Are you looking for ways to learn and find prior art on your own? Welcome to the DIY guide!
Whether you are an inventor, an entrepreneur, a student, a researcher, or a professional, you can identify any existing patents, publications, or other prior art that may be relevant to you or your work.
A prior art search can help you decide whether your invention is novel or non-obvious, which are the essential requirements for obtaining a patent.
If you are an individual researcher, such as a Ph.D. student, or inventor, you may not be able to afford a professional search for prior art. However, you may need it in your project to know what exists out there and what you need to improve upon to get an edge over previous work and stay updated.
If you are looking to make a career in the field of patents, learning how to search for prior art is an important first step.
You should be elated to know that there are many databases and tools available that you can use to your advantage at no cost to you except your internet bill.
It may so happen that you want a quick overview of the prior-art scenario in your field, however, opting for a professional search may be costly and time-consuming. In such a case, knowing how to search prior art using free patent databases and tools can serve you well.
If you are looking to file a patent, conducting various types of prior art searches can help you refine your patent application and, ultimately, save your business and investment. With search, you can mitigate many risks such as patent being denied, later being found to be invalid, etc.
Imagine yourself a Ph.D. student well versed in prior art search. In general, presently filed patents consist of technology from the future, and not all technology from patents is published in a research paper. Thus, it may happen that you might unknowingly end-up duplicating some work already claimed in a patent but not published in a research paper. You can avoid such a situation if you know the art of search. Similarly, by knowing what’s next to come in technology through patents, you can plan your research work that is aligned with future trends and have an edge over other researchers.
In this guide, we will explain to you step-by-step, the art of searching, including tips, resources, and strategies that we have gathered with our experience to search efficiently and effectively.
After learning the basics, you should be able to apply those concepts to any kind of search and modify your search to cater to your specific needs.
So, grab a cup of coffee, sit relaxed and start learning with us.
How to Conduct Prior Art Search Yourself for Free?
In this section, we will learn how to conduct a prior art search. Once we learn the basics of prior-art search, these can be applied to all kinds of searches, be it invalidation search, freedom to operate search, patent landscape search, etc.
Concepts of the search for prior art can be learned in one day; however, excellence will only come with time and more practice. More practice helps you gain experience and insights that you otherwise wouldn’t have.
You should also know that we are writing this guide assuming you are a search professional and have been tasked with searching for great prior art by your client. However, that assumption is only for setting the tone of the guide and the fundamentals of the search remain the same whatever the purpose of the search is.
For the purpose of your better learning, we will take up an example and use it as an anchor to this guide. In the end, we’ll provide a practice example so that you can get some hands-on experience and apply the skills learned in this article.
Without further ado, let’s look at the steps which you follow while conducting a search for prior arts.
1. Identify and understand the technology
The very first step is to identify what you need to look for in a search.
You may be given a set of technical features forming an invention, the invention itself, or a set of instructions directing you about what to look for.
The problem statement describing what to look for can be as per the requirement and convenience of the client, however, it should be enough to give direction to your search.
Suppose you are given a claim to invalidate by your client. Or you yourself want to assess the validity of a claim that may be of interest to you.
So the very first step is to understand what is being claimed. Unless you understand the claims, you won’t be able to search for them properly.
If you know the patent number of the claim then read the various parts of the patent document including the abstract, background, technical field of the invention, problem to be solved, a summary of the invention, and description along with drawings, claims, etc.
While reading the patent document, you come across many concepts that you may not be well-versed in. To better understand the concepts in question, don’t hesitate to browse the internet, watch related videos, and discuss with colleagues and experts.
Additionally, if possible, take input from the client himself about the technology and invention in particular.
Not only this, but you can also look for the file history or file wrapper of the patent application.
1.1 File history or file wrapper
From the file history, you can find out what objections were raised by the patent examiner, what were the responses from the applicant, what was the search strategy of the patent examiner, how the claims were amended over the course of the examination, and for what novel, inventive, and patentable subject matter the application was granted a patent.
You can check the file history of patent applications from IP5 offices (USPTO, EPO, JPO, KIPO, and SIPO) on Global Dossier.
1. For a US patent application file history, you should mainly look for “non-final rejection”, “final rejection”, “notice of allowance”, “list of references cited by examiner”, “examiner’s search strategy and results”, “applicant arguments/ remarks made in an amendment”, etc. among other things which you consider necessary.
Non-final and final rejection reports consist of objections raised by the patent examiner for the patent application.
In the notice of allowance, you may find “reason for allowance” which helps you understand the feature of the claim that played an important role from the novelty and inventive step point of view. Knowing this, you can focus on that feature in your search strategy.
The image below is showing a screenshot of the file history of a US patent application on Global Dossier.
2. Similarly, for the file history of a European patent application, you can mainly look for “European search opinion,” “supplementary search report,” “amended claims with annotations,” “amended description with annotations,” “annex to the communication,” “amendments received before examination,” “text intended for grant,” etc. among other things that you consider necessary.
Here, the European search opinion and annex to the communication are objection reports by the patent examiner to the patent applicant which you definitely want to check out.
Since European Patent Office is one of the IP5 offices, you can find the file histories of patent applications on Global Dossier.
3. For a file history of patent applications in the Indian Patent Office, you can visit INPASS.
You may also go through the cited documents, i.e., perform citation analysis, to see what the alternative ways are to claim similar technology. You can identify some keywords and different ways of representing the same technical concept.
By now, we believe your understanding of the claimed technology is complete.
1.2 Simplify the claim
It’s time to test your understanding of the claims of the patent that you acquired in the previous section.
Try to simplify the claim such that it can be understood by a layman or by a person who is not well-versed in the world of technology and patents.
It’d be even better if you could think of a real-world example for the invention and then incorporate that example into your simplification.
Again, use that simplification to write the crux of the invention in one or at most two lines.
If you can’t do this, you need to further work on your understanding of the claim(s).
For example, let’s consider a claim that we’ll be using throughout this guide:
“A method comprising:
US84128185
executing an online service by a first device;
initiating communication, from the first device to a second device, over a near field communication link during the execution of the online service to coordinate the transfer of information relating to the online service; and
initiating the transfer of information relating to the online service to the second device based on the communication for execution of the online service by the second device independently from the first device.”
We can simplify the claim even further by writing something like this:
– a method for transferring information from first device to second device over a near-field communication link, while executing an online service.
– The second device will then use the transferred information to execute the service independently of the first device.
Next, we can write the claimed invention in just one line:
| Transferring an executing service over NFC from one device to another. Or Switching devices for a service during execution over NFC. |
These may not be the only ways to write the claim above in simplified form. You may arrive at your own simplified statements, including one-liners that represent the crux of the claim.
Let’s look at one example to see how the claimed invention is implemented:
Real World Example: Suppose you are listening to music on your smartphone (first device) and want to continue listening to your speaker (second device) with great sound. For this, you initiate communication from the smartphone to the speaker over near-field communication technologies such as Bluetooth. As a result, the song that is currently playing on the smartphone and the playback information is transferred to your speaker. This way, you can listen to the song on your speaker independently of your smartphone.
Note: The purpose of simplifying the claim to one or at most two lines and exemplifying it is to fully understand the claim so that you can know exactly what you are searching for.
Many times you will feel that you can’t simplify the claim any further. However, with practice, you will improve.
After identifying and absorbing the claim or technology of the invention, you are ready to move on to the next stage of the search, which is to gather material for a search strategy.
2. Brainstorm and Gather the Relevant Data
In the previous step, you did everything that you could to understand the claim, the patent, and its various aspects. It’s time to brainstorm your options. You should now look for all the keywords, classes, top assignees, top inventors, citations, etc. so as to form an informed search strategy.
2.1 Find keywords
As we discussed previously, you come across many keywords and their synonyms while
- browsing the internet,
- asking questions on forums,
- watching related videos,
- discussing with colleagues and experts,
- performing citation analysis,
- searching for classifications,
- exploring AI tools,
- going through the problem statement of client,
- going through thesaurus, etc.
You should collect these keywords.
Note: There are many things that AI tools can do. In fact, in our opinion, it’s best for you to play with and explore AI tools for their utility in the field of patents.
Upon looking at the claim, we can extract the following keywords or concepts: executing, online service, initiate communication, near field communication, coordinate, transfer of information, independently, devices.
Among these keywords, many are not useful in search. Since we use a keyword or class to filter the search results, our aim is to find and use only those keywords that matter.
For that, our next approach should be to devise a relationship among keywords to form a query in such a way that you can communicate effectively with patent search engines or databases about what you are looking for.
Now let’s go back again to our one-line statement of the claims, i.e.
Transferring an executing service over NFC from one device to another.
or
Switching devices for an service during execution over NFC.
From this, we can clearly see that near-field communication, service, execution, and transfer are the important keywords.
You may be wondering, why “device” is not a keyword of interest.
We can think of two reasons for this:
1. the ‘device’ is too broad as a term to contribute something meaningful to the search query and has a large number of technical synonyms like mobile phone, smartwatch, computer, tablet, etc.
We use keywords to filter and shortlist the patent documents. For example, out of millions of patent documents, a keyword-based query can filter and shortlist the results up to a certain number, say 10 or 20, or 100 documents. As humans, we can easily analyze these 10, 20, or 100 patent documents.
So, including the keyword “device” and the large number of its synonyms in the search query doesn’t make sense from the filtering and shortlisting point of view, as every communication-related invention or patent has these devices mentioned in them.
If you can’t filter or differentiate search results using a keyword, then why use it in the first place?
2. We know that the crux of the claim i.e., “a service execution and its transfer over near-field communication,” can only occur in a device similar to that of our invention.
Therefore, if we include keywords such as “service,” “execution,” “transfer,” and “NFC,” then the search result documents will automatically have “devices” in them similar to that of our claim.
By following such logic, you can decide on redundant keywords.
After what we have done so far, we have been left with the following keywords: “Near field communication,” “service,” “execution,” transfer
We can even remove “execution” from the list of search keywords. Any patent document disclosing a service in a device is going to talk about its execution; otherwise, what is the point of talking about a service?
Therefore, the use of the keyword “service” can ensure that the prior art document also has the keyword “execution” or its synonym.
Let’s take a few more similar examples for understanding.
- Suppose you want to search for lights in a vehicle then simply using the keyword “headlight” will ensure that you don’t have to use the keyword “vehicle” separately.
- Another example would be, if you use the keyword “computer” then you don’t have to worry about the keyword “processor”.
It’s like, if a document talks about a processor, then it automatically includes a computer. If a document talks about a headlight, then it automatically includes a vehicle.
So, by this logic, even the keyword “execution” becomes redundant at least for the time being. Because we can always come back and examine whether to include or exclude keywords based on the results of the previous query.
We are now left with only the keywords: “Near field communication,” “service,” and “transfer”
After identifying the actionable keywords, it’s better if we look at their synonyms, gathered through the various approaches mentioned previously.
| Keywords | Synonyms |
| Near Field Communication | NFC, Bluetooth, RFID, Zigbee, UWB (Ultra-wideband), infrared communication, Wi-Fi direct, Bluetooth Low Energy (BLE), Z-wave, Wireless USB (WUSB), FireWire, Li-Fi, SHARP (short-range-device-profile), short-range communication, dedicated short-range communication (DSRC), low energy communication (D2D services, PROXIM services, WIFI Direct services, BLE services) Note: If it is about invalidation, you can ignore technologies that came after the patented invention because for invalidation we need prior technologies. |
| Service | application, session, app, software application, web application, call Note: these services can be anything running on a device to serve. For example, stream services, social media platforms, online marketplace, online banking services, cloud storage services, online gaming services, navigation and mapping services, online education platforms, voice assistants, health and fitness tracking apps, video calling apps, etc. |
| Transfer | switch, migrate, transmit, relocate, shift, move, handover, pass on, delegate, substitute |
2.2 Find classification
Next, your job is to find classes that define the features of your invention. Classification can be a wonderful tool if you happen to find accurate classes.
If you are given a granted patent for invalidation search then most likely, it will have classes assigned to it.
However, you should know that there is always the possibility of finding a better class than the ones assigned to the granted patent.
So, you may start the search with the classes assigned to the granted patent, but remember to look for better classes, if there are any, in the next iterations of the search.
What do we mean by all this? How do you find one or more classes of claims for a patent?
There are many ways to do all this. Mostly, we use IPC and CPC for searching prior-art.
The CPC classification system has been developed by the European Patent Office and the US Patent Office. It is a more accurate, precise, and relatively new classification system. It consists of about 200,000 subgroups.
The IPC classification system, on the other hand, has 70,000 subgroups and is used by patent offices all over the world to classify their patent documents.
In many areas, you will find that IPC and CPC overlap with each other.
There are other classifications also, such as USPC (the US patent classification, which is slowly being phased out), Japanese FI/F-term classification, etc. But mainly, we are going to focus on IPC and CPC classes for the purpose of this guide.
So, what are the ways you can find classes for your patent or claim?
2.2.1 Espacenet- patent classification
- You may visit the espacenet classification search and type the keywords in the search box. As a result, some of the CPC classes are retrieved in order of relevancy. From there, you can select the CPC that you think is useful for your search. or
- You may search the class manually with a top-down approach.
- For example, to search CPC for the feature of “theft alarm due to breaking glass”, you first move to class G: Physics; then in Class G, you move to G08 which includes alarm system in its definition; then upon manually screening through classes, you come across G08B13/00 which talks about theft alarms; then you further look inside this class. Finally, you end up reaching G08B13/04 which is for a theft alarm due to breaking glass. Basically, you start from image 2.2 and reach image 2.3 below.
2.2.2 IPCCAT
You may find IPC classes by using the AI tool IPCCAT provided by the WIPO. In this tool, you simply provide the invention in text form, and the tool gives you 3 or 5 (see box 4: Number of predictions in image 2.4) probable IPC classes as a search result.
It may so happen that the text entered in the tool is not properly representing the invention, in which case this tool may not give an accurate IPC.
To overcome this, you have to make sure that you provide a better description of the invention to this tool and don’t just copy and paste the abstract or claim into it.
Copying and pasting the text directly from the abstract of claims may work sometimes, but you have to be vigilant and change the text if it doesn’t give the right IPC.
For example, upon pasting the whole claim in box 2 (selected by Search tab 1), we get 3 IPC predictions (box 3), namely H04W 4/80, H04L 9/32, and H04B 5/00.
Note: You can change the classification level in box 5 of image 2.4 above. You have the option to choose class (e.g., H04), subclass (e.g., H04W), main group (e.g., H04W 4/00), subgroup (e.g., H04W 4/80).
To further investigate H04W 4/80, either we can click on it in box 3 of image 2.4 or search for it in box 2 of image 2.5 by clicking scheme tab 1.
We have chosen the path view (box 3 of image 2.5) among many views for taking a screenshot, which is now image 2.5.
We find that H04W 4/80 relates to services using short-range wireless communication, e.g., near-field communication, which is accurate. The definition statement we get for this IPC upon clicking icon D (Box 4 in image 2.5) is as given below:
| This place covers: Providing communication services using short-range communication, e.g. nearfield communication [NFC], radio frequency identification [RFID], dedicated short-range communication [DSRC], or low energy communication, e.g. BLE for enabling proximity services (D2D services, PROXIM services, WIFI Direct services, BLE services) |
Looking at the definition of the classes helps gather keywords and synonyms just like the box above.
Next predicted IPC H04L 9/32 (image 2.4) relates to arrangements for secret or secure communication including means for verifying the identity or authority of a user of the system, which is a little off from the topic and is not relevant to us.
The last predicted IPC is H04B 5/00. It relates to the near-field transmission system (transmission systems of a limited range wherein the information signal transmission is inductive or capacitive, rather than electromagnetic, e.g., the inductive loop type) which is also relevant to our claim.
This way, we have gathered two IPCs, namely H04W 4/80 and H04B 5/00, and since both relate to near-field communication, we will use them with the boolean operator “OR”.
2.2.3 Observe the classes of retrieved results in response to search queries
Another way to search classes is to make a query based on the keywords and available classess and run the query. You need to do this exercise over and over again by refining the query until you see some relevant results. When you observe that you have started to get relevant results, you need to look at the classes of search results.
It’s recommended that you make broad search queries. We will discuss it later in this article.
Most of the prior art search tools and databases have this feature of grouping retrieved results by classes.
You need to pick up the relevant class out of them and use that class to further refine your search query.
While you add the class in your search query, don’t forget to remove the corresponding keyword of that class.
For example, if you are using IPC H04W 4/80 in your search query, you don’t need to use any keywords related to near-field communication covered under the definition of this class.
By doing multiple iterations and refinements of the query, you will come across many good classes.
Suppose you are searching for invalidation of a patent then that patent will already have classes assigned to it. You can pick up those classes and use them in your search queries. If you think that you can find even better classes than the ones assigned to the patent, then you can do so by using the above procedure.
Basically, use anything available at your disposal to your advantage.
Note: You don’t need to use any cut-off date in your query while searching classes.
2.2.4 Find classes using a search query
Let’s look at an example of how you can find classes using search queries.
For the claimed invention, we have already collected two IPCs, which are H04W 4/80, H04B 5/00, and relate to near-field communication.
We see that the patent for the claim has many assigned IPCs. One of them is H04L67/148: Migration or transfer of sessions.
You can search for this class yourself, too. For example, let’s run this query in Google Patents:
((NFC OR “near-field communication” OR Bluetooth OR RFID) ((service OR session OR application) NEAR10 (transfer OR switch OR migration))) before:filing:20090913
As Google Patent shows the data of the first 1000 results, the top 5 CPCs of those 1000 results are H04W, H04L, G06K, G06Q, and H04M. See the image below:
From here, you can literally check manually by going to the espacenet cpcbrowser or IPCCAT.
Note: Google Patents also has the feature of grouping the results by classification.
Let’s consider H04L for further exploration.
We see that H04L: Transmission of digital information. We know that our claimed invention also transfers or transmits service information, i.e., digital information. Upon manually scanning this H04L, we find H04L 67/00: Network arrangements or protocols for supporting network services or applications. Since we know this is a relevant IPC, we can stop here or dig deeper into it.
Further, when we go inside, we observe the relevant class, i.e., H04L 67/14: Session management. Since this class has child classes, we went one step ahead and found H04L 67/148: migration or transfer of sessions.
| The complete definition of H04L67/148 would be Electric communication technique-> transmission of digital information-> Network arrangements or protocols for supporting network services or applications-> session management-> migration of transfer of sessions. |
Thus, we have H04L 67/14 and H04L 67/148 which can be used in the search.
You might argue why H04L 67/14 is relevant in addition to H04L 67/148?
- One argument is that H04L 67/14 is the parent class of H04L 67/148 and therefore has broad coverage. We can use the parent class with a search operator that allows us to cover the lower hierarchy of the parent class. For example, in the Google Patents query (CPC=H04L 67/14/low), the ‘low’ operator allows you to cover the class H04L 67/14 and its child classes in the hierarchy.
- Another argument comes from its definition, i.e. session management, when seen with the definition of H04L, i.e. transmission of digital information.
- So, H04L 67/14 involves prior arts that have session management involving the transmission of digital information which is relevant to our search.
Since we are searching with Google Patents, the top 5 CPCs of the first 1000 retrieved results shown are at the class level, i.e., H04L, H04W, etc.
However, when you search in paid databases, you will see complete classes up to the subgroup level, i.e., H04L67/148. So, in those search tools or databases, for retrieved results of a query, you can directly observe the top-most classes up to the subgroup level without having to go and check the hierarchy of classes manually.
Anyway, whatever method you apply, the aim is to get the closest class so that you can perform the search effectively.
Now, we have the following keywords: “near field communication,” “service,” and “transfer”.
Moreover, we have the following classes: H04W 4/80, H04B 5/00 (both relating to near-field communication), and H04L67/148 (relating to the migration of session i.e. online service in our case).
Now that we have keywords and classes, we can look for the prior art documents efficiently.
Since we have Google Patents as a free tool, we will learn to search by running queries on it. The fundamentals of searching for prior art are the same in every search tool or engine except that they have different ways of writing syntax, different operators, etc. for similar things.
If you know the fundamentals, you can perform a search on any database after learning its syntax, operators, command instructions, etc.
3. Start Searching for prior art
So far, you have acquired a good understanding of the claimed invention. Additionally, you have collected keywords along with their synonyms and classifications. It’s time that you start searching, now.
Let’s first quickly see what the cut-off date restriction is for a search.
3.1 Cut-off date restriction
As you are aware that we are searching for prior documents. So, there has to be a date to restrict our search then only we can look for documents prior to that date. We call it cutoff date.
If we are talking about the invalidation of a US patent, then we have to remember the following eligibility criteria for a document to be considered prior art.
1) If the prior art document is a US patent document, then its filing date should be prior to the cutoff date; and 2) if the prior art is non-patent literature or a non-US patent document, then its publication date should be prior to the cutoff date.
There are many other factors also in deciding what date will be used for a document to treat it as a prior art but the basic concept should be clear to you to get started searching.
Generally, for invalidation of an Indian patent, prior art should be published before the cutoff date.
3.2 Query 1
(H04L67/148) ((“near field communication”) OR (NFC)) before:filing:20090913
In the above query, you can see, all essential features of the claim have been covered.
Because CPC used covers the transfer of service from one device to another and the Keywords used cover the transfer of service occurring on near-field communication.
The date used in query 1 is such that prior art should have a filing date before the cutoff date because we are searching for invalidating a US patent.
However, one thing to note in this query is that we have not used synonyms for “near-field communication.” That’s because we are initially focused on a narrow search. You will see more about it in the coming sections.
After analyzing the retrieved results, we found some good documents.
One of them is US7983614B2: Handover for audio and video playback devices.
Let’s call it document D1.
The following table is important for you to understand how you map the features of a claim to the prior-art disclosure. You should know that you can cite any part of the prior art if that part happens to disclose the feature(s) of the claim.
When we prepare a report for the client or for any other purpose, we use a similar format with variations as per the requirement. However, we have decided to teach you what goes into preparing a report at the end of the search and how you correlate the retrieved results with the features of the invention or claim.
3.3 Prior art mapping
| Key features of our claim | What prior D1 discloses corresponding to the features of our claim |
| A method comprising: | a handover method (see claim of D1) |
| executing an online service by a first device; | Playback of a media work is handed over from a source playback device to a target playback device while the source playback device is in a play mode (see abstract of D1) Researcher’s Comment: play mode means the service of playing is being executed by the first device, i.e., the source device. It is not mentioned that the music is being played online but it has also not been denied. It is common to listen to music online. Therefore, it can be concluded that implicitly D1 discloses online service as playing music. In fact, in column 4, lines 19-20, D1 talks about both devices having access to the same content and doesn’t limit that it has to be offline or online content. So, the teaching of D1 from this disclosure is that both devices should have access to the same content doesn’t matter how. Therefore, it also covers “online” music as well. |
| initiating communication, from the first device to a second device, over a near field communication link during the execution of the online service to coordinate the transfer of information relating to the online service; and | In an exemplary embodiment, the playback devices are equipped with a near-field communication (NFC) interface that enables data transfer when the source playback devices are placed in close proximity. When the source playback device is in a play mode and detects a compatible target playback device, the source playback device transfers playback information to the target playback device. (see summary of D1). Researcher’s Comment: source playback device (i.e. the first device) detects (i.e. initiates) the target device (i.e. the second device) to transfer the playback information (coordinating the transfer of information relating to the online service) while in play mode (I.e. during the execution) over NFC. |
| initiating the transfer of information relating to the online service to the second device based on the communication for execution of the online service by the second device independently from the first device. | receiving, responsive to the detection of the source playback device, playback information at said target playback device from said source playback device for a media work currently being played by said source playback device; using the playback information to continue playback of the same media work by the target playback device. (see claim 1 of D1). Researcher’s Comment: responsive to the detection (i.e. initiating), playback information (i.e. information relating to the online service) is received by the target device or transmitted by the source device (i.e. transfer of information). Target device continues to play (I.e. execute) the same media using playback information (i.e. second device after receiving the information relating to the online service, executing the online service independently from the first device). |
3.4 Narrow and broad prior art search
While we are at this stage of the search, it’s good to know what a narrow search and a broad search are.
Every patent professional or prior art searcher has their own search style. Some professionals at the beginning of the search project go for the narrow search and then later conclude the search by conducting the broad search.
Some professionals also do the opposite. They start the search with a broad approach and later arrive at a narrow search.
Some professionals use a mixed approach from the beginning of the project.
What approach is to be followed comes down to the practice of the searcher, which comes with experience. However, for you, it is important to understand the difference.
| Narrow Search | Broad Search |
| It is useful when the time to search is short. Because this approach retrieves small patent datasets to analyze. | Preferred when the time to search is sufficient so as to conduct an exhaustive and reliable search to a certain degree. Because this approach retrieves a large patent dataset. |
| Boolean operator “AND” is used, frequently. Because it narrows down the scope of the search. For e.g., CPC=(H04L67/00 AND H04W15/00) | Boolean operator “OR” is used, frequently Because it broadens the scope of the search For e.g., CPC=(H04L67/00 OR H04W15/00) |
| Whether it is keyword, class, or other Search field, we use only those fields that are highly relevant and do not give noise in results. Example: for the above claim, in query 1, we used the highly relevant keyword “near field communication” or NFC and not other synonyms that we wrote in the table. We could use other synonyms also in the same query, but we didn’t. There is no fixed rule. Search is hit and trial but with a certain degree of confidence that comes with experience. In fact, we did use other synonyms of NFC but not all in query 2. | We try to cover as much as possible so that knowingly, we don’t miss any prior art and are able to conduct an exhaustive Search. For this, we use all possible synonyms, alternatives, variations of a keyword or at least the important ones of the keyword. The same is true for classes and other search fields. |
| The use of wildcards is somewhat less than the broad search for obvious reasons. We prefer to use exact search fields in this approach. For e.g., CPC=(H04L67/148) | To widen the scope of the search we use the wildcards. For example, to denote zero or more characters, we use a truncation operator such as ‘*’ in the Google Patent search engine. So, if we use ident* then it will represent all words made using ident such as identity, identification, identify, etc. Similarly, query CPC=(H04L67/low) will find documents that contain this CPC or its hierarchical child classes of in Google Patents. Every search tool or engine has different operators for the same purpose. For example, instead of “low“, some tools may use “+” or “*” sign to cover the CPC and its hierarchical child classes. for e.g., H04L67+ or H04L67* |
3.4.1 Broad prior art search
Let the relevant classes be C1, C2, C3, and C4. These classes may have been found by you or may already be assigned to the granted patent claims.
Let the keywords for features be K1, K2, K3, and K4. Further let the synonyms for K1 be K1’and K1’’, for K2 be K2’ and K2’’, for K3 be K3’and K3’’, for K4 be K4’ and K4’’.
Then you can formulate the following broad queries:
Q1: (C1 OR C2 OR C3 OR C4)
Suppose class C2 is not well defined or properly assigned to the granted patent, then you can use a wildcard to refer to it as class C2+, for example. This way, you can cover the hierarchy of class C2 in the hope that some relevant child classes will be covered by it.
Different search tools use different symbols for wildcard operators. Here we have used the ‘+’ sign to indicate the operation that covers the child classes, including the class itself. Google Patents uses ‘low’ operator for the same purpose.
We can now make another query when C2 is not properly assigned or well-defined as:
Q2: (C1 OR C2+ OR C3 OR C4)
Q3: (K1 OR K1’ OR K1’’) OR (K2 OR K2’ OR K2’’) OR (K3 OR K3’ OR K3’’) OR (K4 OR K4’ OR K4’’)
Since the above examples of queries are broad in nature, they will retrieve a large set of results. Naturally, it is difficult to analyze large datasets of patents. However, many search tools and engines have sorting features by relevancy. This way, you can analyze the top-most results after sorting.
Moreover, you can benefit from this large set of retrieved patent results by observing their classification. Most search tools provide the facility to group retrieved results according to classes. There, you can find relevant classes and further incorporate the same class into your search strategy.
Another thing you can do with a broad query is to observe who the top assignees, inventors, and jurisdictions are. These important findings can be used for targeted search which we will see next sections.
Another thing you can do with various broad combinations of classes and keywords is you can find multiple relevant documents.
You can combine these relevant documents to arrive at the subject matter of the claim that you are searching for.
By proving that a person skilled in the art (in India) or a person ordinarily skilled in the art (in the US) had access to these multiple documents, i.e., prior art, he will find it obvious to come up with the claimed invention.
In other words, the claimed invention doesn’t fulfill the criteria of non-obviousness there should not have been patented in the first place.
3.5 Let’s try some more queries
Let’s broaden the scope of the search compared to query 1. In query 1, we limited ourselves to session migration over near-field communication and NFC. Now, let’s use other modes of communication from the synonym table.
Query 2:
(H04L67/148) (Bluetooth OR “short range communication” OR wifi OR PROXIM OR “Bluetooth low energy” OR BLE OR RFID OR Zigbee OR UWB OR “ultra wideband” OR “z wave” OR firewire OR “LI-Fi”) before:filing:20090913
Another query we can make is using H04L67/14 as it also relates to session management while involving the transmission of digital information.
Query 3:
(H04L67/14) ((“near field communication”) OR (NFC)) before:filing:20090913
Query 4:
(H04L67/148) (Bluetooth OR “short range communication” OR wifi OR PROXIM OR “Bluetooth low energy” OR BLE OR RFID OR Zigbee OR UWB OR “ultra wideband” OR “z wave” OR firewire OR “LI-Fi”) before:filing:20090913
Upon running these queries in Google Patents, we got many relevant results. Some of them are given below for you to analyze.
US20110055627A1, WO2007149625A2, US20090210536A1.
For exercise, make a feature mapping table similar to this. Assess whether these prior arts take away the novelty or inventive step of the claimed invention.
For simplicity, when all the features of a claim are disclosed in one document, the claim is considered not novel and if you can combine multiple documents to arrive at the features of the claim, then the claim is considered obvious or not inventive.
Another query that you can make is to use only CPCs and sort the results by relevancy to check the top results, as shown below.
Query 5
(H04L67/14 OR H04L67/14 OR H04W 4/80 OR H04B 5/00) before:filing:20090913
Similarly, you can make a query using classes related to near-field communication and keywords related to service transfer.
3.6 Search strategy considering novelty and inventive step (non-obviousness)
So far, we have made search queries that aim to find all the features of the claim in one prior art document because we have included either classes or keywords related to each feature. Such a prior art document that discloses all the features of a claim is said to be a novelty-destroying document.
But that is not the only approach that you can take. If you are able to find a novelty-destroying document, then that is the best-case scenario. However, many a time, it happens that you don’t find a document disclosing all the features of a claim. In that case, you make a conscious decision to search for the prior documents that prove the claimed invention obvious.
For obviousness, we generally, combine more than one document to conclude that the inventor looking to solve a problem would find the solution obvious in view of those multiple prior art documents. This is called the mosaicing of documents.
Therefore, we can argue that the claimed invention doesn’t fulfill the criteria of non-obviousness.
Note: Normally, for a claim to be patentable it has to be new (i.e. novel), inventive and useful. To invalidate the claim, we try to prove that the claim was not novel or inventive, and therefore, should not have been granted a patent in the first place.
It is always recommended to incorporate searching for obviousness to perform an exhaustive search. Because you may never know that you may end up finding a novelty destroying prior art while searching the documents for obviousness which is a blessing in disguise.
Generally, in a patent or in an invention, you will find that not all features are that new and non-obvious, as some of the features are already part of common knowledge. It’d be one or two features that make the whole claim novel and inventive.
3.6.1 Identify features of a claim contributing to novelty and non-obviousness
If you look at claims of a European Patent, there are some features before the phrase “characterized in that” and some after that. Look at the sample claim below, for example.
Basically, this is the way to write a claim in two-part form. The features before the phrase “characterized in that” represent the subject matter already known from the prior art and the features after the phrase represent the novel and inventive features.
Except for European patents, you may or may not find such characterizing phrase. However, it is well established that an invention is built upon existing knowledge i.e. prior art.
So, to represent the invention as a whole, some features of a claim belong to the prior art and some belong to the solution over the prior art.
It is your job to identify what are the characterizing features if you want to excel in prior-art searching.
Once you identify such features, you’ll know what features contribute to the novelty and non-obviousness of the claim. Focusing on these features in your search strategy will make your job efficient and effective.
There is also a practice of using the word “wherein” to indicate characterizing features that you may keep in mind.
For US patent, if the patent is already granted, you will find notice of allowance in its file wrapper. Therein, you may find one or more features of the claim that were considered novel and non-obvious.
In fact, by reading the communication between a patent examiner and the applicant of the patent, you can identify important features of a claim. You can find file histories of the top 5 patent offices from Global Dossier.
If you are searching for the features of an invention provided by the client and there is no granted patent to look for, in such cases, the client himself tells you about the important features to be focused upon.
3.6.2 Search by features
Now consider that there are 4 features of a claim, namely F1, F2, F3, and F4. Next, you find the features contributing to the novelty and non-obviousness of the claim or invention.
When you search for these features, you will also find corresponding keywords and classifications, where multiple keywords or classifications may exist to represent one feature.
In previous sections of this article, we focused on searching for a document that had all the features of the claim. However, that approach is good for a novelty-destroying document.
In the event that we are not able to find such a document, we need to take a second approach and look for multiple documents. By combining them (i.e., mosaicing), it can be proven that the invention in the claim is obvious to a person skilled in the art (in India) and to a person ordinarily skilled in the art (US).
So, to look for documents that have some of the features individually and all features when combined together, we can apply the following strategies:
F1 AND (F2 OR F3 OR F4)
F1 AND F2 AND (F3 OR F4)
F1 AND F2 AND F3 AND F4
To look for multiple relevant documents having different-different features of the claim, you can incorporate broad approach to search.
3.7 Assignee-based search
Assignee-based search is targeted at a key player that you find to be dominating the field of the claimed invention that you are searching for.
One way to find who are the top players in the field of the claimed invention is to run broad queries and see analyze the retrieved results.
Another way is to check citations and prior art if any.
Because if you have found a relevant prior art then there are chances that the assignee or owner of the prior art may have similar patent or non-patent literature in the same technology. Thus, you can unearth good documents.
Most of the tools, including the free ones, provide the facility to search and group results by assignee. So, it should not be an issue as such.
The next thing you can do is to look for documents based on inventors.
3.8 Inventor-based search
It is common sense to conclude that most researchers do not jump into their research fields abruptly, and their growth in their respective technological fields is gradual.
Keeping this in mind, you can first find out who the top players are in the field of the claimed invention that you are searching for. Then you can incorporate those key players into your search strategy.
One way to find the key inventors and pioneers in the concerned field is to run broad queries and observe the retrieved results.
Another way is to identify good prior art and citations and see their inventors.
Now, that you have the names, you can use the search tools and databases at your disposal to look for the patent and non-patent literature by those top players.
3.9 Jurisdictions and non-English search
Mostly, you search for worldwide English documents. However, some technologies are more developed in some nations than others. Similarly, there may be a lot of literature lost in the translation.
By running a broad query, you may observe which country is most advanced in the concerned technology. Similarly, with experience, talking with experts, seeing the origins of top citations, and doing research on the internet, you may identify the leading countries in the field.
Therefore, you may focus your efforts on the inventions and related literature emerging out of that country, and in the process, you may unearth some relevant prior arts.
Further, a lot of technical literature is available in non-English documents. For example, there may be a plethora of scientific and technical literature available in German, Chinese, Korean, and Japanese.
Missing such opportunities may cost you or your client a lot. So, incorporating language-based literature in your search may turn the tide of your search. Moreover, a language-based approach helps you perform an exhaustive search.
4. Citation Analysis
If you are given a granted patent claim to search for invalidity, the first thing you should do is latch on to the citations.
Citations are very useful in understanding the invention and analyzing what a prior that you are going to search for may look like. These are a great way to understand the evolution of technology. Using these citations you can identify important classifications, keywords, and their synonyms, top players, top inventors, etc.
The next you can do is to perform backward and forward analysis.
4.1 Backward citation analysis
Backward citations to a patent are those documents or prior arts which existed before the patent and were referenced during the patent application phase. There are many ways a citation may be attached to a patent.
For example, documents may be cited by the examiner, by the applicant himself, by the opposition party, by the PCT examiner in PCT phase, by novelty search, etc.
These backward citations are important because the origin of the patented invention lies hidden in these documents. What can better resource to understand the background of the invention than backward citations?
Further, crucial insights such as key players and inventors, relevant classifications, etc. are readily available in these citations for us to exploit.
4.2 Forward citation analysis
Forward citations to a patent are those documents that came into existence after the patent was published and subsequently made reference to the very patent.
Analyzing forward citations can help you identify technical advancements in the field and establish the trajectory of the technology when looked at in combination with backward citations.
Further, it’s not only the backward analysis that gives you information about keywords, assignees, inventors, classification, etc. You can leverage forward analysis as well.
The document cited in the patent (backward) or documents citing the document (forward) are considered the first level of citations.
There are search tools that allow a second level of analysis by checking the backward and forward citations of the first level of citations of the patent, such as:
- Backward citations of backward citations of the patent
- Forward citations of backward citations of the patent
- Backward citations of forward citations of the patent
You can create a bucket of all citations and filter them using classes, keywords, cut-off date, etc. to shortlist relevant documents among them.
Further, you can find out the key assignees, key inventors, and relevant classes from the group of these citations. The information thus retrieved can be incorporated into your strategy for productive search.
Sometimes, it may happen that you are not able to locate good prior art through the normal search strategies that we have discussed earlier. In such a case, you may find it in citation analysis. Therefore, its importance cannot be stressed enough for an effective, efficient, and exhaustive search.
5. Analyze, Map, and Prepare Report
Your search project is about to finish, but before that, you need to present your work in an organized and comprehensible way so that a reader can easily understand your findings and recommendations in order to make an informed decision.
In the search process itself, you shortlist the relevant documents, analyze them, mark them according to relevancy, etc. Once you are done with every kind of approach that you can take during the search to make it exhaustive, you map all the features of the searched claim to relevant documents.
In a report, you try to communicate everything that may be of importance to the reader of the report. So, you include:
- a list of key players in the field,
- top inventors,
- important classes with their definition: IPC, CPC, US classes, etc. (you may also include FI and F-terms Patent Classification, )
- keywords used for search,
- search queries that were run on various patent databases,
- list of most relevant prior arts found by you,
- list of prior arts that are worth mentioning although not the most relevant,
- checklist of key features disclosed in those prior arts,
- list of forward and backward citations,
- list of family members of prior arts,
- claim feature mapping to prior arts, etc.
- visual representations if any such as done in landscape study, etc.
The report, consisting of the items in the list above and any others you think should be included, can be prepared in word, excel, or presentation format.
In the list above, all but one may not be clear to you. That is the “checklist of key features”.
It is nothing but a simple table having key features of a claim on the left side and indications of them being present in the prior art on the right side as follows:
| Key features of a claim | Prior art 1 | Prior art 2 |
| A method comprising | ✔ | ✔ |
| executing an online service by a first device; | ✔ | ✔ |
| initiating communication, from the first device to a second device, over a near field communication link during the execution of the online service to coordinate the transfer of information relating to the online service; and | ✔ | X |
| initiating the transfer of information relating to the online service to the second device based on the communication for execution of the online service by the second device independently from the first device. | ✔ | ✔ |
In the table above, if all the key features are present (✔) in a relevant document then it is a novelty-destroying prior art. If one or more features are not present (X) in a relevant document then it is to be combined with other similar documents so as to conclude that the searched claim is obvious.
Note: In addition to the report above, you may highlight relevant text in PDF files of prior art documents with your comments. This will be really helpful for a person reading the PDF version of relevant documents to relate them with the key features of a searched claim.
There is neither a fixed format for preparing a report nor a list of what to include in one. It depends on what your client, the reader of your report, or you yourself want. Accordingly, you can design your report to be as helpful as possible to the reader.
7. Example For Your Practice
For your practice, below is an invention that you can search for and find similar prior art documents. It’s simple enough to understand and therefore, should be easy to implement the learning you have gained in this guide.
Brief about invention: Smartphones are our extended organs nowadays. It is very common to listen to the songs or talk to someone over a phone with headphones on. It becomes dangerous when we use headphones while driving or walking on a busy road.
So, our invention provides a modified headphone or earphone that can pick up sounds louder than a predetermined decimal limit, such as vehicle horns. In response, the headphones or earphones can alert the user about an approaching vehicle. See the figure below representing the invention.
We advise you to approach this example step by step according to the instructions that you have read in this post. We wish you luck and congratulate you on learning a new skill.
Tip: We have taken the same example as above for testing whether Artificial intelligence (AI) tools like chatGPT, PQAI, and Google’s Bard can search for patent and non-patent literature. In this article, we found a few good prior arts which you can refer to.
7. Conclusion
In this search guide for prior arts, we discussed at length the whole procedure from the start to finish.
We started with the understanding of the technology features that were to be searched then we went into the collection of relevant material needed to form search strings. It required us to look for keywords, their synonyms, how to exclude redundant keywords, classifications, etc. Since classes are important to conclude the search exhaustively, we looked at ways to find better classes to incorporate into search strategy.
When we had gathered enough information, it was time to start forming search queries and looking for relevant documents.
To make the sense of found relevant documents, we went on to map the features of the searched claim to the disclosure of a prior art document.
Since there is no fixed search approach, we looked at narrow and broad searches. After that, we looked at how a relevant document can be interpreted as a novelty-destroyer or how we can combine multiple documents to prove the searched claim obvious.
Other approaches to the search were also explored such as assignee-based search, inventor-based search, jurisdiction and non-English search, citation analysis, etc.
Just like every project, in the end, we looked at how to prepare a great report.
To put the learning of this guide into practice, we also provided a practice example. We hope you sincerely go through the example and implement all the teachings of this guide.
We hope that you have understood the nuances of the search and will build upon this knowledge in future to excel.
We have a plethora of resources for learning patents. You must checkout them here.