会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 2. 发明授权
    • Method and apparatus for transparent backtracking
    • 用于透明回溯的方法和装置
    • US06446057B1
    • 2002-09-03
    • US09730481
    • 2000-12-04
    • Paul William Vaughan
    • Paul William Vaughan
    • G06N100
    • G06N5/003
    • A technique is used in embodiments of the invention such that backtracking programs can be written in a general purpose computer language (e.g., C++ or Java) without requiring the control structure of the program to reflect the structure of the decision tree. A data state and a control state are restored during backtracking. For restoring the data state, embodiments of the invention keep track of the changes made to variables and the point in execution at which the changes are made. When backtracking occurs, the data state can be restored by undoing the changes to the desired point in execution. For restoring the control state, the method of the invention provides a “failure” exception state that is invoked upon failure in the program (e.g., a failure to find a solution in a search program). The failure exception is “caught” by catch points established in the execution stack. The failure exception is passed up the execution stack until a point is reached prior to the failure at which execution should be re-initiated. Since the control structure of the search program need not have the same form as the decision tree, part of the control state for the desired decision point may no longer exist on the execution stack, so the catch point may not be directly associated with the desired point but merely preceed it. The remaining part of the control state is restored by re-executing the program in a special re-execution mode until the desired state is achieved and another alternative may be chosen.
    • 在本发明的实施例中使用技术,使得回溯程序可以以通用计算机语言(例如,C ++或Java)编写,而不需要程序的控制结构来反映决策树的结构。 在回溯期间恢复数据状态和控制状态。 为了恢复数据状态,本发明的实施例跟踪对变量所做的改变以及进行改变的执行点。 当发生回溯时,可以通过撤销执行中所需点的更改来恢复数据状态。 为了恢复控制状态,本发明的方法提供了在程序失败时调用的“失败”异常状态(例如,在搜索程序中找不到解决方案)。 失败异常被执行堆栈中建立的捕获点“捕获”。 故障异常被传递到执行堆栈,直到在重新启动执行的故障之前达到一个点。 由于搜索程序的控制结构不需要与决策树具有相同的形式,所以在执行堆栈上可能不再存在所需决策点的控制状态的一部分,所以捕获点可能不会与所需的 但只是前提。 通过以特殊的重新执行模式重新执行程序来恢复控制状态的剩余部分,直到达到期望的状态并且可以选择另一个替代方案。
    • 4. 发明授权
    • Workflow and exception system
    • 工作流和异常系统
    • US08341645B1
    • 2012-12-25
    • US11805593
    • 2007-05-24
    • David OxenstiernaDenise SommervilleWilliam VaughanPeter ZopeAlfred ShortJohn HendyRaju PatelLawrence Beck
    • David OxenstiernaDenise SommervilleWilliam VaughanPeter ZopeAlfred ShortJohn HendyRaju PatelLawrence Beck
    • G06F11/30
    • G06F11/0709G06F11/0793G06Q10/06
    • An exception management system comprising a centralized exception engine for processing exception messages from a plurality of source systems across a firm (or enterprise) and for generating workflow items where needed from the exception messages. The centralized exception engine may comprise an exception backbone that receives the exception messages from the source systems and processes the exception messages based on business rules. The exception messages may adhere to an exception message definition format. Users may access the exception information via a user interface. The exception engine may send alerts to certain users in certain circumstances based on the business rules. A flow manager of the exception backbone may process the exception messages by: (i) enriching the exception messages with additional information; (ii) prioritizing the exception messages; (iii) routing the exception messages to one or more users; and/or (iv) persisting exception information extracted from the exception messages in a persistent store. A workflow engine may communicate workflow messages to the source systems or directly with the exception backbone when workflow is required for one of the exception messages.
    • 一种异常管理系统,包括用于处理跨越企业(或企业)的多个源系统的异常消息并用于从异常消息生成需要的工作流项目的集中式异常引擎。 集中式异常引擎可以包括从源系统接收异常消息并基于业务规则处理异常消息的异常主干。 异常消息可能遵循异常消息定义格式。 用户可以通过用户界面访问异常信息。 异常引擎可能会在某些情况下根据业务规则向某些用户发送警报。 异常骨干的流程管理器可以通过以下方式处理异常消息:(i)使用附加信息丰富异常消息; (ii)排除异常信息的优先级; (iii)将异常消息路由到一个或多个用户; 和/或(iv)持久存储中从异常消息中提取的异常信息。 工作流引擎可以将工作流消息传送到源系统,或者在异常消息之一需要工作流时直接与异常主干通信。
    • 6. 发明授权
    • Learning method
    • 学习方法
    • US06419496B1
    • 2002-07-16
    • US09535883
    • 2000-03-28
    • William Vaughan, Jr.
    • William Vaughan, Jr.
    • G09B700
    • G09B7/02
    • A learning method to enhance long-term retention of a large number of learned facts or principles. The method presents items to a User repeatedly and, depending on the recall of the User, the time between presentations of a properly recalled item increases. Items being presented for the first time and those presented a subsequent time before being properly answered are designated Unseen Items and O-Level Items, respectively. During each Session, a raw score is assigned to the item (related to the User's ability of recall with the item). When its raw score becomes sufficiently high, the item is considered a P-Level Item, after which it is presented less frequently and the rate of presentation continually decreases as the User demonstrates that he or she knows the correct response for the particular P-Level Item. The number of Sessions that are skipped before the particular P-Level item is re-presented to the User is based on the User's Level L in recalling the correct answer to the item. In this way, known items are repeatedly presented to the User with an ever-increasing interval between presentations. If the User fails to recall the item, the Level of the item is decremented and the time to next presentation decreased. As the User is working with this method, the progress of the User is recorded in a form available for review.
    • 一种增强长期保留大量学习事实或原则的学习方法。 该方法向用户反复呈现项目,并且根据用户的回忆,正确调用的项目的呈现之间的时间增加。 第一次呈现的项目以及在适当回答之前呈现的后续时间分别被指定为“未看到的项目”和“O级项目”。 在每个会话期间,原始分数被分配给项目(与用户与项目的召回能力相关)。 当它的原始分数变得足够高时,该项目被认为是P级项目,之后它被更少地呈现,并且呈现率不断降低,因为用户证明他或她知道对于特定P级别的正确响应 项目。 在将特定P级别项目重新呈现给用户之前跳过的会话数量是基于用户级别L在调用对该项目的正确答案。 以这种方式,已知的项目在演示之间以不断增加的间隔重复地呈现给用户。 如果用户未能回收该项目,则该项目的级别递减,下一次演示的时间减少。 正如用户正在使用此方法一样,用户可以以可查看的形式记录用户的进度。