Hao H Liu

US Patent:

7933283, Apr 26, 2011

Filed:

Mar 4, 2008

Appl. No.:

12/073317

Inventors:

Hao Howard Liu - Milpitas CA, US
Eugene W. Lee - San Jose CA, US

Assignee:

Cortina Systems, Inc. - Sunnyvale CA

International Classification:

H04L 12/54
G06F 15/167

US Classification:

370412, 370428, 709213

Abstract:

Shared memory management apparatus and techniques are disclosed. A shared memory provides queues for storing communication traffic that is to be scheduled for transmission. Based on amounts of communication traffic stored in the queues, a determination is made as to whether or not communication traffic is to be removed from the shared memory, and if so, communication traffic is removed. Traffic may be selected for removal based on one or more of: a scheduling algorithm for scheduling communication traffic out of the queues for transmission, thresholds associated with the queues, and latency of the communication traffic stored in the queues. Removal of communication traffic from the shared memory is coordinated with scheduling of communication traffic in some embodiments.

US Patent:

20200234113, Jul 23, 2020

Filed:

Apr 25, 2019

Appl. No.:

16/395083

Inventors:

- San Francisco CA, US
Hao LIU - Palo Alto CA, US

International Classification:

G06N 3/08

Abstract:

A method for deep reinforcement learning using a neural network model includes receiving a distribution including a plurality of related tasks. Parameters for the reinforcement learning neural network model is trained based on gradient estimation associated with the parameters using samples associated with the plurality of related tasks. Control variates are incorporated into the gradient estimation by automatic differentiation.

US Patent:

20200175632, Jun 4, 2020

Filed:

Nov 30, 2018

Appl. No.:

16/207004

Inventors:

- San Francisco CA, US
Hao Yu Liu - Union City CA, US
Benjamin Han - San Francisco CA, US
Julia Yu - San Francisco CA, US
Le Guan - San Francisco CA, US
Xiaoyuan Xu - San Mateo CA, US
Mayank Gulati - San Francisco CA, US
Charles Parker Spielman - San Francisco CA, US
Chirag Chhagan Chheda - San Francisco CA, US
David Chouinard - San Francisco CA, US

International Classification:

G06Q 50/30
G06Q 10/06

Abstract:

The disclosed computer-implemented method may include dynamically selecting transportation options to present to a transportation requestor device based on current transportation network conditions and transportation requestor device history. In some embodiments, transportation network may have many different ways of arranging a transportation requestor's trip, such as private rides, shared rides, immediate rides, and delayed rides. In some examples, the requestor's choice of transportation option may have an impact on the transportation network. In anticipation of or in response to a transportation request, the method may determine which transportation options will better benefit the transportation network and determine which transportation options to display to the requestor and/or the prominence with which the transportation products are displayed. Various other methods, systems, and computer-readable media are also disclosed.

US Patent:

20200104404, Apr 2, 2020

Filed:

Sep 29, 2018

Appl. No.:

16/147737

Inventors:

- Redmond WA, US
Lanhui Long - Sunnyvale CA, US
Haochen Wei - Sunnyvale CA, US
Yiheng Wang - Sunnyvale CA, US
Hao Liu - Mountain View CA, US
Sourav Maji - Sunnyvale CA, US
Cindy Chen - Sunnyvale CA, US

International Classification:

G06F 17/30
H04L 29/08
G06F 9/54

Abstract:

Techniques for migrating clients from one technology stack to another are provided. In one technique, while a legacy service is hosted that is actively serving requests from multiple clients, a new service is initiated and one or more clients send requests to the new service. The legacy service reads data from and writes data to a legacy database in response to the requests. The new service forwards, to the legacy service, a first set of client requests that were directed to the new service. A new database is synchronized with the legacy database. After synchronization, the legacy service forwards, to the new service, a second set of client requests, which the new service processes.

US Patent:

20180349947, Dec 6, 2018

Filed:

Jun 2, 2017

Appl. No.:

15/612586

Inventors:

- Redmond WA, US
Xinruo Jing - Foster City CA, US
Luke Kopakowski - San Francisco CA, US
Hao Liu - Sunnyvale CA, US

International Classification:

G06Q 30/02

Abstract:

Techniques for avoiding race conditions in a content delivery system are provided. Multiple content delivery campaigns that are affected by at least one remapping in a set of remappings are identified. Each remapping in the set of remappings maps an old identifier to a new identifier that is different than the old identifier. For at least some of the plurality of content delivery campaigns, identification data that identifies the content delivery campaign is sent to a particular service. In response to the particular service receiving the identification data, the particular service retrieves targeting data associated with the content delivery campaign. The particular service applies a subset of the set of remappings to the targeting data to generate updated targeting data. The particular service causes the updated targeting data to be stored in association with the content delivery campaign.

US Patent:

20180159924, Jun 7, 2018

Filed:

Dec 6, 2016

Appl. No.:

15/370190

Inventors:

- Redmond WA, US
Hao Liu - Sunnyvale CA, US
Rongsheng Liang - Cupertino CA, US
Jianhong Fang - Milpitas CA, US
Shangcheng Ying - Shanghai, CN
Hailin Wu - Sunnyvale CA, US

International Classification:

H04L 29/08
H04L 12/26

Abstract:

Techniques for assigning entities to data centers are provided. In one technique, usage data associated with a particular entity that is assigned to a first data center is generated. The usage data indicates activity of multiple users relative to the particular entity. It is determined, based on the usage data, whether the particular entity is to be assigned to another data center. If so, then the particular entity is assigned to a second data center.

US Patent:

20180159927, Jun 7, 2018

Filed:

Dec 6, 2016

Appl. No.:

15/370143

Inventors:

- Redmond WA, US
Yingfeng Oh - Cupertino CA, US
Hao Liu - Sunnyvale CA, US
Dominic Law - Sunnyvale CA, US
Xiaokang Zhang - Milpitas CA, US

International Classification:

H04L 29/08

Abstract:

Techniques for handling the dynamic reassignment of entities among multiple data centers are provided. In one technique, for each task of a plurality of tasks that are associated with a first data center of a plurality of data centers, an entity identifier that is associated with the task is identified. A call to an entity routing service is initiated, where the call includes the entity A response to the call is received from the entity routing service, where the response indicates a particular data center. An entity processor that is different than the entity routing service processes the task only if the particular data center is the first data center. If the particular data center is different than the first data center, then a data structure that stores the task is updated to indicate the particular data center and the task is eventually processed in the particular data center.

US Patent:

20180139273, May 17, 2018

Filed:

Nov 14, 2016

Appl. No.:

15/351028

Inventors:

- Sunnyvale CA, US
Joseph Florencio - San Jose CA, US
Timothy Jack Showalter - Newark CA, US
Alan Choi - San Francisco CA, US
Rongsheng Liang - Cupertino CA, US
Hailin Wu - Sunnyvale CA, US
Hao Liu - Sunnyvale CA, US
Jianhong Fang - Milpitas CA, US
Xiao Bao - Mountain View CA, US
Mihir Gandhi - Sunnyvale CA, US
Yiwen Sun - Santa Clara CA, US

International Classification:

H04L 29/08
H04L 12/721
H04L 12/733

Abstract:

In some embodiments, the disclosed subject matter involves an entity routing service to route user requests for an application service to a particular data center based on the user's entity status. The user's entity status is defined by at least the application service requested, and may include the user's organization, geographic area and other criteria. The routing may be effected at the frontend application server level rather than at the backend. Other embodiments are described and claimed.