DETAILED ACTION
This communication is in response to the amendment filed 4/1/26 in which claims 1, 8, 12, 13, 16 were amended. Claims 1-6 and 8-21 are pending. Claim 7 was previously canceled.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 12/22/25 was filed after the mailing date of the non-final action on 10/1/25. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Response to Arguments
Applicant, in pertinent part, argues:
Quass generally relates to an automated system for locating, interpreting, and electronically populating web-based forms in order to retrieve information that is otherwise hidden behind those forms. The system of Quass transforms HTML into structured object models and then uses automated decision-making to determine how to submit forms and access deeper content. A key aspect of Quass is its use of classifiers, which evaluate the object model to decide which forms should be filled out and how each field should be populated. However, the classifications of Quass are not related to, "inferring a plurality of field type classifications of an element on the electronic form by determining a plurality of field type probabilities corresponding to the plurality of field type classifications of the element;" and "based on the inferred plurality of field type classifications of the element, inferring the classification of the electronic form," as recited in independent claim 1.
App. Resp. 9-10.
Applicant’s arguments with respect to claim 1 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 103
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 non-obviousness.
Claims 1-8, 12, and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Iasi (US 2016/0019197 A1; published Jan. 21, 2016) in view of Quass (US 2002/0083068 A1; published Jun. 27, 2002) and Bourdev (US 7,343,551 B1; published Mar. 11, 2008).
Regarding claim 1, Iasi discloses [a] processing server system configured to infer a classification of an electronic form, the processing server system comprising:
one or more hardware processors; and (see ¶ 101)
memory storing computer instructions, (see ¶ 103) the computer instructions when executed by the one or more hardware processors configured to perform:
extracting information from raw source code and contextual information of the electronic form; (¶ 64 (“a form filler program is not able to “read” labels in the same way as a human does. Form fillers must rely on the underlying code that generally does not use standardized form field names. We see the “FIRST NAME” label, but a form filler sees the internal program code for this same field which could be some arbitrary name such as “FIELD102.” For example, an HTML code would look something like: <h2>FIRST NAME</h2>, <input type=“text”, name=“FIELD102”/>.”), ¶ 66 (“Any field on any form, PDF or Web-based, can be accurately identified given that there are sufficient terms describing it. FFMI can gather the context of the fields (for example, the section header “Name” 602 vs. “Name of Relative” 702) to assist in this goal. For example, the top level form category may be derived by the occurrences of certain keywords within the PDF document.”), ¶ 75 (“In a first step 1102, keyword content is extracted from the form.”)).
Although Iasi teaches “FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.” ¶ 68, Iasi does not expressly disclose based on the extracted information, inferring a plurality of field type classifications of an element on the electronic form by determining a plurality of probabilities… (but see Quass ¶ 67 (“This example appliance category classifier illustrates only one of the ways in which classifiers 166 in FIG. 4 could be employed in accordance with the invention. In general, a classifier could use any combination of information [extracted information] obtained from an object model 165, an XHTML document 163, an HTML document 161, support components 167, and other classifiers 166. The information available from an object model can be particularly useful if the object model exposes features that tend to indicate which classification is best, such as the descriptive text used by the simple appliance category classifier.”), ¶ 68 (“A classifier does not necessarily have to produce a yes-or-no decision. A classifier might choose from multiple classifications. For example, a classifier might classify a FormField object 224 [element] as one of: (1) spin through all values; (2) choose one particular value; (3) don't change anything [inferred field type classifications of the element]. For classification (2), the particular value chosen might be identified by a support component 167 or by another classifier 166. Classification (3) might be the decision the classifier reverts to if it cannot pick (1) or (2) with sufficient confidence. A classifier might also return a confidence level for its classification, perhaps to be used in resolving conflicting classifications from multiple classifiers. For example, if a classifier identifies more than one form per document that should be filled out, the one whose "fill it out" decision has the highest confidence might be chosen.”), ¶ 70 (“Flowchart 250 is only one of the ways in which classifiers 166 could perform their classification task. Classifiers might use advanced techniques from the broad field of machine learning, which can make them especially useful in complex situations. For example, a classifier might compute whether a SubmitButtonField 233 is the correct submit button to press by using a machine learning technique that can take into account a large number of features. Such features might include whether the button's text contains indicative keywords like “submit” or “search”, whether the button's text contains contraindicative keywords like “reset” or “e-mail”, whether there are other submit buttons in the form, whether the button is the first button in the form, etc. The presence or absence of these features might be combined mathematically to compute an overall probability, with the classification being made according to whether the probability exceeds a threshold. The classifier might have been previously trained how to best combine the features by examining examples of forms whose correct submit buttons have already been correctly identified, and adjusting parameters in order to best classify those examples.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Quass to provide confidence values associated multiple classifications for a form field element of a form, at least because doing so would allow resolving conflicting classifications from multiple classifiers. See Quass ¶ 68.
Although Iasi and Quass teach that the classifiers use advanced techniques from machine learning to compute an overall probability that a field is of a certain type based on HTML/XHTML source code, they do not expressly teach doing so by determining a plurality of probabilities corresponding to the plurality of inferred field type classifications (but see Bourdev col. 3, ll. 44-50 (“The generated likelihood assessments can be adjusted based on a determined characteristic of the current form field object. The determined characteristic of the current form field object can be a determined type for previously entered values. The determined characteristic of the current form field object can be a field type distribution indicating probabilities that the current form field object is of a given type.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi and Quass to incorporate the teachings of Bourdev to obtain a field type distribution indicating probabilities that the current form field object is of a given type, at least because doing so would allow a classifier to provide more than a yes-or-no decision about a field type. See Quass ¶ 68.
Iasi further discloses based on the inferred plurality of field type classifications of the element, inferring the classification of the electronic form; (¶ 70 (“The form category can be determined using a machine learning classification engine as applied to keywords extracted from the document,”), ¶ 75 (“The result of the machine learning algorithm is the generation of candidate categories and confidence values for each category in step 1106. Depending on the engine used in step 1104, there may be more than one good fitting category that the document fits within.”))
suggesting a downstream action based on the inferred plurality of field type classifications of the element and the inferred classification of the electronic form; and (¶ 68 (“FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.”), ¶ 74 (“The matching form values are then transmitted, in step 1010, from the remote server to the browser application, where the browser extension application operates to enter all of the field values into the respective fields in order to complete the online form.”))
selectively performing the downstream action (¶ 74 (“Essentially, when the user presses the “FILL” button in the browser extension menu, the server can fill values to the fields based on the underlying terms previously assigned to each field.”)).
Regarding claim 2, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi does further discloses wherein the extracting of information further comprises extracting metadata of the electronic form, the metadata indicating previous versions or a lineage of the electronic form (¶ 91 (“In some embodiments, as part of the series of validity checks, the form filler application will attempt a cursory check to determine if the same form has been encountered in the past. For example, in some embodiments, the form filler application can search one or more storage locations or media (e.g., a memory store, a server, or a database) for prior data, including prior form field attribute values. In some embodiments, the one or more storage locations or media are remote (e.g., a remote memory store, a remote server, or a remote database). Therefore, in some embodiments, the form filler application determines whether it has already encountered the same form in the past based at least in part on remote data. This is accomplished by performing field-level checks which compare either the name or ID attribute with the field names on the server and see if it exists.”)).
Regarding claim 3, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses wherein the computer instructions, when executed by the one or more hardware processors, are configured to perform determining one or more probabilities corresponding to the one or more inferred classifications of elements and the inferred classification of the electronic form (¶ 46 (“When employing multiple term mapping it is important to note that the terms in the form field do not necessarily need to precisely match the terms in the database. The mapping engine is able to use any number of probabilistic search or other matching algorithms…”)).
Regarding claim 4, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses wherein the downstream action comprises autofilling or autocompleting one or more elements on the electronic form (¶ 74 (“Essentially, when the user presses the “FILL” button in the browser extension menu, the server can fill values to the fields based on the underlying terms previously assigned to each field.”)).
Regarding claim 5, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses wherein the contextual information comprises any textual features and media components within the electronic form, (¶ 66 (“Any field on any form, PDF or Web-based, can be accurately identified given that there are sufficient terms describing it. FFMI can gather the context of the fields (for example, the section header “Name” 602 vs. “Name of Relative” 702) to assist in this goal.”)) and relative positions of the any textual features and media components (¶ 57 (“in step 504, the system assigns multiple relevant terms for each field within a document by decomposing the document's structure and identifying a field's hierarchical position within the document”)).
Regarding claim 6, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses wherein the contextual information comprises inferred or verified classifications of previous elements or forms of an immediately preceding form, wherein the immediately preceding form, following submission, populates the electronic form (¶ 48 (“an MTM-based form filler will use a best-fit algorithm to match the closest and best matching value for the field. MTM supports a range of mapping algorithms that can be improved over time.”), ¶ 54 (“the automatic form mapper will look at each document's structure and assign a field name that is based on the field's hierarchical context within the form. Once we identified the form category 410/412, we could then identify sections 406/408, subsections, and finally the field name 402/404 itself. So what you end up with is a set of “keywords” that uniquely locate and identify that field's hierarchical context within the document.”)).
Regarding claim 8, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses wherein the computer instructions, when executed by the one or more hardware processors, are configured to perform:
detecting an update to the electronic form, the update comprising a new or modified element; (¶ 92 (“the form filler application can determine whether an electronic form has been changed to include new form fields. In some embodiments, the form filler application is configured to look for the addition of specific types of form fields (e.g., personal and financial data inputs)”))
extracting new information from new raw source code and new contextual information of the new or modified element; (¶ 64 (“a form filler program is not able to “read” labels in the same way as a human does. Form fillers must rely on the underlying code that generally does not use standardized form field names. We see the “FIRST NAME” label, but a form filler sees the internal program code for this same field which could be some arbitrary name such as “FIELD102.” For example, an HTML code would look something like: <h2>FIRST NAME</h2>, <input type=“text”, name=“FIELD102”/>.”), ¶ 66 (“Any field on any form, PDF or Web-based, can be accurately identified given that there are sufficient terms describing it. FFMI can gather the context of the fields (for example, the section header “Name” 602 vs. “Name of Relative” 702) to assist in this goal. For example, the top level form category may be derived by the occurrences of certain keywords within the PDF document.”))
based on the new information, inferring one or more classifications of the new or modified element; (¶ 66 (“The aim of FFMI is to automatically generate a sufficiently unique set of terms to describe a given field. For example, a form filler will see a field named “FIRST NAME CLOSEST RELATIVE” instead of “FIELD103,” and this makes filling the field with the correct value much easier.”))
based on the inferred one or more classifications of the new or modified element, inferring an updated classification of the electronic form; (¶ 70 (“The form category can be determined using a machine learning classification engine as applied to keywords extracted from the document,”))
suggesting a second downstream action based on the inferred one or more classifications of the new or modified element and the updated classification of the electronic form; and (¶ 68 (“FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.”), ¶ 74 (“The matching form values are then transmitted, in step 1010, from the remote server to the browser application, where the browser extension application operates to enter all of the field values into the respective fields in order to complete the online form.”))
selectively performing the second downstream action (¶ 74 (“Essentially, when the user presses the “FILL” button in the browser extension menu, the server can fill values to the fields based on the underlying terms previously assigned to each field.”)).
Regarding claim 12, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses wherein the inferring of the plurality of field type classifications of the element on the electronic form and the inferring of the classification of the electronic form are performed using one or more machine learning components, (¶ 8 (“The mapping algorithms may be improved over time and use a variety of approaches, such as machine learning and artificial intelligence methods, to associate field names with field values to be input into the form fields.”)) and the one or more machine learning components are trained iteratively, using a first training dataset comprising previously inferred or verified classifications of elements and forms and a second training dataset comprising incorrectly inferred classifications of the prior elements and the prior forms by the one or more machine learning components following the training using the first training dataset (¶ 50 (“Fields are matched using a scalable learning engine that improves over time as the mapping engine learns how to map form fields by analyzing the field values entered into those form fields.”), ¶ 75 (“In one embodiment, the Machine Learning Classification engine will be trained with a cross-section of different documents in advance of its application. Additionally, additional sets of training data may be built and incorporated on a continuous basis with data that users input into forms and fields.”)).
Regarding claim 21, Iasi, in view of Quass and Bourdev, discloses the invention of claim 1 as discussed above. Iasi further discloses the computer instructions when executed by the one or more hardware processors further configured to reclassify a second element on the electronic form based on the inferred classification of the electronic form (¶ 54 (“However, the automatic form mapper will look at each document's structure and assign a field name that is based on the field's hierarchical context within the form. Once we identified the form category 410/412, we could then identify sections 406/408, subsections, and finally the field name 402/404 [“second element”] itself.”)).
Claims 16-20 are rejected under 35 U.S.C. 103 as being unpatentable over Iasi in view of Bourdev and Quass.
Regarding claim 16, Iasi discloses [a] client device configured to infer a classification of an electronic form, the client device comprising one or more hardware processors (¶ 101); and
memory storing computer instructions, (¶ 103) the computer instructions when executed by the one or more hardware processors configured to perform:
receiving a plugin, the plugin comprising a machine learning component that classifies one or more elements within the electronic form by determining one or more probabilities…classifies the electronic form based on the one or more inferred classifications of the one or more elements; and (¶ 50 (“In one embodiment, the system utilizes machine learning methods to accurately associate database values to form fields. Fields are matched using a scalable learning engine that improves over time as the mapping engine learns how to map form fields by analyzing the field values entered into those form fields.”), ¶ 61 (“Disclosed herein are systems and methods for identifying and labeling form fields using machine learning algorithms. A form field mapping and identification engine identifies a form category using a machine learning classification algorithm, then determines and maps form labels to form fields using seeded values and optical scanning in order to produce an enhanced field name that incorporates a readable label associated with each field.”), ¶ 95 (“As demonstrated herein, the forms may be populated directly on the user's device, through a form filler application that can be deployed as a browser extension, add-on browser application, or as an application programming interface (API) interacting with a third party service or application.”), ¶ 68 (“As will be described in further detail below, FFMI uses the advanced concepts of a Scalable Machine Learning Framework and its advanced concepts to assign a set of descriptive terms to each field in a document. Current form fillers perform a strict one-for-one literal match between the field name in the form and the field name in its database containing the value to assign to the field. FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.”)).
Iasi does not expressly disclose determining one or more probabilities corresponding to one or more inferred field type classifications of the one or more elements (but see Bourdev col. 3, ll. 44-50 (“The generated likelihood assessments can be adjusted based on a determined characteristic of the current form field object. The determined characteristic of the current form field object can be a determined type for previously entered values. The determined characteristic of the current form field object can be a field type distribution indicating probabilities that the current form field object is of a given type.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi and Quass to incorporate the teachings of Bourdev to obtain a field type distribution indicating probabilities that the current form field object is of a given type, at least because doing so would allow performing a probabilistic best-fit assignment of names to a form field instead of being restricted to a one-for-one exact match to a preset field type.
Iasi further discloses executing the plugin, wherein the executing of the plugin comprises:
extracting information from raw source code and contextual information of the electronic form; (¶ 64 (“a form filler program is not able to “read” labels in the same way as a human does. Form fillers must rely on the underlying code that generally does not use standardized form field names. We see the “FIRST NAME” label, but a form filler sees the internal program code for this same field which could be some arbitrary name such as “FIELD102.” For example, an HTML code would look something like: <h2>FIRST NAME</h2>, <input type=“text”, name=“FIELD102”/>.”), ¶ 66 (“Any field on any form, PDF or Web-based, can be accurately identified given that there are sufficient terms describing it. FFMI can gather the context of the fields (for example, the section header “Name” 602 vs. “Name of Relative” 702) to assist in this goal. For example, the top level form category may be derived by the occurrences of certain keywords within the PDF document.”), ¶ 75 (“In a first step 1102, keyword content is extracted from the form.”)).
Although Iasi teaches “FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.” ¶ 68, Iasi does not expressly disclose based on the extracted information, inferring a plurality of field type classifications of an element on the electronic form by determining a plurality of field type probabilities…(but see Quass ¶ 68 (“A classifier does not necessarily have to produce a yes-or-no decision. A classifier might choose from multiple classifications. For example, a classifier might classify a FormField object 224 [element] as one of: (1) spin through all values; (2) choose one particular value; (3) don't change anything [plurality of field type classifications]. For classification (2), the particular value chosen might be identified by a support component 167 or by another classifier 166. Classification (3) might be the decision the classifier reverts to if it cannot pick (1) or (2) with sufficient confidence. A classifier might also return a confidence level for its classification, perhaps to be used in resolving conflicting classifications from multiple classifiers. For example, if a classifier identifies more than one form per document that should be filled out, the one whose "fill it out" decision has the highest confidence might be chosen.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Quass to provide confidence values associated multiple classifications for a form field element of a form, at least because doing so would allow resolving conflicting classifications from multiple classifiers. See Quass ¶ 68.
Although Iasi and Quass teach that the classifiers use advanced techniques from machine learning to compute an overall probability that a field is of a certain type based on HTML/XHTML source code, they do not expressly teach determining a plurality of probabilities corresponding to the plurality of inferred field type classifications (but see Bourdev col. 3, ll. 44-50 (“The generated likelihood assessments can be adjusted based on a determined characteristic of the current form field object. The determined characteristic of the current form field object can be a determined type for previously entered values. The determined characteristic of the current form field object can be a field type distribution indicating probabilities that the current form field object is of a given type.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi and Quass to incorporate the teachings of Bourdev to obtain a field type distribution indicating probabilities that the current form field object is of a given type, at least because doing so would allow a classifier to provide more than a yes-or-no decision about a field type. See Quass ¶ 68.
Iasi further discloses based on the inferred field type classifications of the element, inferring the classification of the electronic form; (¶ 70 (“The form category can be determined using a machine learning classification engine as applied to keywords extracted from the document,”), ¶ 75 (“The result of the machine learning algorithm is the generation of candidate categories and confidence values for each category in step 1106. Depending on the engine used in step 1104, there may be more than one good fitting category that the document fits within.”))
suggesting a downstream action based on the inferred plurality of field type classifications of the element and the inferred classification of the electronic form; and (¶ 68 (“FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.”), ¶ 74 (“The matching form values are then transmitted, in step 1010, from the remote server to the browser application, where the browser extension application operates to enter all of the field values into the respective fields in order to complete the online form.”))
selectively performing the downstream action (¶ 74 (“Essentially, when the user presses the “FILL” button in the browser extension menu, the server can fill values to the fields based on the underlying terms previously assigned to each field.”)).
Regarding claim 17, Iasi, in view of Bourdev and Quass, discloses the invention of claim 16 as discussed above. Iasi further discloses wherein the extracting of information further comprises extracting metadata of the electronic form, (¶ 64 (“a form filler program is not able to “read” labels in the same way as a human does. Form fillers must rely on the underlying code that generally does not use standardized form field names. We see the “FIRST NAME” label, but a form filler sees the internal program code for this same field which could be some arbitrary name such as “FIELD102.” For example, an HTML code would look something like: <h2>FIRST NAME</h2>, <input type=“text”, name=“FIELD102”/>.”), ¶ 66 (“Any field on any form, PDF or Web-based, can be accurately identified given that there are sufficient terms describing it. FFMI can gather the context of the fields (for example, the section header “Name” 602 vs. “Name of Relative” 702) to assist in this goal. For example, the top level form category may be derived by the occurrences of certain keywords within the PDF document.”)) the metadata indicating previous versions or a lineage of the electronic form (¶ 91 (“In some embodiments, as part of the series of validity checks, the form filler application will attempt a cursory check to determine if the same form has been encountered in the past. For example, in some embodiments, the form filler application can search one or more storage locations or media (e.g., a memory store, a server, or a database) for prior data, including prior form field attribute values. In some embodiments, the one or more storage locations or media are remote (e.g., a remote memory store, a remote server, or a remote database). Therefore, in some embodiments, the form filler application determines whether it has already encountered the same form in the past based at least in part on remote data. This is accomplished by performing field-level checks which compare either the name or ID attribute with the field names on the server and see if it exists.”)).
Regarding claim 18, Iasi, in view of Bourdev and Quass, discloses the invention of claim 16 as discussed above. Iasi further discloses wherein the computer instructions, when executed by the one or more hardware processors, are configured to perform determining one or more probabilities corresponding to the inferred classification of the electronic form (¶ 46 (“When employing multiple term mapping it is important to note that the terms in the form field do not necessarily need to precisely match the terms in the database. The mapping engine is able to use any number of probabilistic search or other matching algorithms…”)).
Regarding claim 19, Iasi, in view of Bourdev and Quass, discloses the invention of claim 16 as discussed above. Iasi further discloses wherein the downstream action comprises autofilling or autocompleting one or more of the one or more elements (¶ 74 (“Essentially, when the user presses the “FILL” button in the browser extension menu, the server can fill values to the fields based on the underlying terms previously assigned to each field.”)).
Regarding claim 20, Iasi, in view of Bourdev and Quass, discloses the invention of claim 16 as discussed above. Iasi further discloses wherein the contextual information comprises any textual features and media components within the electronic form, (¶ 66 (“Any field on any form, PDF or Web-based, can be accurately identified given that there are sufficient terms describing it. FFMI can gather the context of the fields (for example, the section header “Name” 602 vs. “Name of Relative” 702) to assist in this goal.”)) and relative positions of the any textual features and media components (¶ 57 (“in step 504, the system assigns multiple relevant terms for each field within a document by decomposing the document's structure and identifying a field's hierarchical position within the document”)).
Claims 9 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over Iasi, Quass, and Bourdev as applied to claim 8, in further view of Dalle (US 2018/0181866 A1; published Jun. 28, 2018).
Regarding claim 9, Iasi, in view of Quass and Bourdev, discloses the invention of claim 8 as discussed above. Iasi does not expressly disclose wherein the update to the electronic form is responsive to a change in an entity being monitored or tracked by the electronic form. However, Dalle teaches retraining an autofill application using a received user input correcting a first value provided in a first field of an electronic form. Dalle Abstract.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Dalle to retrain the machine learning classification engine using a correction received from a user input correcting a first value providing in a first field of the electronic form, at least because doing so would enable the machine learning system to observe the manual correction and help retrain the autofill system. Dalle ¶ 14.
Regarding claim 11, Iasi, in view of Quass and Bourdev, discloses the invention of claim 8 as discussed above. Iasi does not expressly disclose wherein the update to the electronic form is in response to a user input or a user action within the electronic form. However, Dalle (US 2018/0181866 A1; published Jun. 28, 2018) teaches retraining an autofill application using a received user input correcting a first value provided in a first field of an electronic form. Dalle Abstract.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Dalle to retrain the machine learning classification engine using a correction received from a user input correcting a first value providing in a first field of the electronic form, at least because doing so would enable the machine learning system to observe the manual correction and help retrain the autofill system. Dalle ¶ 14.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Iasi, Quass, and Bourdev as applied to claim 8 above, and further in view of Lucas (US 2020/0126663 A1; published Apr. 23, 2020).
Regarding claim 10, Iasi, in view of Quass and Bourdev, discloses the invention of claim 8 as discussed above. Iasi does not expressly disclose wherein the update to the electronic form comprises an automatic switch between different versions of the electronic form at particular time intervals. However, Lucas teaches periodically checking NLP/MLA models being used to map form fields. Lucas ¶¶ 231, 237.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Lucas to periodically check the form classification machine learning model, at least because doing so would make help make sure that the models are up-to-date. Lucas ¶ 237.
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Iasi in view of Bourdev and Miao (US 2017/0091651 A1; published mar. 30, 2017).
Regarding claim 13, Iasi discloses [a] processing server system configured to identify one or more classifications of an electronic form, the processing server system comprising:
one or more hardware processors; and (¶ 101)
memory storing computer instructions, (¶ 103) the computer instructions when executed by the one or more hardware processors configured to perform:
distributing a plugin to a client device, the plugin comprising a machine learning component that classifies one or more elements within the electronic form by determining one or more probabilities…and classifies the electronic form based on the one or more inferred field type classifications of the one or more elements; (¶ 47 (“Multiple Term Mapping (MTM) supports a continuum of mapping engines ranging from a simple single term literal match method all the way up to advanced machine intelligence frameworks. As more advanced mapping technologies are developed, it will be possible to easily incorporate those technologies without having to re-design the mapping applications (extensions, form fillers, etc.) or having to re-map existing forms.”), ¶ 95 (“As demonstrated herein, the forms may be populated directly on the user's device, through a form filler application that can be deployed as a browser extension, add-on browser application, or as an application programming interface (API) interacting with a third party service or application.”), ¶ 68 (“As will be described in further detail below, FFMI uses the advanced concepts of a Scalable Machine Learning Framework and its advanced concepts to assign a set of descriptive terms to each field in a document. Current form fillers perform a strict one-for-one literal match between the field name in the form and the field name in its database containing the value to assign to the field. FFMI provides a way for Multiple Term Mapping (MTM)-based form fillers to read all of the terms assigned to a form field and perform a probabilistic best-fit for a similar set of terms in its database. There is no need to be restricted to a one-for-one exact match, which opens the door for more fields to be auto-filled more accurately and with greater relevancy.”)).
Iasi does not expressly disclose determining one or more probabilities corresponding to one or more inferred field type classifications of the one or more elements (but see Bourdev col. 3, ll. 44-50 (“The generated likelihood assessments can be adjusted based on a determined characteristic of the current form field object. The determined characteristic of the current form field object can be a determined type for previously entered values. The determined characteristic of the current form field object can be a field type distribution indicating probabilities that the current form field object is of a given type.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi and Quass to incorporate the teachings of Bourdev to obtain a field type distribution indicating probabilities that the current form field object is of a given type, at least because doing so would allow performing a probabilistic best-fit assignment of names to a form field instead of being restricted to a one-for-one exact match to a preset field type.
Although Iasi teaches deploying a form filler application on client devices, Iasi does not expressly disclose distributed machine learning. However, Miao discloses receiving feedback from the client device regarding a performance of the machine learning component; (¶ 22 (“. . . user feedback from users of the clients”), ¶ 23 (“. . . a local version of statistical model 108 that is adapted to input data on the corresponding client”))
transmitting an indication to perform further training on the machine learning component based on the received feedback; (¶ 22 (“. . . the clients may be electronic devices (e.g., personal computers, laptop computers, mobile phones, tablet computers, portable media players, digital cameras, etc.) that produce updates 114-116 to statistical model 108 based on user feedback from users of the clients.”))
obtaining an updated machine learning component based on the further training; and (¶ 22 (“. . . statistical model 108 may be trained and/or adapted to new data received on the clients.”), ¶ 24 (“server 102 may merge updates 114-116 into subsequent global versions of statistical model 108.”),
distributing the plugin having the updated machine learning component to the client device (¶ 24 (“After a new global version of statistical model 108 is created, server 102 may transmit the new global version to the clients to propagate updates 114-116 included in the new global version to the clients.”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Miao to retrain the classification machine learning model based on user feedback from users of the clients, at least because doing so would prevent overfitting of the form classification machine learning model to input data of individual users. Miao ¶ 25.
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Iasi, Bourdev, and Miao as applied to claim 13 above, and further in view of Lee (US 2020/0234109 A1; published Jul. 23, 2020).
Regarding claim 14, Iasi, in view of Bourdev and Miao, discloses the invention of claim 13 as discussed above. Although Iasi teaches a machine learning classification engine to correctly infer correct form-filling, Iasi does not expressly disclose wherein the received feedback comprises erroneous inferences of classifications of the one or more elements or erroneous inferences of classifications of the electronic form. However, Lee teaches receiving user feedback from a user of a client computing device to dynamically modify the operation of a classification model. Lee ¶ 100. The output of the classification may be used to generate a notification to a recipient of the communication information them of the classification generated by the model and user feedback may be received back indicating whether or not the classification was correct or not. ¶ 114.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Lee to provide user feedback whether the classification of a form field is correct or not, at least because doing so would enable dynamically modifying the operation of the machine learning classification engine.
Claim 15 is rejected under 35 U.S.C. 103 as being unpatentable over Iasi, Bourdev, and Miao as applied to claim 13 above, and further in view of Han (CN 112418392 A; published Feb. 26, 2021).
Regarding claim 15, Iasi, in view of Bourdev and Miao, discloses the invention of claim 13 as discussed above. Iasi does not expressly disclose downscaling the machine learning model. However, Han discloses wherein the computer instructions, when executed by the one or more hardware processors, are configured to perform:
storing a trained machine learning component within the processing server system; and (1:40-41 (“construct multiple first neural networks according to the at least one set of first parameter combinations”))
wherein the distributing of the plugin comprises
determining or obtaining one or more storage or processing attributes or constraints of the client device; and (1:41-43 (“obtain A constraint range, the constraint range includes a numerical range that identifies the computing capability of the computing device, the constraint range may be a numeric range determined according to the information of the computing capability of the computing device”))
selectively downscaling the machine learning component relative to the stored trained machine learning component based on the one or more storage or processing attributes or constraints of the client device (1:43-48 (“according to the mapping relationship, the second parameter combination corresponding to the constraint range is obtained, The mapping relationship includes the relationship between multiple parameters and the evaluation results of the multiple first neural networks, and the evaluation result is the result obtained by evaluating the structure of each first neural network in the multiple first neural networks; The two parameters are combined to obtain the target neural network”)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iasi to incorporate the teachings of Han to obtain a target neural network based on the size or performance constraints of the computing device, at least because doing so would enable running the trained machine learning classification engine on different hardware device. See Han 1:24-28.
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 SHAHID KHAN whose telephone number is (571)270-0419. The examiner can normally be reached M-F, 9-5 est.
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, Usmaan Saeed can be reached at (571)272-4046. 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.
/SHAHID K KHAN/Primary Examiner, Art Unit 2146
Prosecution Insights