Qi Tang Zhang

20220383202, Dec 1, 2022

May 26, 2021

17/331067

- Armonk NY, US
Qi Zhang - West Harrison NY, US
Petr Novotny - Mount Kisco NY, US

G06N 20/20

A contribution evaluation method, system, and computer program product that evaluates the contribution of each participant-computing device in the federated learning scheme based upon the quality of model updates received from the participant, measured by the accuracy improvement of FL model with applying the participant's model updates.

20220269977, Aug 25, 2022

Feb 22, 2021

17/180972

- Armonk NY, US
Qi Zhang - West Harrison NY, US
Petr Novotny - Mount Kisco NY, US
Taesung Lee - Ridgefield CT, US

G06N 20/00
H04L 29/06

Mechanisms for performing intelligent federated machine learning (ML) model updates are provided. A plurality of ML model updates, and a plurality of dataset sketch commitment data structures (sketches), are received from a plurality of participant computing systems. Each sketch provides statistical characteristics of a corresponding local dataset used by a corresponding participant to train a local ML model. A potentially malicious participant identification operation is performed based on an analysis of the plurality of sketches to identify one or more potentially malicious participants based on differences in sketches. ML model updates received from participant computing systems identified as potentially malicious participants are discarded to thereby generate a modified set of updates. The federated ML computer model is updated based on the modified set of updates.

20220255990, Aug 11, 2022

Feb 5, 2021

17/168917

- Armonk NY, US
Qi Zhang - West Harrison NY, US
Petr Novotny - Mount Kisco NY, US
Nitin Gaur - Round Rock TX, US

H04L 29/08

A processor may record one or more transaction proposals regardless of validity of the one or more transaction proposals. The processor may separate a transaction validation into a series of steps. The processor may validate, asynchronously, one or more transactions respectively associated with each of the one or more transaction proposals based on the transaction validation.

20220188295, Jun 16, 2022

Dec 14, 2020

17/120603

- Armonk NY, US
Qi Zhang - West Harrison NY, US
Lei Yu - SLEEPY HOLLOW NY, US
Nitin Gaur - Round Rock TX, US

G06F 16/23

A processor may define an available resource set in the blockchain network. The available resource set may be the one or more peers. The processor may collect one or more metrics associated with the one or more peers in the blockchain network. The processor may analyze the one or more metrics and may identify a first workload level for the one or more peers. The processor may determine an optimal status for a first particular peer of the one or more peers, based in part on the available resource set and the first workload level. The processor may compare the optimal status to a current status of the first particular peer. The processor may determine if the optimal status and the current status are different. The processor may execute a status change of the first particular peer from the current status to the optimal status.

20230019920, Jan 19, 2023

Jul 14, 2021

17/375185

- Armonk NY, US
Yohei Ueda - Shinagawa, JP
Tatsuhiro Chiba - Bunkyo-ku, JP
Marcelo Carneiro Do Amaral - Tokyo, JP
Sunyanan Choochotkaew - Koto, JP
Qi Zhang - West Harrison NY, US

G06F 9/46

A computer-implemented method for detecting bottlenecks in microservice cloud systems is provided including identifying a plurality of nodes within one or more clusters associated with a plurality of containers, collecting thread profiles and network connectivity data by periodically dumping stacks of threads and identifying network connectivity status of one or more containers of the plurality of containers, classifying the stacks of threads based on a plurality of thread states, constructing a microservice dependency graph from the network connectivity data, aligning the plurality of nodes to bar graphs to depict an average number of working threads in a corresponding microservice, and generating, on a display, an illustration outlining the plurality of thread states, each of the plurality of thread states having a different representation.

20210398012, Dec 23, 2021

Jun 17, 2020

16/903872

- Armonk NY, US
Qi Zhang - Elmsford NY, US
Lei Yu - Sleepy Hollow NY, US
Hong Min - Poughkeepsie NY, US

G06N 20/00
G06F 16/242
G06F 16/28
G06F 16/21

In accordance with an embodiment of the invention, a method is provided for performing data pre-processing operations during data preparation of a machine learning lifecycle. The method includes defining one or more data pre-processing functions for applying to data stored in a dataset, executing one or more learn functions for learning the data, and executing one or more transform functions for transforming the data. Each of the one or more learn functions generates a first Structured Query Language (SQL) statement representing a definition of corresponding learn function for corresponding defined data pre-processing function. Each of the one or more transform functions generates a second SQL statement representing a definition of corresponding transform function for corresponding defined data pre-processing function. The dataset is stored in a database.

20210240673, Aug 5, 2021

Feb 5, 2020

16/783155

- Armonk NY, US
Qi Zhang - Elmsford NY, US

G06F 16/182
H04L 29/06
G06F 16/9035
H04L 9/06

An example operation may include one or more of transmitting an inquiry to a blockchain peer, where the inquiry comprises a trigger to invoke a chaincode installed on the blockchain peer, receiving a response from the blockchain peer based on a chaincode execution of the blockchain peer, determining a latency value of the blockchain peer based on the response, and storing an identifier of the blockchain peer among identifiers of a group of blockchain peers, where an order of the identifiers of the group of blockchain peers are arranged based on respective latency values.

20210226800, Jul 22, 2021

Jan 20, 2020

16/747496

- Armonk NY, US
Yuan Yuan - Singapore, SG
Qi Zhang - Elmsford NY, US

H04L 9/32
H04L 29/06
H04L 9/30
H04L 9/06

An example operation may include one or more of storing a first hashed timelock request in a first storage structure, where the first hashed timelock request is hashed based on a first secret, generating a second secret based on the first secret and a public key of a client, hashing the second secret to generate a hashed second secret, and transmitting a request for a second hashed timelock request to the client, where the request comprises the generated hashed second secret.