Wen-Syan Li has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
Abstract: A system and method for analyzing documents, such as posts, on-line reviews and comments from people based on topics of the documents, to determine general sentiment of users is disclosed. Topics from the documents and their corresponding sentiment polarities are extracted. The documents are regarded to be constituted by a series of topics. The sentiment for a topic is represented by a quadruple (k, so, h, i), where k is the topic, so is the sentiment opinion, h is the comment or post holder, and i is the document. A quintuple (k, sup, p, n, ne) is used to illustrate the topics and corresponding sentiments and is stored in S, where sup indicates the frequency of the topic, and p (positive), n (negative) and ne (neutral) are different types of opinions of the users. From the quintuple set S, every topic is related to three kinds of sentiment opinions (positive, negative, and neutral), enabling determination of popular topics in documents as well as the users' sentiment polarities.
Abstract: Disclosed herein are technologies for providing recommendations as to particular products and/or services that are customer specific and general, based on customer preference and inquiry. The recommendations are provided as part of an online shopping system. In accordance with one aspect, an item query is received from a customer, and analyzed by a query analyzer to determine if the query is a general item query or a specific item query. A search may be performed for items based on the item query in an items database listing items offered for purchase. If the query is the general item query, customer preference is determined from results of the search. If the query is the specific item query, the items from the results of the search are grouped based on cost performance. The items of the search result are ranked and provided to the customer.
January 14, 2015
Date of Patent:
May 21, 2019
Zhan Shi, Yu Cheng, Gufei Sun, Wen-Syan Li
Abstract: A technology for multi-objective optimization is provided. In accordance with one aspect, historical data and real-time data of a network of interest are provided in an integrated database. User input containing a problem to be solved and user preference for solving the problem is received. An optimization function is identified for generating at least one solution to the problem. The optimization function is identified based on the user preference. In response to identifying a single objective optimization function, the optimization function is initiated from a single objective optimization function library. In response to identifying a multi-objective optimization function, the optimization function is initiated from a multi-objective optimization function library. The optimization function retrieves the historical and real-time data based on the user preference for solving the problem. A result comprising at least one solution to the problem is provided.
Abstract: Systems and methods are provided for tracking parts to be used in a production line. First and second digital outputs that represent, respectively, a first physical property of a first part and a second physical property of a second part, are received by one or more data processors. The first part has a first expected delivery date, and the second part has a second expected delivery date. The first and second physical properties are independently selected from the group consisting of humidity, temperature, and shock. The first digital output is compared to a first predetermined value representing damage to the first part. After a determination that first part is damaged and will not be available on the first expected delivery date, an alert is generated. The one or more data processors output the alert to at least one of a display screen and a computer-readable medium.
Abstract: Disclosed herein are technologies for facilitating placement of charging stations. In accordance with one aspect, sensor data of electric vehicles is received. A charging demand distribution over a set of locations is determined based on the sensor data. Candidate locations are selected from the set of locations based on the charging demand distribution. Placement of charging stations at one or more of the candidate locations is then optimized.
Abstract: Disclosed herein is a framework for determining a service provider to service a customer request. In accordance with one aspect, a request from a customer is received by a decision module. The decision module further receives a response to the request from a first service provider, determines the average frequency of responses for a given time interval during the time and location of the request, estimates arrival time of the first service provider and near-by service providers to the location of the customer, ranks all service providers based on arrival time, and estimates a probability of a shorter customer wait time.
Abstract: A framework for intelligent memory replacement of loaded data blocks by requested data blocks is provided. For example, various factors are taken into account to optimize the selection of loaded data blocks to be discarded from the memory, in favor of the requested data blocks to be loaded into the memory. In some implementations, correlations between the requested data blocks and the loaded data blocks are used to determine which of the loaded data blocks may become candidates to be discarded from memory.
November 21, 2014
Date of Patent:
September 4, 2018
Nairu Fan, Tianyu Luwang, Conglun Yao, Wen-Syan Li
Abstract: Techniques are described for forming a machine learning model vector, or just model vector, that represents a weighted combination of machine learning models, each associated with a corresponding feature set and parameterized by corresponding model parameters. A model vector generator generates such a model vector for executing automated machine learning with respect to historical data, including generating the model vector through an iterative selection of values for a feature vector, a weighted model vector, and a parameter vector that comprise the model vector. Accordingly, the various benefits of known and future machine learning algorithms are provided in a fast, effective, and efficient manner, which is highly adaptable to many different types of use cases.
Abstract: The embodiments provide a system for decision-making criteria-based recommendations. The system may include a decision engine configured to receive a request for recommendations for an option problem associated with a product or service category, and determine options among a plurality of options for the product or service category based on preference information. The preference information may include activated decision-making criteria and corresponding weight values. The corresponding weight values may represent a relative importance of each activated decision-making criterion. The decision engine may be configured to determine options among the plurality of options including calculating scores for the plurality of options based on, in part, the activated decision-making criteria and the corresponding weight values and selecting the options among the plurality of options based on the calculated scores.
Abstract: A method of manufacturing at least a first product and a second product with at least a first machine and a second machine at minimum cost in an environment in which a cost of energy used by the first machine and the second machine varies as a function of time may include generating multiple chromosomes, determining fitness scores of each of the chromosomes, randomly generating, with probabilities based on the fitness scores, new chromosomes, determining fitness scores of the new chromosomes, selecting one of the new chromosomes with an optimal fitness score, and manufacturing at least the first product and the second product with at least the first machine and the second machine according to a schedule based on the selected new chromosome.
Abstract: In accordance with aspects of the disclosure, systems and methods are provided for managing production resources including scheduling production events for production resources used to manufacture products relative to time intervals while maintaining collaboration among the production resources. The systems and methods may include retrieving information related to each production resource, evaluating each production event for each product to determine a sequence of the production events, and generating potential production scheduling schemes for use of each production resource within the time intervals while maintaining collaboration among the production resources. The systems and methods may include generating a production schedule for the production events within the time intervals based on the potential production scheduling schemes for use of each production resource within the time intervals while maintaining collaboration among the production resources.
Abstract: The present description provides data analysis to provide highly individualized recommendations for users. The recommendations received by a user will include a combination of frequent recommendations and occasional recommendations. The frequent recommendations refer to recommendations made for the same product or type of product repetitively across multiple sets of recommendations, while the occasional recommendations recommend a given product or type of product only rarely across the multiple sets of recommendations. The described techniques and systems automatically determine, through data analysis, which products should be considered frequent and which should be occasional, as well as which products should not be recommended at all (blacklisted). Moreover, when generating a given recommendation set, the described techniques and systems automatically determine an optimal or near-optimal ratio of frequent/occasional recommendations.
Abstract: A candidate position selector determines locations for M candidate positions along pathways within an area defining potential routes for moving objects within the area. A coverage rate calculator calculates a coverage rate for a device layout of N media capture devices among the M candidate positions, including calculating the coverage rate as a ratio of the moving objects captured by the N media capture devices in the device layout to a total number of the moving objects in the area within a defined period of time. A layout selector iteratively tests different device layouts of the N media capture devices among the M candidate positions, based on corresponding coverage rates calculated by the coverage rate calculator for each tested device layout, until a termination condition is reached and a final device layout of the tested device layouts is selected.
Abstract: A query analyzer may be used to identify a plurality of queries and associated query characteristics, the query characteristics including a corresponding dataset of data units for each query, wherein the data units are to be loaded from an offline memory into a memory of an in-memory database. Then, a sequencer may be used to select a selected dataset for loading into the memory of the in-memory database, based on dependencies between the plurality of queries as included within the query characteristics, and on relationships, if any, between the datasets.
July 10, 2014
Date of Patent:
May 29, 2018
Leo Hu, Yuan Fang, Yun Jin, Wen-Syan Li
Abstract: The present disclosure involves systems, software, and computer implemented methods for mutual reinforcement of edge devices with dynamic triggering conditions and/or computation frequencies. In one example, a first edge device in multiple edge devices in an Internet of Things (IoT) system monitors at least a first parameter. The first edge device determines whether a first condition from one set of conditions is satisfied based on at least the monitored first parameter. In response to determining that the first condition is satisfied, the first edge device automatically transmits a signal to a second edge device in the multiple edge devices to initiate or stop monitoring of a second parameter by the second edge device.
Abstract: The present description provides data analysis for machine maintenance scheduling. For example, dynamic maintenance intervals are assigned for each machine being scheduled. Then, a system is provided for implementing a particle swarm optimization for finding an optimized maintenance schedule. In the optimization, an objective function is defined for maximizing production while minimizing relative maintenance cost.
Abstract: Software techniques are described for optimizing resource assignments among assignees of those resources, by using dynamic priority updating over a plurality of rounds of resource assignment. In particular, in example implementations, a greedy algorithm is used to optimize the resource assignments over multiple rounds based on ranked selections within each selection set of each assignee, including dynamically updating the priority of the assignees at each round, based on assignment results from one or more preceding assignment rounds.
Abstract: A computer system for controlling operation of a vehicle travelling from a starting location to a target location includes a memory and a semiconductor-based processor forming one or more logic circuits configured to generate a travel itinerary by selecting different highway segments on a map for the vehicle to drive over at respective speeds from the starting location to the target location subject to an arrival time deadline for arriving at the target location. Each of the highway segments has a distance that can be driven at the respective speed in less than a continuous driving time limit. The computer system sends the travel itinerary to a vehicle operator to drive the vehicle over the different highway segments with instructions to make a stop at the end of each highway segment to stay below the continuous driving time limit.
Abstract: Techniques are provided for rendering network applications in a highly-customized manner, in which, for example, user interactions with one or more network applications using devices having different screen sizes are analyzed and used to assign user preferences and priorities with respect to the one or more network application(s). In this way, users may be provided with desired and useful content in a convenient manner, while application providers may have their content rendered in a manner that increases a likelihood of achieving an intended result (e.g., consummating a sale or other transaction, or eliciting some other desired reaction from the user).
Abstract: Techniques for managing display of content include identifying a document to present to a user on a graphical user interface (GUI); extracting at least one property of the document, the property including metadata associated with the document; determining a correlation between the document and one or more document topics associated with the user; determining an amount of time to display a portion of the document to the user through the GUI based at least in part on the property and the correlation; and displaying the portion of the document to the user through the GUI for the determined amount of time.