Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
2. Applicant's arguments filed 2/23/2026 have been fully considered but they are not persuasive.
I. Applicant argues that Di Fabbrizio fails to disclose the newly added limitation of amended claim 1 requiring that “the plurality of scored intent hypotheses are scored by increasing a confidence value for each scored intent hypothesis in response to determination that the corresponding scored intent hypothesis is a child of a current intent, matches a recently completed task, or matches an on-hold task.” The Examiner acknowledges that Di Fabbrizio, by itself, does not expressly disclose increasing the confidence values of the scored intent hypotheses based on the specific contextual criteria now recited in amended claim 1. Di Fabbrizio discloses receiving from the NLU module a categorization of the user request in a list of possible call types, wherein the input is “typically also associated with some kind of confidence score.” Di Fabbrizio further discloses using a confidence threshold to determine whether a node is lit, for example, “light the node if the TaxShelter intention is returned by the NLU with confidence greater than $threshold” (see Di Fabbrizio, ¶19, ¶23-24). Therefore, Di Fabbrizio teaches scored intent hypothesis and threshold-based filtering, but does not expressly disclose the newly added confidence-increasing step. Accordingly, In view of Applicant’s amendment, the previous rejection of claim 1 under 35 U.S.C. 102 over Di Fabbrizio alone has been withdrawn and replaced with a rejection under 35 U.S.C. 103. However, the amended limitation is rendered obvious by the combination of Di Fabbrizio and Bharara. Di Fabbrizio remains relied upon as the primary reference because it teaches a dialog manager that receives NLU-generated possible intents/call types with confidence scores, uses those scores to light nodes in a rooted dialog tree, disambiguates between multiple possible paths to same intent and sets a current focus node/path location (see Di Fabrizzio, ¶17-19, ¶27 and ¶36). Bharara teaches the missing rescoring aspect. Specifically, Bharara discloses calculating a confidence or correlation score for an input statement to intent mapping, wherein “the confidence score may indicate the strength of the correlation between the input statement and the intent” (see Bharara ¶40). Bharara further teaches analyzing the sequence, order or hierarchy of intents in a command tree and “changing or adjusting the confidence scores based on the results of the analysis” (see Bharara ¶99). Bharara also teaches analyzing timing and recency of prior commands and that “if a command mapping is determined to have last been used recently, then the confidence score may remain the same or increase” (see Bharara ¶100). Therefore, Bharara expressly teaches increasing or adjusting confidence scores based on hierarchy/sequence and historical recency.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed inventions to Modify Di Fabbrizio’s dialog disambiguation method to include Bharara’s sequence and history based confidence adjustment. Both references are from the same field of endeavor, namely natural language dialog system that map user utterances or commands to intents using an NLU/dialog systems that map user utterances or commands to intents using an NLU/dialog management framework. Di Fabbrizio uses automated speech recognition, natural language understanding and a dialog manager to determine and disambiguate user intent in a spoken dialog service (see Di Fabbrizio, ¶17-18). Bharara likewise concerns natural language command translation, intent determination, confidence scoring and command trees in dialog like software interfaces (see Bharara ¶36-40). The motivation to combine is provided by Bharara which teaches that confidences score adjustment using stored mappings and historical usage improves future command translation, stating that “by storing successful mappings using the previously successful mappings in future command translation, the significant correlation framework may gradually adapt over time to provide more accurate command translation” (see Bharara, ¶64).
Regarding Applicant’s argument regarding claim 6 are also unpersuasive because claim 6 remains obvious for the same reasons as amended claim 1. Di Fabbrizio further teaches that the focus node represents the current dialog state, that the focus may remain unchanged, move to a child node, or be used to disambiguate between different paths to the same intent (see Di Fabbrizio, ¶20, ¶27, ¶31 and ¶36).
Regarding Claim 9, the rejection does not rely on Di Fabbrizio alone. Di Fabbrizio teaches the confirmation/disambiguation framework, including trigger conditions, prompts, user responses and context shifts to different objective (see Di Fabbrizio ¶18 ad ¶38-40). Vibbert supplies the task and slot management features, including modifying intent slots management features, including modifying intent slots, switching tasks and suspending or resuming tasks (see Vibbert, ¶[0029], ¶[0110] and ¶[0098]). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the claimed accept and reject logic into Di Fabbrizio. Di Fabbrizio, Bharara and Vibbert are combinable because they all are directed to spoken dialog and natural language understanding systems. Bharara motivates the confidence adjustment by teaching sequence and timing based score adjustment and Vibbert motivates slot task handling by teaching confirmation to reduce ambiguity and modification of intent slots as state in Vibbert ¶85: “dialog engine 550 may continuously modify expectation agenda 570 having bound user input values to concepts or having analyzed needs for dialog focus shifts in order to continuously improve the expectations of any dialog agents or agencies using the bounded concepts or dialog focus shifts.”
Accordingly, the rejection of claim 1-20 is maintained.
Claim Rejections - 35 USC § 103
3. In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
4. Claims 1-8 are rejected under 35 U.S.C. 103 as being unpatentable over Di Fabbrizio (US 2005/0165607), in view of Bharara (US 2020/0349228) and further in view of Vibbert (US 2016/0042735).
Regarding Claim 1:
Di Fabbrizio discloses a method for selecting, by a dialog manager, a current path location of an instance of a task from a list of a plurality of scored intent hypotheses with distinct intents and different task paths (Di Fabbrizio: ¶[0018]-[0019] discloses a dialog manager, which are given call types (CTs) (i.e., intents and tasks to be performed), there is a list of possible CTs that are associated with some kind of confidence score. Each CT is associated with a path in a tree (i.e., different task paths)), the method comprising the steps of:
performing a first filtering of the list of the plurality of scored intent hypotheses to comprise hypotheses meeting a pre-determined threshold (Di Fabbrizio: ¶[0023]-[0024] specifically use a confidence threshold to decide whether a particular CT hypothesis is considered, lighting the node if a confidence is greater than a threshold is exactly filtering the list of scored hypotheses down to those meeting a pre-determined threshold (i.e., a first filtering));
from the first filtered list, determining an intent and a task path for each of the hypotheses meeting the pre-determined threshold, wherein each of the hypotheses meeting the pre-determined threshold have the same intent (Di Fabbrizio: ¶[0019] and ¶[0027] discloses that each CT is associated with a path through the tree from the root to a leaf, interpreted as a “task path.” These hypothesis have the same intent, ¶[0036] discloses “based on the user utterance, the method establishes two lit nodes 230 and 240 for tax shelter information.” Therefore, after confidence threshold lighting logic, there are sets of lit nodes (filtered hypotheses that share the same underlying intent);
determining a simplest distinct task path to each hypothesis and presenting each of the simplest distinct paths in a second list (Di Fabbrizio: ¶[0036] the algorithm moves the focus to the lowest common ancestor and presents the user with the two distinct branch choices, small business or mid to large size business, this is the same as presenting a second list of distinct path options that differentiate the hypotheses);
performing a second filtering of the second list by selecting a single distinct path (Di Fabbrizio: ¶[0036] the second list is the set presented for disambiguation (small business node, large business note), the users answer “a small business” causes the system to establish that node as the new focus node and cut the competing branch corresponding to node 240, this is a second filtering step); and
confirming the intent of the hypothesis and configuring the selected single path to the current path location (Di Fabbrizio: ¶[0036] discloses that after the user picks small business, the method eventually moves focus to the leaf tax shelter node in that branch, this is a confirmation of intent and configuration of the current path).
Di Fabbrizio does not explicitly disclose wherein the plurality of scored intent hypotheses are scored by increasing a confidence value for each scored intent hypothesis in response to determination that the corresponding scored intent hypothesis is a child of a current intent, matches a recently completed task, or matches an on-hold task.
However, Bharara discloses wherein the plurality of scored intent hypotheses are scored by increasing a confidence value for each scored intent hypothesis in response to determination that the corresponding scored intent hypothesis is a child of a current intent, matches a recently completed task, (Bharara: ¶99 discloses the sequence (or order, or hierarchy) of the intents is analyzed, ¶100 discloses the sequence of intents may include comparing the intent structure in the command tree against the sequence of structure of previously executed intents, also that it may include changing or adjusting the confidence scores based on the results of the analysis).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed inventions to modify Di Fabbrizio’s dialog disambiguation method to include Bharara’s sequence and history based confidence adjustment. Both references are from the same field of endeavor, namely natural language dialog system that map user utterances or commands to intents using an NLU/dialog systems that map user utterances or commands to intents using an NLU/dialog management framework. Di Fabbrizio uses automated speech recognition, natural language understanding and a dialog manager to determine and disambiguate user intent in a spoken dialog service (see Di Fabbrizio, ¶17-18). Bharara likewise concerns natural language command translation, intent determination, confidence scoring and command trees in dialog like software interfaces (see Bharara ¶36-40). The motivation to combine is provided by Bharara which teaches that confidences score adjustment using stored mappings and historical usage improves future command translation, stating that “by storing successful mappings using the previously successful mappings in future command translation, the significant correlation framework may gradually adapt over time to provide more accurate command translation” (see Bharara, ¶64).
The combination of Di Fabbrizio and Bharara does not explicitly disclose the current task matches an on-hold task.
However, Vibbert discloses the current task matches an on-hold task (Vibbert: ¶98 discloses the task manager is configured to create, suspend, resume or abort and exist different tasks if another task is running the engine may determine if the current task can be suspended in favor of a new task, once the newly requested task has completed execution, task execution engine may resume running the old task that was suspended, i.e., on hold).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the claimed accept and reject logic into Di Fabbrizio. Di Fabbrizio, Bharara and Vibbert are combinable because they all are directed to spoken dialog and natural language understanding systems. Bharara motivates the confidence adjustment by teaching sequence and timing based score adjustment and Vibbert motivates slot task handling by teaching confirmation to reduce ambiguity and modification of intent slots as state in Vibbert ¶85: “dialog engine 550 may continuously modify expectation agenda 570 having bound user input values to concepts or having analyzed needs for dialog focus shifts in order to continuously improve the expectations of any dialog agents or agencies using the bounded concepts or dialog focus shifts.”
Regarding Claim 2:
Di Fabbrizio discloses the method of claim 1, wherein the plurality of scored intent hypotheses are determined by:
passing input received from a conversation through a natural language understanding engine (Di Fabbrizio: ¶[0017] user utterance goes to ASR then text is passed to an NLU module);
determining a listing of possible intents (Di Fabbrizio: ¶[0019] and ¶[0024] discloses the NLU outputs an XML structure with a class name (e.g., TaxShelter) and attributes score, offset plus other context variables. These structured values are tied to each recognized intent (call type) (i.e., parameters, attributes and named entities bound to a particular intent)), of which each of the possible intents have one or more associated slot values (Vibbert: ¶[0110] discloses slot fields and that the system seeks user confirmation by providing prompts that can be chosen to confirm the specific slot content the user supplied);
associating a confidence value with each possible intent (Di Fabbrizio: ¶[0019] discloses that each candidate class (intent) has a score attribute interpreted as a confidence value);
scoring the confidence values by:
configuring the natural language understanding engine with a listing of all available intents (Di Fabbrizio: ¶[0019] discloses that there are categorization that the NLU computes of the user request in a list of possible intents ¶[0027] discloses a central data structure for a disambiguation algorithm is a tree where the leaves represent a successful classification of the user’s needs and each internal node is a choice of categories. Altogether this teaches a list of call types and categories defined in the tree),
retrieving stored information associated with historical conversations, from a memory of the dialog manager (Di Fabbrizio: ¶[0039] discloses that the lit node records, states and contexts and digression table are stored conversation information in the memory of the system from prior turns),
for each new intent, increasing the confidence score for a possible intent that is: a child of a current intent, matching a recently completed task (Bharara ¶[0040] discloses a confidence score may indicate the strength of the correlation between the input statement and intents, ¶[0099] further disclose a sequence of intents in the command tree may be analyzed which involves comparing the intent structure in the command tree against the sequence or structure of previously executed intents and changing or adjusting confidence scores based on these results, ¶[0100] discloses changing time based on recency. Altogether this teaches that boosting the score of an intent that is in the expected sequence (e.g., a child of the current intent in the tree) and boosting the score of intents corresponding to recently performed/completed tasks), or matching an on-hold task, and
determining a separate score for each distinct task path to an intent (Di Fabbrizio: ¶[0036] the system chooses a focus based on which path through the tree is supported, using rules like single lit direct descendent or lowest common ancestor. This means each candidate path (sequence of intents) has a score or amount of support from the NLU, each path by the NLU is essentially evaluated and scored and Boolean conditions are used on each node as well in order to find the optimal path), and
producing the list of the plurality of scored intent hypotheses (Di Fabbrizio: ¶[0019] discloses a list of possible intents that are associated with confidence scores).
Di Fabbrizio, Bharara and Vibbert are combinable because they are from the same field of endeavor of spoken dialogue systems and dialogue management. All disclose dialog managers that process user utterances via automatic speech recognition and natural language understanding to disambiguate and confirm user intent. It would have been obvious to one of ordinary skill in the art before the effective filing date to include to Di Fabbrizio’s confirmation workflow to also confirm and manage slot values as taught by Vibbert to improve accuracy and reduce ambiguity. The suggestion/motivation for doing so is provided by Vibbert’s teachings in ¶[0070] that “Conversational strategies such as turn taking behaviors, managing timing and order in which information is presented and asked from the user, grounding behavior such as seeking confirmation or reducing ambiguity in the conversation”.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to disclose for each new intent, increasing the confidence score for a possible intent that is: a child of a current intent, matching a recently completed task, or matching an on-hold task. Di Fabbrizio, Bharara and Vibbert are within the same field of endeavor, dialogue services using ASR. Di Fabbrizio discloses using ASR, an NLU module and a dialog manager to map user utterances to tasks and control dialog flow with associated confidence scores. Bharara discloses a system for natural language command translation to intents in dialog like systems and teaches assigning confidence or correlation scores to mappings and then adjusting those scores using historical dialog information including sequencing and timing of prior intents. The motivation to combine these references is stated within Bharara’s ¶[0100]: “Analyzing the command timing at 430 may include changing or adjusting the confidence scores based on the results of the analysis.” Which motivates increasing the confidence of hypotheses that are consistent with the current analysis and context.
Regarding Claim 3:
The combination of Di Fabbrizio, Vibbert and Bharara further discloses the method of claim 2, wherein the stored information associated with historical conversations comprises one or more of:
a path to the current task in an intent hierarchy (Di Fabbrizio: ¶[0027], ¶[0039] discloses maintaining a focus node and the lit node context state in the digressions table is effectively storing the path from the root to the current node in the intent hierarchy.);
a list of paths to recently completed tasks (Bharara: ¶100 discloses the sequence of intents may include comparing the intent structure in the command tree against the sequence of structure of previously executed intents); and
a list of paths to tasks currently on-hold (Vibbert: ¶98 discloses the task manager is configured to create, suspend, resume or abort and exist different tasks if another task is running the engine may determine if the current task can be suspended in favor of a new task, once the newly requested task has completed execution, task execution engine may resume running the old task that was suspended, i.e., on hold).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed inventions to modify Di Fabbrizio’s dialog disambiguation method to include Bharara’s sequence and history based confidence adjustment. Both references are from the same field of endeavor, namely natural language dialog system that map user utterances or commands to intents using an NLU/dialog systems that map user utterances or commands to intents using an NLU/dialog management framework. Di Fabbrizio uses automated speech recognition, natural language understanding and a dialog manager to determine and disambiguate user intent in a spoken dialog service (see Di Fabbrizio, ¶17-18). Bharara likewise concerns natural language command translation, intent determination, confidence scoring and command trees in dialog like software interfaces (see Bharara ¶36-40). The motivation to combine is provided by Bharara which teaches that confidences score adjustment using stored mappings and historical usage improves future command translation, stating that “by storing successful mappings using the previously successful mappings in future command translation, the significant correlation framework may gradually adapt over time to provide more accurate command translation” (see Bharara, ¶64).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the claimed accept and reject logic into Di Fabbrizio. Di Fabbrizio, Bharara and Vibbert are combinable because they all are directed to spoken dialog and natural language understanding systems. Bharara motivates the confidence adjustment by teaching sequence and timing based score adjustment and Vibbert motivates slot task handling by teaching confirmation to reduce ambiguity and modification of intent slots as state in Vibbert ¶85: “dialog engine 550 may continuously modify expectation agenda 570 having bound user input values to concepts or having analyzed needs for dialog focus shifts in order to continuously improve the expectations of any dialog agents or agencies using the bounded concepts or dialog focus shifts.”
Regarding Claim 4:
The proposed combination of Di Fabbrizio, Bharara and Vibbert further discloses the method of claim 3, wherein the stored information associated with historical conversations is used by the dialog manager (Di Fabbrizio: ¶[0019] discloses an NLU outputs confidence scored hypotheses, the dialog manager keeps history so it can return to a prior context) to further refine confidence scores for each new intent at runtime (Bharara: ¶[0097]-[0100] discloses calculating confidence scores for intent mappings, then these scores are adjusted based off of sequence and timing).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further refine confidence scores for each new intent at runtime. Di Fabbrizio and Bharara are within the same field of endeavor, dialogue services using ASR. Di Fabbrizio discloses using ASR, an NLU module and a dialog manager to map user utterances to tasks and control dialog flow with associated confidence scores. Bharara discloses a system for natural language command translation to intents in dialog like systems and teaches assigning confidence or correlation scores to mappings and then adjusting those scores using historical dialog information including sequencing and timing of prior intents. The motivation to combine these references is stated within Bharara’s ¶[0100]: “Analyzing the command timing at 430 may include changing or adjusting the confidence scores based on the results of the analysis.” Which motivates increasing the confidence of hypotheses that are consistent with the current analysis and context. Therefore, it would have been obvious to combine Di Fabbrizio in view of Bharara to obtain the invention in claim 4.
Regarding Claim 5:
The combination of Di Fabbrizio, Vibbert and Bharara further discloses the method of claim 2, wherein the conversation is between a human user and a bot (Di Fabbrizio: ¶[0017] discloses a user interacting with a dialog management system, which is a bot).
Regarding Claim 6:
The proposed combination of Di Fabbrizio, Bharara and Vibbert further discloses the method of claim 1, wherein the current path location comprises one of:
staying with what was previously the current path location based on the conversation with the dialog manager (Di Fabbrizio: ¶[0031] discloses the dialog manager processes input and stays with the previous focus node as the current path location when the input doesn’t push it deeper);
activating a child task of what was previously the current path location (Di Fabbrizio: ¶[0020] discloses if a new direct descended of the focus is selected, it is made the new focus);
re-opening a recently closed path (Vibbert: ¶[0059] t);
switching away from the previously current path (Vibbert: ¶[0059] the task manager may switch between multiple tasks that they manage based on the nature of the dialog);
switching back to a previously on-hold path (Vibbert: ¶[0098] discloses a suspended task, i.e., a task that is on hold and then resuming the previously suspended tasks, that is switching back to an on old task); and
disambiguating between different paths to the same intent (Di Fabbrizio: ¶[0036] discloses there are two different paths leading to the same high-level intent “tax shelter information” and the method disambiguates between them with a targeted prompt for the user).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the claimed accept and reject logic into Di Fabbrizio. Di Fabbrizio, Bharara and Vibbert are combinable because they all are directed to spoken dialog and natural language understanding systems. Bharara motivates the confidence adjustment by teaching sequence and timing based score adjustment and Vibbert teaches task switching and suspending or resuming operations in the same field of dialog management. Vibbert motivates the desire for this functionality in ¶85: “dialog engine 550 may continuously modify expectation agenda 570 having bound user input values to concepts or having analyzed needs for dialog focus shifts in order to continuously improve the expectations of any dialog agents or agencies using the bounded concepts or dialog focus shifts.”
Regarding Claim 7:
The combination of Di Fabbrizio, Vibbert and Bharara discloses the method of claim 6, wherein disambiguation comprises applying a corresponding weight towards a path location that is one of: recently completed, on-hold, a partial match to the current path location, a full match to the current path location (Bharara: ¶[0097]-[0100] discloses increasing confidence for mappings used recently, and comparing the current intent sequence against sequences stored in the intent sequence repository and adjusts scores when the current path partially or fully matches a known sequence).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to disclose applying a corresponding weight towards a path location that is recently completed. Di Fabbrizio and Bharara are within the same field of endeavor, dialogue services using ASR. Di Fabbrizio discloses using ASR, an NLU module and a dialog manager to map user utterances to tasks and control dialog flow with associated confidence scores. Bharara discloses a system for natural language command translation to intents in dialog like systems and teaches assigning confidence or correlation scores to mappings and then adjusting those scores using historical dialog information including sequencing and timing of prior intents. The motivation to combine these references is stated within Bharara’s ¶[0100]: “Analyzing the command timing at 430 may include changing or adjusting the confidence scores based on the results of the analysis.” Which motivates increasing the confidence of hypotheses that are consistent with the current analysis and context.
Regarding Claim 8:
The combination of Di Fabbrizio, Vibbert and Bharara the method of claim 6, wherein disambiguation comprises examination of lowest non-common denominator partial path locations and presenting said path locations as a list to the user (Di Fabbrizio: ¶[0036] discloses finding the lowest common ancestor of lit nodes and sets it as focus, this lowest common ancestor (LCA) is the lowest level where the paths diverge, i.e., the lowest non-common denominator partial path locations are the child branches below the LCA where the paths diverge, i.e., the candidate path choices that the system needs the user to pick between).
5. Claims 9-12 are rejected under 35 U.S.C. 103 as being unpatentable over Di Fabbrizio in view of Vibbert.
Regarding Claim 9:
Di Fabbrizio discloses a method for managing confirmation of a new intent and multiple slot values at the same time (Di Fabbrizio: ¶[0018] discloses a dialog management (DM) module that receives natural language understanding (NLU) results from an NLU module and uses prompts to clarify and confirm what the user meant), the method comprising the steps of:
defining one or more trigger phrases and one or more confirmation prompts (Di Fabbrizio: ¶[0027] discloses triggering conditions associated with the node and its descendants based specifically on user dialog (i.e., trigger phrases); ¶[0041], Table in Col 5 and ¶[0018]-[0019] using the DM to disambiguate and clarify user intentions (i.e., confirmation prompts));
triggering a phrase (Di Fabbrizio: ¶[0017]-[0018] and ¶[0021] discloses a user utterance (triggering phrase) causes the NLU to categorize and the DM to light relevant nodes and proceed with clarification);
selecting, by the dialog manager, the one or more confirmation prompts (Di Fabbrizio: ¶[0019] table of prompts, ¶[0036] prompts the user to disambiguate, in other words, the DM selects appropriate confirmation/clarification prompts based on the current focus node and conditions);
presenting the one or more confirmation prompts to the user (Di Fabbrizio: ¶[0019], ¶[0031]-[0034] discloses the DM presents the selected confirmation/clarification prompts to the user);
in response to the user answering the one or more confirmation prompts, mapping by the natural language understanding engine the user response to the new intent (Di Fabbrizio: ¶[0017]-[0018] discloses NLU maps the user’s replay to an intent, the DM uses that for the next step in the dialogue);
simultaneously confirming, by the dialog manager, the new intent (Di Fabbrizio: ¶[0018] ¶[0041] discloses confirmation handling by a confirmation sub dialog to confirm the newly determined intent); and
determining whether the new intent is a task intent, wherein if it is determined that the new intent is a task intent, rejecting the original intent (Di Fabbrizio: ¶[0039]-[0040] discloses shifting to a different objective and continuing with appropriate messaging. The system detects that a new/different task (intent) has arisen and performs a context shift. During this process it accepts/rejects the original path or the new path).
Di Fabbrizio does not explicitly disclose the following portions of the limitations above:
triggering a phrase that includes one or more slots;
selecting... confirmation prompts slots that match those provided by a user;
mapping… to the new intent and slot values;
simultaneously confirming… the new intent and slot values;
…rejecting the original intent and slots and accepting the new intent and slots… wherein if it is determined that the new intent is a non-task intent, accepting the original intent, otherwise, rejecting the original intent/slots and the new intent/slots.
However, Vibbert discloses:
triggering a phrase that includes one or more slots (Vibbert: ¶[0110] discloses that the dialog task manager involves intent slots tied to user phrases and commands);
selecting... confirmation prompts comprising slots that match those provided by a user (Vibbert: ¶[0110] discloses slot fields and that the system seeks user confirmation by providing prompts that can be chosen to confirm the specific slot content the user supplied);
mapping… to the new intent and slot values (Vibbert: ¶[0109]-[0110] discloses intent mapping and explicit change of slot values based on ranking and reranking);
simultaneously confirming… the new intent and slot values (Vibbert: ¶[0070] discloses seeking confirmation to reduce ambiguity ¶[0110] discloses removing adding or modifying intent slots, taken together, these teach confirming both the intent and its associated slot fields in the same confirmation interaction);
…rejecting the original intent and slots and accepting the new intent and slots… wherein if it is determined that the new intent is a non-task intent, accepting the original intent, otherwise, rejecting the original intent/slots and the new intent/slots (Vibbert: ¶[0110] discloses that when the dialog manager changes paths/intents it can accept/reject and update the associated slots).
Di Fabbrizio in view of Vibbert are combinable because they are from the same field of endeavor of spoken dialogue systems and dialogue management. Both disclose dialog managers that process user utterances via automatic speech recognition and natural language understanding to disambiguate and confirm user intent. It would have been obvious to one of ordinary skill in the art before the effective filing date to include to Di Fabbrizio’s confirmation workflow to also confirm and manage slot values as taught by Vibbert to improve accuracy and reduce ambiguity. The suggestion/motivation for doing so is provided by Vibbert’s teachings in ¶[0070] that “Conversational strategies such as turn taking behaviors, managing timing and order in which information is presented and asked from the user, grounding behavior such as seeking confirmation or reducing ambiguity in the conversation”. Therefore, it would have been obvious to combine Di Fabbrizio with Vibbert to obtain the invention as specified in claim 9.
Regarding Claim 10:
The proposed combination of Di Fabbrizio and Vibbert further discloses the method of claim 9, wherein the simultaneously confirming comprises the dialog manager selecting a confirmation prompt variation (Di Fabbrizio: ¶[0019]-[0020] discloses confirmation handling through template/tailored prompts conditioned on the user’s input) whose slots match those provided by the user (Vibbert: ¶[0110] discloses slot fields and that the system seeks user confirmation by providing prompts that can be chosen to confirm the specific slot content the user supplied, i.e., the additional dialogue confirmation necessarily matches the user slot fields).
Di Fabbrizio in view of Vibbert are combinable because they are from the same field of endeavor of spoken dialogue systems and dialogue management. Both disclose dialog managers that process user utterances via automatic speech recognition and natural language understanding to disambiguate and confirm user intent. It would have been obvious to one of ordinary skill in the art before the effective filing date to include to Di Fabbrizio’s confirmation workflow to also confirm and manage slot values as taught by Vibbert to improve accuracy and reduce ambiguity. The suggestion/motivation for doing so is provided by Vibbert’s teachings in ¶[0070] that “Conversational strategies such as turn taking behaviors, managing timing and order in which information is presented and asked from the user, grounding behavior such as seeking confirmation or reducing ambiguity in the conversation”. Therefore, it would have been obvious to combine Di Fabbrizio with Vibbert to obtain the invention as specified in claim 10.
Regarding Claim 11:
The proposed combination of Di Fabbrizio and Vibbert further discloses the method of claim 9, wherein an answer to the one or more confirmation prompts is mapped by the natural language understanding engine to the intent (Di Fabbrizio: ¶[0019]-[0020] discloses confirmation handling prompts that are given to the user and fed back to the NLU which classifies it into an intent (call type) with confidence, satisfying mapping to an intent) and slot values (Vibbert: ¶[0110] discloses slot fields and that the system seeks user confirmation by providing prompts that can be chosen to confirm the specific slot content the user supplied).
Di Fabbrizio in view of Vibbert are combinable because they are from the same field of endeavor of spoken dialogue systems and dialogue management. Both disclose dialog managers that process user utterances via automatic speech recognition and natural language understanding to disambiguate and confirm user intent. It would have been obvious to one of ordinary skill in the art before the effective filing date to include to Di Fabbrizio’s confirmation workflow to also confirm and manage slot values as taught by Vibbert to improve accuracy and reduce ambiguity. The suggestion/motivation for doing so is provided by Vibbert’s teachings in ¶[0070] that “Conversational strategies such as turn taking behaviors, managing timing and order in which information is presented and asked from the user, grounding behavior such as seeking confirmation or reducing ambiguity in the conversation”. Therefore, it would have been obvious to combine Di Fabbrizio with Vibbert to obtain the invention as specified in claim 11.
Regarding Claim 12:
The proposed combination of Di Fabbrizio and Vibbert further discloses the method of claim 9, wherein a context-switch message is provided to the user if the user triggers a different task intent during processing of the new intent (Di Fabbrizio: ¶[0039] “mixed initiative dialog must also allow the user to change the focus” of the request to different objectives at any time…context shift table will point the dialog execution to a different module that will handle the spurious request” this explicitly discloses detecting a new task intent mid-flow and shifting context, ¶[0040] further discloses an explicit message presented when the user triggers a different task intent, it informs the user of the context switch and continues).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to IAN SCOTT MCLEAN whose telephone number is (703)756-4599. The examiner can normally be reached "Monday - Friday 8:00-5:00 EST, off Every 2nd Friday".
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Hai Phan can be reached at (571) 272-6338. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/IAN SCOTT MCLEAN/ Examiner, Art Unit 2654
/HAI PHAN/ Supervisory Patent Examiner, Art Unit 2654
Prosecution Insights