eBay Inc.
[2019] APO 10
•19 March 2019
IP AUSTRALIA
AUSTRALIAN PATENT OFFICE
eBay Inc. [2019] APO 10
Patent Application: 2015315634
Title:Enhanced Search Query Suggestions
Patent Applicant: eBay Inc.
Delegate: M. G. Kraefft
Decision Date: 19 March 2019
Hearing Date: Written submissions filed on 7 February 2019
Catchwords: PATENTS – section 45 – examiner’s objection – whether claimed invention is a manner of manufacture – generating enhanced search query suggestions – not evident that claimed enhancement requires more than mere generic computer implementation – claimed invention not a manner of manufacture – possibly patentable subject matter in body of specification – sub-regulation 13.4(1)(g) applied to provide applicant with three months to obtain acceptance of application.
Representation: Patent attorney for the applicant: FPA Patent Attorneys
IP AUSTRALIA
AUSTRALIAN PATENT OFFICE
Patent Application: 2015315634
Title:Enhanced Search Query Suggestions
Patent Applicant: eBay Inc.
Date of Decision: 19 March 2019
DECISION
The claimed invention is not for a manner of manufacture.
There may be patentable subject matter in the body of the specification from which valid claims could be drafted to overcome the above finding.
Pursuant to sub-regulation 13.4(1)(g), the final date for acceptance is three (3) months from the date of this decision. The applicant is invited to file amendments and any supporting submissions addressing the findings in this decision. In the event that such amendments and/or submissions are filed, the response will be dealt with in the normal manner of such responses to an examiner’s report.
REASONS FOR DECISION
BACKGROUND
The present matter concerns a patent application by eBay Inc. (“the applicant”). The applicant filed the patent application on 28 August 2015. The application number is 2015315634. The application is based on two earlier US applications. The earliest claimed priority date is 11 September 2014.
The application has been subjected to four examination reports. The first two examination reports were issued on the same date. Those two reports are substantially the same aside from the second report having some minor changes of wording within an objection, and having corrections of some typographical errors that existed in the first report. The substantive objection in those reports was that the claimed invention lacked an inventive step.
The applicant subsequently filed a statement of proposed amendments to amend the claims of the specification and to amend some pages of the description.
In the third report, the examiner maintained an objection that a number of claims lacked an inventive step and introduced an objection that all the claims did not define a manner of manufacture. Despite responsive submissions from the applicant, the examiner maintained the latter objection in the fourth report.
The applicant subsequently requested to be heard.
While the final date for acceptance of the application was 17 November 2018, patent sub-regulation 13.4(1)(g) may be available to extend the time for gaining acceptance to 3 months from the date of the present decision.
SPECIFICATION
The specification describes the present invention as relating generally to generating search query suggestions and, more particularly, to enhancing search query suggestions. As background, the specification mentions that search queries conventionally have included search query suggestions provided to a user while the user is entering a search query into a search field. The specification further mentions that, in some systems, search query suggestions are known as Auto Complete. Search query suggestions may thus be used to populate and formulate search queries from search queries stored in a search system. The suggestions often represent the search system determining suggested terms related to the terms or partial terms input into the search query. Moreover, search systems commonly use global scoring functions to evaluate the overall performance of a search query across all users to rank the popularity and efficiency of a particular term or set of terms used in a search query.
The specification, as proposed to be amended, ends with 20 claims. Claims 1, 11 and 13 are independent claims. These claims may be found at Annex A at the end of this decision.
APPLICABLE LAW
The application is governed by the Patents Act 1990 (“the Act”) as amended by the Intellectual Property Laws Amendment (Raising the Bar) Act 2012 (“the Raising the Bar Act”). Amendments to Sections 7, 40 and 49 of the Act apply to the present case as a consequence of Schedule 1, items 55(1)(d) and 55(4)(a), and Schedule 6, item 133(7)(d), of the Raising the Bar Act. The application was filed after 15 April 2013.
Section 18 of the Patents Act 1990 provides that:-
(1)Subject to subsection (2), an invention is a patentable invention for the purposes of a standard patent if the invention, so far as claimed in any claim:
(a) is a manner of manufacture within the meaning of section 6 of the Statute of Monopolies; and …
The standard of proof that applies is the balance of probabilities (subsection 49(1)). I must accept the application if satisfied on the balance of probabilities that the application complies with the Act. If I am not so satisfied, then I can refuse the application.
CASE LAW
In National Research Development Corporation v Commissioner of Patents (“NRDC”), [1959] HCA 67, (1959) 102 CLR 252, the High Court provided a statement of the law in this regard. At page 275, “… a process, to fall within the limits of patentability which the context of the Statute of Monopolies has supplied, must be one that offers some advantage which is material, in the sense that the process belongs to a useful art as distinct from a fine art …- that its value to the country is in the field of economic endeavour”. In discussing the “vendible product” proposition put forward by Morton J in Re G.E.C’s Application, (1942) 60 RPC 1, the High Court in NRDC upheld the validity of a patent for the use of previously unknown properties of a known chemical to effect a new purpose. At page 277:-
“The effect produced by the appellant’s method exhibits the two essential qualities upon which ‘product’ and ‘vendible’ seem designed to insist. It is a ‘product’ because it consists in an artificially created state of affairs, discernible by observing over a period the growth of weeds and crops respectively on sown land on which the method has been put into practice. And the significance of the product is economic; for it provides a remarkable advantage … for one of the most elemental activities by which man has served his material needs, the cultivation of the soil for the production of its fruits.”
The High Court though was not laying down a precise formulation that can be applied unthinkingly. In D’Arcy v Myriad Genetics Inc (“Myriad”), [2015] HCA 35, at [23]:-
“This Court in NRDC did not prescribe a well-defined pathway for the development of the concept of ‘manner of manufacture’ in its application to unimagined technologies with unimagined characteristics and implications. Rather, it authorised a case-by-case methodology.”
That case-by-case approach must have regard to the substance of the claimed invention, not simply the form of the claim. The point was made succinctly in the Myriad case by Gageler and Nettle JJ. At [144]:-
“Whatever words have been used, the matter must be looked at as one of substance and effect must be given to the true nature of the claim.”
In Commissioner of Patents v RPL Central Pty Ltd (“RPL”), [2015] FCAFC 177, the Full Court of the Federal Court stated the same thing in the context of an invention that was in substance a scheme. At [96]:-
“A claimed invention must be examined to ascertain whether it is in substance a scheme or plan or whether it can broadly be described as an improvement in computer technology. The basis for the analysis starts with the fact that a business method, or mere scheme, is not, per se, patentable. The fact that it is a scheme or business method does not exclude it from properly being the subject of letters patent, but it must be more than that. There must be more than an abstract idea; it must involve the creation of an artificial state of affairs where the computer is integral to the invention, rather than a mere tool in which the invention is performed.”
Moreover at [98]:-
“It is not a question of stating precise guidelines but of deciding, in each case, whether the claimed invention, as a matter of substance not form, is properly the subject of a patent”.
In Research Affiliates LLC v Commissioner of Patents, [2014] FCAFC 150, the Full Court of the Federal Court noted that a consistent approach in UK decisions, as distinct from previous Board of Appeal decisions of the European Patent Office, could be of assistance in the Australian context. At [23] of Research Affiliates:-
“In our opinion, it is more helpful to consider the analysis of the issue in the UK decisions which, with respect, provide a consistent approach. Despite being in the context of the statutory exclusion of computer programs ‘as such’, the UK decisions are of assistance in understanding the distinction to be drawn in the Australian context between an unpatentable business method and a claimed invention which may be patentable if the invention results in an ‘artificial effect’, within the understanding of that concept as explained in NRDC.”
At [36], the court further indicated that applying a test of a “technical contribution”, as opined for example in Aerotel Ltd v Telco Holdings Ltd; Macrossan’s Application, [2006] EWCA Civ 1371; [2007] RPC 7, can be useful in an analysis of an “artificial effect”.
SUBMISSIONS
After presenting the background to the present alleged invention, discussing the content of the specification and outlining the relevant legal principles, the applicant contested the examiner’s characterisation of the substance of the claimed invention.
For example, in the third report, the examiner described the substance of the alleged invention as a scheme for suggesting keywords or queries to a user searching for information on a computer. By contrast, the applicant submitted the substance of the invention must be determined by reference to the claim as a whole. Initially the applicant noted that the specification drew a distinction between a search query and a search query suggestion. A search query is the final query, typically comprising one or more words or phrases, or one or more complete sentences, that is submitted to the search engine for action. A search query suggestion is a suggestion to the user of possible terms that could be added to a partial search query to complete it before submitting it to the search engine. Furthermore the applicant submitted that, considered as a whole, the substance of the invention resided in clearly defined, computer-executed steps for automatically generating search query suggestions. Firstly this involved a server device generating a first search query suggestion based on a determination that a query portion received from a client device corresponds to at least one token portion stored in a token pool that includes a plurality of tokens also generated by the server device. Secondly the server device generated a second search query suggestion based at least in part on an expansion of at least a portion of the generated first search query suggestion, the expansion being performed based at least in part on a session history associated with the client device.
The applicant further submitted that, more importantly, once the generated search query suggestions are entered into an Internet search engine, they allow the user to locate desired information more quickly than would be the case if the user’s own search query was used or a search query generated using prior art auto-complete techniques. The applicant thus described the present invention as involving ingenuity in devising an algorithm that, once executed, results in tangible benefits in the form of more relevant and personalised search query suggestions and consequent improvement in the use of Internet search engines.
The applicant concluded that the problem of improving the quality of automatically generated search query suggestions is a technical one because solving the problem allows users to utilise search engines more efficiently to more quickly locate desired information.
DISCUSSION
Claimed Invention
Claim 1 defines a method using a system architecture comprising a server device, a client device and a token pool that stores a plurality of tokens generated by the server device. The devices are in communication with one another such that at least a search query portion from the client may be compared for correspondence with at least one token portion in the token pool. The server device generates at least a first search query suggestion based at least in part on a determination of such correspondence. The server device further generates at least a second search query suggestion based at least in part on an expansion of at least a portion of the first search query suggestion. The server device performs that expansion based at least in part on a session history associated with the client device. Both search query suggestions are then communicated by the server device to the client.
For example, paragraph [0018] of the specification is illustrative. It explains that the system may enhance search query suggestions by biasing, for example expanding, removing and re-ranking, standard search query suggestions with the user’s past search queries to generate more relevant and personalized search query suggestions. The user’s preferences in several areas like brand, size, gender, etc. may already be captured via past searches. Using this information and modifying the standard search query suggestions to take into account such implicit preferences may be more preferable to the user and reduce the time to quickly reach relevant search queries.
In exemplary embodiments, a native result module may generate one or more first search query suggestions, which may be first auto-complete results, based on the client’s search query, or at least a search query portion, and on tokens or token portions stored in the token pool. An expansion result module may further generate one or more second search query suggestions based on tokens or token portions and on the client’s session history. In this way, the second search query suggestions represent additional auto-complete results predicted or expanded by a search enhancement system to enhance the first search query suggestions. The specification presents an example at [0018] along the following lines. If the user previously entered the search query “gucci® handbag”, then in response to typing “belt” in the search box a search query suggestion for “gucci® belt” may be shown.
Claim 11 is directed to a computerized system comprising a client device, a token pool associated with the client device, and a number of modules. Four of the modules broadly perform equivalent functions to those of the server device of claim 1. Two further modules in claim 11 though perform the additional functions of merging at least the first and second generated search query suggestions and ordering the merged list based at least in part on a boosting operation that modifies a score determined for each search query suggestion in the merged list.
Claim 13 defines a non-transitory machine-readable storage medium comprising executable instructions that, when executed by a processor of a machine, cause the machine to essentially perform the method of claim 1.
Substance of Invention
The specification admits the generation of search query suggestions or the auto-complete function was known in the art ([0003], [0017] and [0018]). At [0003] for example, it is noted that, conventionally, search queries have included search query suggestions provided to a user while the user is entering a search query into a search field. At [0017], the specification states that the auto-complete feature is an important and integral part of any search system. It helps a user quickly populate and formulate search queries by suggesting popular and well-known search queries from the search system. At [0018], the specification discusses the enhancement or biasing of standard search query suggestions (my emphasis). Consequently, the substance of the invention is not in receiving a query portion from a client and generating a first search query suggestion itself.
It would also appear there is nothing of substance in the claimed method of such generation based on determination of the correspondence of at least the received search query portion with a stored token portion in a token pool. In this respect, some definitions from the specification appear relevant. At [0031], it is noted that query items may be understood as portions of a query, such as a keyword or portion of a keyword in a query. Paragraph [0036] further describes a query portion as all or a portion of a search query received from the client device. The query portion includes at least a token portion. The token portion may include all or a portion of a word, or a word set, for example a sentence or a paragraph, or similarly a number or a number set. A global token pool may be understood to be a set of tokens, token portions, query portions or search queries, including words and phrases or parts thereof, used by users or not previously used by users ([0037]). Moreover paragraph [0050] describes a token in terms of a sub-list of words in the query or the query portion. For example, the query portion “prada® eyeglasses” may have three tokens; “prada®”, “eyeglasses” and “prada® eyeglasses”. A token portion may be understood as a portion or fragment of a token. For example, where the token is “eyeglasses”, a token portion is a part of the term, such as “eye” or “eyegl”. All of this, including the matching of query portions with stored token portions to provide the auto-complete functionality for search queries, would appear to have been known in the art at the relevant time. For example Google® provided this functionality with Google Instant® in 2010.
The claimed invention then defines the generation of the second query suggestion based at least in part on an expansion of at least a portion of the first query suggestion. This may suggest an iterative, two-step sequential approach of generating the second query suggestion directly from the first, although that is not necessarily so. The claims in their broadest sense merely state the expansion is performed based at least in part on session history. Similarly the descriptive part of the specification reflects a separation between generation of the first and second search query suggestions. As mentioned above in respect to [0018] of the specification for example, the previous search query “gucci® handbag”, in response to now typing “belt” in the search box presents a search query suggestion for “gucci® belt”. Similarly, at [0062], the search enhancement system enhances search query suggestions, for example the native queue, by first expanding the user query based on session history and the token pool (my emphasis).
Figure 3, reproduced below, appears to illustrate a sequential flow of operations of the search enhancement system in some embodiments.
The flow includes the operations of receiving a query portion, generating a first search query suggestion set, accessing a token pool and generating a second search query suggestion set. On the other hand, the modules for performing those operations do so with elements of independence. At [0048], in operation 310, the receiver module receives a query portion including at least a token portion from the client device. In operation 320 (paragraph [0051]), the native result module generates a first search query suggestion set based on the token portion. From [0053], in operation 330, the access module accesses a token pool associated with the client device. At [0054], in various embodiments, the token pool is a pool that records tokens or token portions in a session history of a user. In some instances, the token pool may be distributed among session histories of individual users or client devices. At [0061], in operation 340, the expansion result module generates a second search query suggestion set based on the token portion and the token pool. Notably there appears to be no direct association of first search query suggestion results in generating second search query suggestions.
In summary, one may conclude from the specification, as a whole, that the second search query suggestion is not directly based on the first search query suggestion. Rather the second search query suggestion is merely an expanded version of the first and appears to be generated independently.
The specification indicates that second search query suggestions may represent additional auto-complete results. At [0061], the second search query suggestion set represents additional auto-complete results predicted or expanded by the search enhancement system. Since the first and second search query suggestions may thus both be auto-complete results, then the substance of the claimed invention cannot lie in their independent generation. Furthermore there would appear to be little of substance in the claimed input sources of generation of the first and second search query suggestions, that being a comparison of at least a user query portion with at least a token portion in a token pool in the case of the first search query suggestion, and at least the user’s session history in the case of the second. In the latter respect as claimed, one may also infer that the generation of the second search query suggestion may simply rely on the user’s session history for a word, token or symbol comparison in a similar way to the generation of the first search query suggestion. With reference again to [0061], second search query suggestions may be generated based on a token portion and the token pool associated with a user or user device. It may also be said that the generation of the first and second search query suggestions, either individually or in combination, through word, token or symbol association does not make a technical contribution to the art.
The thrust of the specification is that the second search query suggestions are an enhancement, and by example an expansion, of the first search query suggestions. It may thus be said, as indicated in the applicant’s submissions, that the search query suggestions are more relevant and personalised to allow the user to locate desired information more quickly. It may further be said that such targeted suggestions result in a tangible benefit and produce a technical effect. In the presently claimed invention though, both search query suggestions may be generated from little more than standard auto-complete techniques. In claim 11, the second search query suggestion is merely defined in terms of an expansion of the first search query suggestion. In claims 1 and 13, the claims merely define such expansion being based at least in part on a particular input, that being a client’s session history. It is not evident that the tangible benefit or technical effect arises from any computing functionality requiring more than mere generic computer implementation (RPL at [99]). Alternatively and similar to RPL at [96], there does not appear to be any technical contribution to the art where the computing system is integral to the claimed invention, rather than a mere tool in which the invention is performed.
I conclude the claimed invention is not for a manner of manufacture.
Whether There May Be Patentable Subject Matter In The Specification
There may be patentable subject matter in the body of the specification from which valid claims could be drafted to overcome the above finding. For instance, as mentioned earlier, paragraph [0018] of the specification provides an example. In response to now typing “belt” in the search box, the existence of a previous search query “gucci® handbag”, may result in a search query suggestion for “gucci® belt”. Notably a relevant second search query suggestion may be generated where there is no direct word, or partial word, association between the previous search query or query portion and the current search query or query portion. That is, it would appear the invention may reside in system operations that may use an algorithm that forms associations, through the token pool, and through client session histories and through some further means, to present search query suggestions that are aligned with the search query or query portion by an association other than just a word, token or symbol, or partial word, token or symbol, association. Arguably then, if there is substance in the alleged invention, it may be in a further, more specific method by which the second search query suggestion is generated. For instance, the above example appears to associate a handbag and a belt as the same or a similar class of article, and thus possibly from the same supplier or retailer, and hence offers “gucci® belt” as the search query suggestion. Paragraph [0018] mentions possible associations by user preferences for things like brand, size or gender from past user searches. Paragraph [0019] further discusses the search enhancement system using a global scoring function to rank queries across multiple users by features such as popularity to present search query suggestions based on the global scoring function and the user’s previous search queries, the generated search query suggestions thus reflecting the preferences and intentions of the user. In a similar way to [0018], paragraphs [00109] and [00110] suggest that a user’s previous queries over a particular brand and then subsequent query with just a new product will lead to search query suggestions associating that brand with that particular product.
There may be questions of whether an invention claimed along the above lines would have sufficient support by matter disclosed in the specification and/or whether there is disclosure that is clear enough and complete enough for such invention to be performed by a person skilled in the art (sub-sections 40(2) and (3)). For example, while various embodiments discuss new queries from a user that appear to involve products of the same or similar class as previous product and brand queries from the user, it is not readily apparent from the specification how a class determination is achieved to associate the new product query with a particular brand. Alternatively, it could be said there is no such class determination disclosed and thus one could conclude from the specification that the second search query suggestions associate brands, from previous client session histories, with new product queries even if the products would be completely unrelated to anything supplied or sold under that brand. In the latter respect it is then questionable whether the enhanced search query suggestions would be more relevant and personalised to allow the user to locate desired information more quickly as suggested by the applicant.
In a still further alternative, and using the above example, if the global token pool across multiple users or devices had previously associated the words “gucci®” and “belt” while the present user had not, then it would appear the above-discussed standard auto-complete functionality would generate “gucci® belt” from the query term “belt”, and thus there is no contribution apparent to the art in this case either.
The above matters would appear to be some pertinent issues for the applicant in considering the findings of this decision. Since I have not determined at present that there is no patentable subject matter in the body of the specification, I will allow an opportunity for the applicant to obtain acceptance of the application. The applicant may propose suitable amendments and file any further supporting submissions with a view to achieving that end.
CONCLUSION
The claimed invention is not for a manner of manufacture.
There may be patentable subject matter in the body of the specification from which valid claims could be drafted to overcome the above finding.
Pursuant to sub-regulation 13.4(1)(g), the final date for acceptance is now three (3) months from the date of this decision. The applicant is invited to file amendments and any supporting submissions addressing the findings in this decision. It would also appear advisable that the applicant’s amendments/submissions address the issues related to subs-sections 40(2) and (3) discussed above. In the event that such amendments and/or submissions are filed, the response will be dealt with in the normal manner of such responses to an examiner’s report.
M. G. Kraefft
Delegate of the Commissioner of PatentsAnnex A
1.A computer implemented method, comprising:
receiving, by a server device, at least a query portion from a client device to conduct a search based at least in part on the received query portion;
generating, by the server device, at least a first search query suggestion based at least in part on a determination that at least the received query portion corresponds to at least one token portion stored in a token pool that includes a plurality of tokes (sic) generated by the server device;
generating, by the server device, at least a second search query suggestion based at least in part on an expansion of at least a portion of the generated first search query suggestion, wherein the expansion is performed by the server device based at least in part on a session history associated with the client device; and
communicating, by the server device, at least the generated first and second search query suggestions to the client device responsive to the received query portion.11. A computerized system, comprising:
a receiving module configured to receive at least a query portion from a client device;
a token access module configured to access a token pool associated with the client device, the token pool including a plurality of tokens that were each generated based on search queries received from the client device prior to the received query portion;
a suggestion generation module configured to generate at least a first search query suggestion and at least a second search query suggestion, the first search query suggestion being generated based at least in part on a determination that at least the received query portion corresponds to at least one token portion stored in the token pool, the second search query suggestion being generated based at least in part on an expansion of at least a portion of the generated first search query suggestion;
a merge module configured to merge at least the first generated search query suggestion and at least the second generated search query suggestion into a merged list of search query suggestions;
an organization module configured to order the merged list of search query suggestions based at least in part on a boosting operation that modifies a score determined for each search query suggestion included in the merged list; and
a presentation module configured to cause presentation of at least the first and second generated search query suggestions on the client device based at least in part on a determination that the first and second generated search query suggestions each have a corresponding modified score above a defined threshold.13.A non-transitory machine-readable storage medium comprising processor executable instructions that, when executed by a processor of a machine, cause the machine to perform operations comprising:
receiving at least a query portion from a client device to conduct a search based at least in part on the received query portion;
generating at least a first search query suggestion based at least in part on a determination that at least the received query corresponds to at least a first token portion stored in a first token pool that includes a first plurality of generated tokens, wherein each token in the first token pool was generated based at least in part on a first search query received from the client device;
generating at least a second search query suggestion based at least in part on an expansion of at least a portion of the generated first search query suggestion, wherein the expansion is performed based at least in part on a stored session history associated with the client device; andcommunicating at least the generated first and second search query suggestions to the client device as a response to the received query portion.