to cause a processor or processors to receive a text comprising a plurality of sentences, by a machine learning model, extract a nonverbal message from one of the sentences and add an annotation to the text, the annotation indicating the nonverbal message, and output a version of the text including the annotation.

Abstract: There is included a method and apparatus comprising computer code configured to cause a processor or processors to perform generating one or more aligned inventories, wherein the one or more aligned inventories are generated using one or more word sense inventories, obtaining a word in a context sentence, determining one or more semantic equivalence scores indicating semantic similarity between the word in the context sentence and each of one or more associated glosses in the one or more aligned inventories using a semantic equivalence recognizer model, and predicting a correct sense of the word in the context sentence based on the determined one or more semantic equivalence scores.

Abstract: Described herein are systems and methods that take advantage of a self-servicing mortgage engine that utilizes a language-model-based virtual assistant with access to knowledge databases, loan product databases, and underwriting databases. The mortgage engine integrates rules-based responses with the language model to analyze user input from conversations and provide context-aware interactive guidance, e.g., in the form of easy-to-understand explanations, instructions, and suggestions tailored to user questions. The interactive guidance generates recommendations and actionable outputs for borrowers that reduce the complexities of the lending process and drives the loan application. Advantageously, this increase efficiency and transparency to the borrower, while simultaneously reducing costs to lenders.

Abstract: Described herein are systems and methods that take advantage of a self-servicing mortgage engine that utilizes a language-model-based virtual assistant with access to knowledge databases, loan product databases, and underwriting databases. The mortgage engine integrates rules-based responses with the language model to analyze user input from conversations and provide context-aware interactive guidance, e.g., in the form of easy-to-understand explanations, instructions, and suggestions tailored to user questions. The interactive guidance generates recommendations and actionable outputs for borrowers that reduce the complexities of the lending process and drives the loan application. Advantageously, this increase efficiency and transparency for the borrower, while simultaneously reducing costs to lenders.

Abstract: A method and apparatus that identifies one or more characters within a text; determines one or more informative sections within the text, the one or more informative sections providing information regarding a gender of the one or more characters within the text; selects a most informative section from the one or more informative sections; extracts unlabeled instances corresponding to the gender of the one or more characters from the most informative section; iteratively trains a multi-task model using unlabeled corpora, the multi-task model performing both speaker identification and gender identification; and labels the gender of the one or more characters based on the extracted unlabeled instances and the multi-task model.

Abstract: A method, computer program, and computer system is provided for identifying a speaker in at text based work. Labeled and unlabeled instances corresponding to one or more speakers are extracted. Pseudo-labels are inferred for the extracted unlabeled instances based on the labeled instances. One or more of the unlabeled instances are labeled based on the inferred pseudo-labels.

Abstract: A method including: receiving an input comprising natural language texts; selecting, via a knowledge selector, one of a plurality of knowledge categories from an external memory based on a context of the input; retrieving one or more helpful knowledge pieces from the selected knowledge category; augmenting the input using the one or more helpful knowledge pieces; feeding the augmented input into a text-to-text model; and generating an output answer based on the text-to-text model.

Abstract: A key structure includes a circuit board, a keycap, an elastic member and a metal member. The keycap has a bottom facing the circuit board. The elastic member is disposed between the circuit board and the keycap. The elastic member includes a dome portion, a hollow portion, and a contact pillar. The dome portion touches the bottom of the keycap. The hollow portion is located inside of the dome portion. The contact pillar is located in the hollow portion, and the contact pillar extends from the dome portion to the circuit board. The metal member covers the outside of the contact pillar. The metal member comprises a trigger portion, and the trigger portion is a convex portion extending toward the circuit board.

Abstract: A method, computer program, and computer system is provided for identifying a speaker in at text based work. Labeled and unlabeled instances corresponding to one or more speakers are extracted. Pseudo-labels are inferred for the extracted unlabeled instances based on the labeled instances. One or more of the unlabeled instances are labeled based on the inferred pseudo-labels.

Abstract: A key structure includes a circuit board, a keycap, an elastic member and a metal member. The keycap has a bottom facing the circuit board. The elastic member is disposed between the circuit board and the keycap. The elastic member includes a dome portion, a hollow portion, and a contact pillar. The dome portion touches the bottom of the keycap. The hollow portion is located inside of the dome portion. The contact pillar is located in the hollow portion, and the contact pillar extends from the dome portion to the circuit board. The metal member covers the outside of the contact pillar. The metal member comprises a trigger portion, and the trigger portion is a convex portion extending toward the circuit board.

Abstract: Systems and methods are provided for processing a request for a product. The processing the request includes receiving a preliminary request from a client device for a product including identity information of a requestor and conditions of the product without specifying a particular product. The disclosed technology generates a list of eligible products for the requester to select one product. A determiner receives a request for a product and information needed to approve a use of the product (e.g., underwriting a mortgage application). The disclosed technology automatically retrieves verified information associated with the requestor from various data providers and requests for underwriting by one or more underwriters distributed over the network. The determiner determines a result of the request for the product based on decisions made the underwriters. The client device solicits for uploading additional information and documents needed to fully approve the request for the product.

Abstract: There is included a method and apparatus comprising computer code configured to cause a processor or processors to perform generating one or more aligned inventories, wherein the one or more aligned inventories are generated using one or more word sense inventories, obtaining a word in a context sentence, determining one or more semantic equivalence scores indicating semantic similarity between the word in the context sentence and each of one or more associated glosses in the one or more aligned inventories using a semantic equivalence recognizer model, and predicting a correct sense of the word in the context sentence based on the determined one or more semantic equivalence scores.

Abstract: The present disclosure provides a method for isolating and purifying mesenchymal stem cells from hernia sac and a method for tissue repair by using the mesenchymal stem cells from hernia sac. The mesenchymal stem cells isolated and purified by the method of the present disclosure have high yield, high cell division ability, do not need to use enzymes as necessary reagents, and have high potential for application in regenerative medicine or tissue engineering for tissue repair.

Abstract: A method, computer program, and computer system is provided for identifying a speaker in at text based work. Labeled and unlabeled instances corresponding to one or more speakers are extracted. Pseudo-labels are inferred for the extracted unlabeled instances based on the labeled instances. One or more of the unlabeled instances are labeled based on the inferred pseudo-labels.

Abstract: A keyboard includes a bottom plate and keyswitch devices. Each keyswitch device includes a keycap, a first supporting member, and a second supporting member. The keycap is over the bottom plate. The first and second supporting members are connected between the bottom plate and the keycap and respectively include first and second magnetic attraction portions. The second magnetic attraction portion is configured to attract the first magnetic attraction portion. The first and second magnetic attraction portions respectively have first and second abutting surfaces. When the first abutting surface abuts against the second abutting surface, the keycap is at a highest position relative to the bottom plate. When the keycap moves toward the bottom plate from the highest position, the first abutting surface is at least partially separated from the second abutting surface.

Abstract: A keyboard includes a bottom plate and keyswitch devices. Each keyswitch device includes a keycap, a first supporting member, and a second supporting member. The keycap is over the bottom plate. The first and second supporting members are connected between the bottom plate and the keycap and respectively include first and second magnetic attraction portions. The second magnetic attraction portion is configured to attract the first magnetic attraction portion. The first and second magnetic attraction portions respectively have first and second abutting surfaces. When the first abutting surface abuts against the second abutting surface, the keycap is at a highest position relative to the bottom plate. When the keycap moves toward the bottom plate from the highest position, the first abutting surface is at least partially separated from the second abutting surface.

Abstract: The present invention is related to a method for protection of bladder from damage, and prevention of frequent urination using a Canavalia ensiformis supplement.

Abstract: The present invention provides a method of repairing a defect in an abdominal wall of a subject, including the steps of: making a small incision through a skin layer above the defect in the abdominal wall, introducing a surgical device including a deformable body through the incision to reach a preperitoneum underlying the abdominal wall around the defect, using the surgical device to identify a protruded region of the preperitoneum extending towards the abdominal wall around the defect, and flattening the protruded region of the preperitoneum and separating the protruded region of the preperitoneum from the abdominal wall with the deformable body in an expanded condition. The method results in low recurrence of abdominal wall defect, less pain, and a shorter hospital stay after surgery.

Abstract: The present invention provides a method of repairing a defect in an abdominal wall of a subject, including the steps of: making a small incision through a skin layer above the defect in the abdominal wall, introducing a surgical device including a deformable body through the incision to reach a preperitoneum underlying the abdominal wall around the defect, using the surgical device to identify a protruded region of the preperitoneum extending towards the abdominal wall around the defect, and flattening the protruded region of the preperitoneum and separating the protruded region of the preperitoneum from the abdominal wall with the deformable body in an expanded condition. The method results in low recurrence of abdominal wall defect, less pain, and a shorter hospital stay after surgery.