Inventec Corp.

Inventec (pudong) Technology Corporation

Hewlett Packard Enterprise Development LP

National Taiwan University

Wei-Chao Chen

Trista Pei-Chun Chen

Chih-Fan Hsu

Ming-Ching Chang

Yu-Sheng Lin

Ying-Sheng Luo

Ming-Chi Chang

Feng-Hao Liu

Yi-Chun Chen

Jonathan Hans Soeseno

Daniel Stanley Young Tan

Jing-Lun Huang

Yu-Lun Chang

Chih-Pin Wei

Ke-Han Li

Guilherme Henrique Galelli Christmann

Ghih-Pin Wei

Yu Te Ku

Shao-Yi Chien

Lee A Preimesberger

Jorge Cisneros

Shih-Sung Lin

Shih-Hao Hung

Ke-Min Hu

Yung-Yu Chuang

Hua-Hsiu Chiang

Ke-Li Wu

Chuo-Jui Wu

Shu-Huei Yang

Siang-Yu Lan

The present disclosure provides a system including first to second electronic devices. The first electronic device includes a first processor configured to encrypt a machine learning model to generate an encrypted machine learning model. The second electronic device includes a second processor configured to receive the encrypted machine learning model; and a trusted processing circuit configured to decrypt the encrypted machine learning model to get the machine learning model. The third electronic device includes a database configured to store public keys; and a third processor configured to perform a check operation to check whether a first public key of the trusted processing circuit is in the public keys to determine whether the trusted processing circuit is valid. The first processor is further configured to transmit the encrypted machine learning model to the second processor according to the trusted processing circuit determined valid.

System and method for executing machine learning model inference

A federated learning method using synonym includes: sending a general model to every client device by a moderator; performing a training procedure by each client device, wherein the training procedure includes: encoding private data into a digest; training a client model according to the private data, the digest and the general model, and sending the digest and a client parameter of the client model to the moderator; determining an absent client device of the plurality of client devices by the moderator; generating a synonym of the digest corresponding to the absent client device by a synonym generator; training a replacement model according to the synonym and the digest corresponding to the absent client device by the moderator; and performing an aggregation to update the general model by the moderator according to the replacement model and the client parameter of every client devices except the absent client device.

Federated learning method using synonym

An animation processing method performed a computing device includes: obtaining a character file configured to set a character in an animation, setting a plurality of keyframes in a plurality of frames of the animation, generating a motion file according to the character file and the plurality of keyframes, where the motion file is configured to specify motion information of the character for each frame, loading the character file into a simulator to create the character, and performing a training process based on deep reinforcement learning. The training process includes performing a plurality of actions by the character in an environment set by the simulator and collecting a plurality of data points when the character performs motions; and training a policy model according to the training dataset, where the policy model outputs one of the plurality of motions according to a state of the character and the motion information.

Animation processing method based on physical characteristics

A system for verifying edited image includes: a producer terminal device configured to tile a source image for a plurality of smaller tiled images with individual source image hash values to accordingly calculate an integrated source image hash value, and to execute digitally signing to generate an image tag pair; an editor terminal device configured to receive the image tag pair, to divide the source image into these smaller tiled images according to a tile configuration, to edit part of the smaller tiled images, to include the rest part of these smaller tiled images to generate an edited integral image and further calculate an integrated edit image hash value, and to execute digitally signing to generate a zero-knowledge proof (ZKP) assurance; and, a user terminal device configured to receive the ZKP assurance to accordingly verify whether or not the edited integral image is generated by editing the source image.

System for verifying edited image

A federated learning system using synonym includes a moderator and a plurality of clients. Each client includes an encoder encoding private data into a digest. The client trains a client model according to the private data, the digest and a general model, and sends the digest and a client parameter of the client model to the moderator. The moderator communicably connected to every client includes a synonym generator, sends the general model to every client, and determines an absent client of the plurality of client. The synonym generator generates a synonym of the digest corresponding to the absent client. The moderator trains a replacement model according to the synonym and the digest corresponding to the absent client device, performs an aggregation to update the general model according to a replacement parameter of the replacement model and the client parameter of every client except the absent client.

Federated learning system using synonym

A system for judging input mode of form data is configured to extract a study-use field information and a study-use volume of time stamp from each form data having a ground truth of being manually input or automatically input to accordingly execute a learning calculation to generate a judgment calculation model, further to extract an under-judged field information and an under-judged volume of time stamp from each under-judged form data without the ground truth to accordingly generate a judgment result for predicting that the under-judged form data is manually input or automatically input, and further to define judgment-abnormal form data and a trace back ground truth to re-execute the learning calculation for revising the judgment calculation model when the judgment result does not comply with a feedback ground truth.

System for judging input mode of form data

A data processing method applied in a data center is provided. The data processing method includes obtaining stored procedures of a development environment, a data quality assurance system environment and a production environment and column information of all tables, for each stored procedure, analyzing the stored procedure and determining whether the stored procedure contains a wildcard, converting and restoring the stored procedure into a restored stored procedure with complete column information based on the wildcard, generating column lineage graphs of the development environment, the data quality assurance system environment and the production environment respectively according to the restored stored procedures of the development environment, the data quality assurance system environment and the production environment, comparing the column lineage graphs of the development environment, the data quality assurance system environment and the production environment to generate a comparison result and performing a notification function according to the comparison result.

Data processing method and data processing device

The federated learning system includes a moderator and client devices. Each client device performs a method for verifying model update as follows: receiving a hash function and a general model; training a client model according to the general model and raw data; calculating a difference as an update parameter between the general model and the client model, sending the update parameter to the moderator; inputting the update parameter to the hash function to generate a hash value; sending the hash value to other client devices, and receiving other hash values; summing all the hash values to generate a trust value; receiving an aggregation parameter calculated according to the update parameters; inputting the aggregation parameter to the hash function to generate a to-be-verified value; and updating the client model according to the aggregation parameter when the to-be-verified value equals the trust value.

Method for verifying model update

The present disclosure provides a privacy computing method based on homomorphic encryption, which includes steps as follows. The ciphertext data is received, where the ciphertext data has a floating-point homomorphic encryption data structure, and the floating-point homomorphic encryption data structure of the ciphertext data includes the ciphertext mantissa, exponent parameter and gain parameter. The gain parameter sets the precision of the floating point corresponding to the ciphertext mantissa. The exponent parameter is adapted to multiplication or division. The artificial intelligence model performs operations on the ciphertext data to return the ciphertext result.

Privacy computing method based on homomorphic encryption

A method for permission management includes: generating a plurality of job roles with different permissions according to organization permission table; generating first permission structure directed graph according to the job roles; selecting one of the job roles in first permission structure directed graph as target job role; generating minimum directed spanning graph in first permission structure directed graph according to target job role; determining whether permission of each of the job roles in first permission structure directed graph matches job of each of the job roles in first permission structure directed graph; and adjusting permission and job of each of the job roles to generate second permission structure directed graph if it is determined that permission of each of the job roles in first permission structure directed graph does not match job of each of the job roles in first permission structure directed graph.

Method for permission management

A transition method of locomotion gait of a robot includes: executing a deployment procedure multiple times, each execution includes: randomly selecting a source policy and a destination policy, simulating a transition operation from the source policy to the destination policy, and recording a transition configuration and a transition result to a transition database, where each policy is a neural network model, and a latent state in the transition configuration is a hidden layer of the neural network model of the source policy. The method further includes: training a transition-net according to the transition database, and performing the following steps by a meta-controller disposed on the robot: selecting two gait policies as an active policy and a queued policy, executing the active policy, inputting the two policies to the transition-net to obtain a success probability, and when the success probability is greater than a threshold, executing the queued policy.

Transition method of locomotion gait of robot

A power consumption monitoring device includes a sensor, a storage, and a processor. The sensor is configured to detect a power-consuming device quantity and a power consumption amount. The storage is configured to store the power-consuming device quantity and the power consumption amount. The processor is communicatively connected to the sensor and the storage. The processor is configured to calculate a power-consuming device idling indicator based on the power-consuming device quantity and the power consumption amount in a monitoring time interval, wherein the power-consuming device idling indicator is used for indicating a deviation status of the power-consuming device quantity and the power consumption amount. The processor is further configured to determine whether the power-consuming device idling indicator exceeds a warning threshold. In response to the power-consuming device idling indicator exceeding the warning threshold, the processor is further configured to generate a warning message.

Power consumption monitoring device and method

A setting method of in-memory computing simulator is presented. It involves an in-memory computing device performing various test combinations of neural network models and datasets, recording the corresponding first estimation indices. A processing device then uses these test combinations to execute a simulator with adjustable settings and records the corresponding second estimation indices. The processing device calculates a correlation sum according to the first estimation indices and second estimation indices, and performs an optimal algorithm to search an optimal parameter in a setting space to maximize the correlation sum.

Setting method of in-memory computing simulator

A method for detecting and correcting cycle time includes multiple steps performed by a computing device, and these steps include: obtaining a video from a camera, obtaining a bounding box from an input device, inputting the bounding box and the video to a cycle time detection model to generate a preliminary report, wherein the bounding box is used to set a region of interest in the video, the preliminary report includes a plurality of candidate events, and each candidate event includes a start time and a candidate cycle time, receiving a revision label associated with at least one candidate event from the input device, and tuning a hyper-parameter of the cycle time detection model according to the revision label.

System and method for detecting and correcting cycle time

A processing element array includes N processing elements (PE) arranged linearly, N≥2, and an operating method of the PE array includes: performing a first data transmission procedure, where an initial value of I is 1 and the first data transmission procedure includes: operating, by an iPE, according to a first datum stored in itself, and sending the first datum to other PEs for their operations, adding 1 to I when I<N, and performing the first data transmission procedure again, performing a second data transmission procedure when I is equal to N, which includes: operating, by the JPE, according to a second datum stored in itself, and sending the second datum to other PEs for their operations, reducing J by 1 when J>1 and the (J−1)PE has the second datum, and performing the second data transmission procedure again.

Growing community of inventors

Taipei, Taiwan

Wei-Chao Chen