You are on page 1of 108

Breaking

the Bias
Mastering the art of prior-art searches
Raima Ghosh, Shelza Gupta & Jaskaran Singh
Preface

If you are new to the world of intellectual property, prior-art searches are like solving
murder mysteries - you try to find proofs through clues. A textbook definition would
be uncovering evidences that prove an invention to be public knowledge, before a
particular date.

Research shows the most common process opted for prior-art searches is based a
thinking model. Side effect of following a thinking model is falling for biases. Most
people involved in searches fall for six major kinds of biases. This book introduces
six different ways to combat the biases that arise from a regular thinking model.

On a side note, you do not need to be an IP expert to crack the code. This book is for
you, if you aim to refine your thinking skills.
Note from the authors

At GreyB, a captivating challenge emerged - an opportunity to craft a book with three of


us working as a team. The idea was exciting and new, something we hadn't done
before, for none amongst us had any experience of book authorship before. Yet, united
by our shared passion for the written word, we dared to accept the challenge.

With each page, we experienced an extraordinary journey - not just through words, but
also in learning, friendship, and discovering more about ourselves. With lot of virtual
connects, re-visiting our journey of working in the Patent industry, we set out on a
writing adventure to bring our very first book to life.

As enthusiasts of prior art search, our collective narrative encapsulated not just tales
but essential lessons gained through our collective experiences. It was a testament to
our spirit, our willingness to explore uncharted territories, and our unyielding
commitment to defy limits.
It's easier to think outside the box if
you don't draw one around yourself.

Jason Kravitz
Introduction
According to a study in 2020, there are nearly 16
million active patents in the world.

From a regular toothbrush, to the structure and fabric


of pillows. From the design of pencils to the design
PATENTED of bombs. From the most obvious things to the most
unprecedented ones.

Probably everything is already patented?!


Is there a room for more patents?

Yes, innovation never stops, just like a circle has


infinite tangents. It's a world of making and
breaking. Inventions are born, registered,
implemented and they go obsolete. New inventions
come into the scene, which are improvements to
previous inventions like from toothbrush to electric
toothbrush. Or completely different inventions like
compact discs to cloud.

Hence, patenting inventions will always remain as


the superpower of inventors.

One might say we just contradicted ourselves?!


Yes
No

For an invention to be
patented, it has to be new.
How will one figure out if an invention is patentable?
The answer is search.
How will one figure out if a patent is marketable?
Search.
How will one figure out if a patent is misused?
Search.
Well, what search?
A prior-art search.
Yes
No

= $$$
The first instinct would be to fire up
www.google.com or www.chatgpt.com. That is fair
but patents are seldomly drafted simple. On an
average, it takes a minimum of three to four months
to get good hold on patent understanding.

Ask me anything… There would not have been million dollar litigations
against patents, had it been a button away.

We are not ruling out the possibility of finding a


prior-art with in five clicks. We are saying, if you are
dealing with invention searches, and running out of
coffee - this book might just save you.
The problem that we feel with most invention
search is researchers stay in a vicious loop.

Let’s say for example, you are trying very hard


to remember the name of a song.
How do you find it?
You Google the tune.
You ask a friend who jammed with you on the
same song.
You try to recall the visual and verbal cues to
remember it.

Without even knowing we might be following


a mental model.
The mental model
Some of us might not even know, but while we are in the process of finding a
solution to a problem, we follow a mental path. The selection of this path might
have plenty of reasons. This could be the first way we were taught, could be the way
that worked out well or could be a general approach. When the case in point is
finding prior-art, the linear innovation model is modus operandi.

Let’s break it down. Now, we’ll indulge into the six major phases of this thinking
model and their association with searching prior-art.
Write down the problem.
Think real hard.
Write down the solution.

The Feynman Algorithm


The Linear Innovation Model
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


Note: Looping and iterating back to any of the steps, would
still be considered as a linear approach.
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


Consider a ball is rolling down a platform.
How to stop the ball from rolling away from the platform?
It’s really easy to say “find the problem”, but is it really that easy?

No! When we start to innovate, we first try to figure out what will that
innovation solve.

Prior-art Context: When we have to invalidate a patent or check how strong a


patent is, we identify what’s the problem that the patent is solving - also known
as the novel aspect of the patent.
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


Let’s ask questions to gather information.
What’s the inclination of the platform?
What’s the material of the platform/ball?
What’s the mass of the ball?
In this phase, we gather information about various factors related to the problem, for
example we gather more information about the domain. This builds up a foundation
for the next phases and a concrete understanding.

Prior-art Context: Gathering more information about the domain by analyzing other
patents, research papers from the citations and also referring other databases.
Reading about the key inventors and key companies that contributed to the domain.
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


Is the ball magnetic? Is the surface of the platform
If yes, then we can use a magnet to made of dirt/mud or something
stop it from rolling. malleable? Then, we can dig it up
to make holes.
This phase involves brainstorming and exploring various creative and
innovative solutions to address the identified problem.

Prior-art Context: We generally find list of ways (A, B, C, etc.) that can
attack the novelty of a patent. Then comes the process of making
logics, search strings based on the collected keywords, classes, leads.
This also involves combining more than one angles. (A+B, B+C, etc.)
The more the number of attributes in the string, the narrower its fetch
is.
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


We try to check for the feasibility of various ideas.
For example, if i use a heavy obstacle, will it be enough to stop the ball?
Will a magnet be powerful enough?
Is the inclination too much for the ball to be stopped by the holes?
In this phase, the focus shifts towards refining and developing the ideas generated
in the previous step. It involves evaluating the feasibility, practicality, and potential
impact of each idea.

Prior-art Context: We analyze the collected data to assess its relevance to the
invention in question. This involves evaluating the patent documents and other
literature to determine if they predate the patent application and if they cover
similar technologies or concepts.
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


Testing idea 1: Using an obstacle to stop the ball
Testing idea 1: The obstacle might trip.
Testing idea 2: Using a magnet to stop the ball
Testing idea 2: The magnetic force might not be strong enough.
Testing idea 3: To make holes in the ground to stop the ball.
Testing idea 3: But the ball still might slip.
This phase involves putting the chosen solution into action, deploying resources,
and executing the plan. Implementation may involve testing, prototyping, and
iterating to ensure that the solution works as intended.

Prior-art Context: Trying and testing out the identified patent or non-patent
references, if they will work in actuality. We compare the discovered prior art with
the invention's claims to determine if any of the prior art anticipates or renders the
patent claims obvious.
Identifying the Problem

Gathering Information

Generating Ideas

Developing Solutions

Implementing the Solutions

Evaluating the Results


When we broaden our vision, we see that there was another inclined plane
mirrored just next to the first one. So the ball would eventually come to a
pause itself. So, there was no problem in the first place.
The final step in the linear innovation model is to evaluate the results of the
implemented solution. This involves analyzing the outcomes, measuring the
effectiveness of the solution, and assessing its impact on the identified problem.

Prior-art Context: We critically analyze the discovered prior art to determine its
relevance and significance to the patent in question. This helps to evaluate how
strong a patent actually is or could it be invalidated.
The Bias
Since biases limit our creative flow, we have used a box for demonstration.
Each side of this box, describes a bias while
following the mental model.
The bias associated with identifying the problem
Confirmation Bias

We tend to notice things that positively affirm our already held belief (for example,
reading 2000 patents will ensure finding a good prior-art. It worked in past 2-3
projects!), than we do the things that contradict our belief (what if the prior-art does
not exist in patents at all, maybe it is out there somewhere in library of a university!).
The bias associated with gathering information
Similarity Bias

Searchers might be more inclined towards prior art closely related to the
technology or field of the invention they're searching for. This could lead to
overlooking alternative approaches or less obvious prior-art. Can a prior-art for 5G
tech be found in WiFi?
The bias associated with generating ideas
sounds like an oxymoron
Outcome Bias

Outcome bias can lead to overlooking the potential blind spots in search, when
one relevant prior is found. But a better or more relevant prior art can still be
found.
The bias associated with developing solutions
Availability Bias

Less-known or unorthodox sources of prior art might not immediately come to mind,
leading to their exclusion from the search process. For instance, we are exposed to so
much new age tech, WhatsApp, Facebook, Instagram - it might not strike naturally to check
the MSN messenger, a 1999 invention that has been shut down!
The bias associated with implementing solutions
Overconfidence Bias

Overconfidence can lead searchers to favor their initial decided search approach,
even if it's not yielding in good prior art, and blinds the searcher to explore any
alternative strategy.
The bias associated with evaluating the results
Chronological Bias

Searchers have tendency to constantly notice when a reference was published, it


unintentionally leads to giving more weight to the timeline close to the cut-off date of
search. For example, typically, a search focused on WiFi should not have prior art before
1995, or is it?.
The Breaking
We have used handful old prior-art cases that we had cracked.
In retrospect, we broke the biases when it got harder to find art.
Case 01
Confirmation Bias

Procuring mobile phones from


China
In 2016, a major Chinese smartphone manufacturer approached us with a significant
problem. They were facing a patent infringement lawsuit from a competitor based in
China, and the potential consequences were severe – massive fines, costly licensing fees,
or even an injunction that could disrupt their operations.

The patent in question covered a seemingly basic technological concept: an on-screen


display showing call logs, and the cellular network associated with each call. Although
such patents covering general ideas can be valuable, they can also attract disputes from
other companies seeking to use similar technology.
Our client used this technology across all their phones, making the situation even more
critical. To address the issue, they sought to invalidate their competitor's patent by proving
that the patented idea already existed before the patent was granted.

In this case, our client needed to find prior art dating back to February 2006, a challenging
task given the rapid evolution of cellular technology.

During our search, we found some leads including user manuals of an old Chinese phone,
indicating that the required technology predated the competitor's patent. However, as we
know, the user manuals are not very detailed, hence it was not sufficient to invalidate the
patent.
Falling for the bias

We were about to give up and wrap up the project. Our deadline was approaching. There
was pressure to submit the results faster because the case was live in the court, and so
much was at stake.
It appeared that we have done everything in our capacity. Because an extensive search in
the published literature has already been performed. We had read enough literature to
ensure nothing was out there online. And that's how a routine search would have been
concluded.
Overcoming the bias

Unbiased third person review: the team closely working on the search, were somehow
convinced that our search is over. We had explored all the databases, tools we had been
using in our projects.
At that time, we had a discussion to review and brainstorm on the final search output.
This time with a wider team, including few senior members.Thats we changed
everything!
One of the teammates said, indeed we have explored all the literature available online.
What about what's present offline? Our closest result was the manual of a mobile phone.
How about checking the phone itself in person?
Overcoming the bias

This led us to explore Chinese cell phones on various online sites (like Chinese
counterpart for ebay), and on-the-ground investigation in Chinese stores dealing with old
phones.
We eventually got hands on a promising phone model that met our criteria through our
Chinese team mate who found it in an old market there. He then shipped it to us in India
for further examination.
The phone's government-issued production authorization number proved its existence
before the competitor's patent, securing vital evidence for their defense.

Had we succumbed to confirmation bias and stopped the search prematurely, our client
might have faced significant financial losses.
This case highlights the importance of negating confirmation bias in patent search and
decision-making. By staying open-minded, and welcoming different perspectives can
help identify biases and gaps in the search strategy.
Fearlessly explore new avenues to break this side of
the box.
Case 02
Similarity Bias

Prior-art in software
We were working on a prior art search for a patent related to Wi-Fi standards (IEEE 802.11)
and with a priority around January 2004. The patent was related to multiple IDs for access
points, allowing them to broadcast and support multiple network names. However, not all
of these new identifiers could be broadcasted, which created challenges in the scanning
process for new devices, leading to potential connectivity issues and delays.
The patent discussed the technology addressing this problem and presented a solution.

Our investigation revealed that around November 2003, the IEEE standard body made a
change in the Wi-Fi systems, somewhat related to what we were looking for. This intrigued
us to look further into the development of the IEEE standards.
Falling for the bias

As the patent had direct application in IEEE standards, and the inventors had drafted it in
a clever way so as it obstructs one to think beyond what's written on the face of the
patent.
The same happened here. Almost!

Following a familiar path, we were searching in the IEEE standards, and the related
publications, where we might expect to find the essence of the standard.

It blinded us from the valuable sources hidden just beyond the spotlight, i.e., potential for
valuable prior-art hidden within the folds of alternate technologies.
Overcoming the bias

Content Re-examination: Once we had hit the deadend with IEEE exploration, we thought
to take a pause and re-visit the patent we were given to search upon.
This time, instead of solely relying on the subject patent's technology , we adopted a
more comprehensive approach. We explored the reasons behind the patent's need and
eventually stumbled upon a timeline that led to discover a crucial piece of code. This
code, which was elusive through the usual search approach, proved to be instrumental in
our search process, highlighting the significance of avoiding the similarity bias.
Overcoming the bias

Further, we realized that similar problem might also be present in Linux. Then, we
thoroughly examined the open-source code of Linux, and found the actual prior art - an
update to the drivers for the new devices that resolved the issue generated by the IEEE
standard body’s changes.

This incident underscores the importance of understanding the "why" behind a patent to
ensure that potential references are thoroughly evaluated based on their content and
relevance rather than their surface-level similarities.
Look for new sources that hoard old content to break this side of
the box.
Case 03
Outcome Bias

The Dating Game


In the summers of 2021, when the pandemic had paralysed the world, we worked on a
crucial project. Ii was on customer complaint management, and the key feature was
related to the "Whisper" technology within a group chat window.

As we started the search, the first thing that hit the mind was to search for some other
social apps that have a chat system for multiple people. Systems like Zoom call were
something we had in mind that seemed similar.
Falling for the bias

While we followed the standard search procedure we did come across some decent
results like a Microsoft chat system, another chat system in a customer relationship
domain, however, the specific novel feature I was seeking was missing.

Since chat systems all seemed in the right direction, we had an outcome bias hardwired
in the brain. It led to the overlooking of potential blind spots in the search process, and
looking beyond the chat systems.
Overcoming the bias

Reverse inspection: now we were in tough spot. The chat systems were not very helpful
as prior-art. We again took a pause, and went back to our white board for some
brainstorming. This time we decided to follow a reverse approach.
Instead of thinking how a potential prior art should look like, we re-imagined what could
be a potential product overlap in today’s world other than the routine social chat apps.
With this, we were able to figure out that this feature might have existed in some old
game systems, given its prevalence in current-era games.
Overcoming the bias

We quickly shifted the focus to chat systems used in games. Slowly but surely, we began
to uncover more promising results. Eventually, leading to discover a YouTube video from
a user with mere 9 subscribers! Despite its humble origin, the video provided valuable
insights that became a game-changer.

This incidence reminded us that sometimes, unconventional approaches can yield


remarkable results that conventional methods may overlook. By staying open-minded
and pursuing alternative avenues, we was able to find the answers. So, the next time you
find yourself facing a challenging search, remember that a little hint of unorthodox
methodology might just be the key to success, and it certainly won't disappoint you.
Identify the odd points to break this side of the box.
Case 04
Availability Bias

PDA Teardown
In an European invalidation project centered around notification for chat on a handheld
device, similar to the pings we receive these days in major chat platforms on our mobile
phones.
We encountered a situation where a particular chat app appeared to be operating in a
similar direction to the technology in question. As we embarked on our patent search, we
were determined to ensure a comprehensive analysis free from biases.
Falling for the bias

This app that we shortlisted was very similar to what we required. It had almost all the
details. Except for a few small details. Since this app was extremely close match, we had
sifted through the entire literature around this app available online. We even managed to
find names of the developers of this app, and searched for all of their publications. But
were not able to find the missing details.

At this time, it looked like we had covered all the available literature. So the less-known
sources of prior art did not come immediately to mind, leading to their exclusion from the
search process.
Overcoming the bias

Collaboration and Peer Review: we then brainstormed with our team, and tried to look for
in what other ways the missing details can be verified. We figured out that the missing
detail was so minute of a teach feature that no one would really write or publish about it.
At that time, we thought what if we had the app installed in our phones, it would become
so easy to verify. And that's exactly what we did!
We chose to go the extra mile by purchasing two old PDAs with the intention of
mimicking a similar chat scenario and showcasing potential overlaps via video.
Unfortunately, we encountered a significant obstacle as the chat client servers (ICQ,
MSN) were no longer operational, rendering their installation on the PDAs impossible.
Overcoming the bias

Undeterred by the challenges, we aimed to build an artificial server for the chat clients
using available open-source codes. This would allow the PDAs to communicate with
each other effectively, providing valuable insights for our patent search.

By being aware of availability bias and implementing these strategies, we were able to
improve the thoroughness and reliability of our prior art search.
Look out for non-textual evidence to break this side of
the box.
Case 05
Overconfidence Bias

The Apple Smartwatch


Once we had to work on a prior art search against Apple’s smartwatch patent. It
mentioned how a fingerprint sensor was incorporated into the watch and how it was made
waterproof. So, we were planning to devise a solution against it.
Falling for the bias

While Initially, we had covered around 2000 patent literature, but we couldn’t identify
anything similar. So, at this point generally overconfidence bias kicks in. Leading the
searchers to favor the pre-defined search approach, even if it's not yielding desired
results, and discourage exploration of alternative strategies.
Overcoming the bias

Diverse Approaches: We didn’t wanna leave any page unturned, we went back to
analyzing the claim elements and came across some conclusions. The claim didn't limit
itself to a particular type of electronic device i.e. it can have a vast application area. The
claim neither talked about the characteristic features of the "Seal" like the material or so.
Nor the claim limited itself to a fingerprint based biometric sensor like the watch talked
about in the description of the patent.
So, we broadened the search to three different domains, first the smart watches, second
other electronic devices like mobile phones since they have waterproofing as well and
lastly as the claims did not limit to fingerprint sensor so we thought biometric sensors
could be a part of medical device and they might have waterproofing aspects.
Overcoming the bias

So what did this lead us to?

Search approach 1: No results


Search approach 2: Bang on result.
Search approach 3: Another Bang on result.

By actively addressing overconfidence bias, we can improve the quality and reliability of
our prior art searches, leading to more accurate assessments of the novelty of
inventions.
Diversify your understanding of concepts to break this side of
the box.
Case 06
Chronological Bias

Museum Trip
Let me share the fascinating story of how overcoming the chronology bias helped us
crack an important case.

This time around the task was to find prior art to invalidate a design patent for a
decorative chair.
The design was unique and intricate, featuring a ring-shaped structure in a circular
backrest, ornamental rings on its thin legs, and outward-slanting back legs. Conventional
methods, such as focusing on US design and Locarno classifications, yielded some
fruitful results (not too impressive) even after analyzing 36,000 entries.
Falling for the bias

By now, we had already analyzed around 36,000 entries. The deadline was fast
approaching, and we had only two days left to submit our findings.

Since the target chair was a decorative one, mostly used in wedding or other gatherings.
We looked at multiple places, even with the event planners, to get leads on finding such a
decorative chair. The design looked so recent and new era, we were not able to even
think to find it in ancient times.
Overcoming the bias

Cross-Check with Historical Context: As fate would have it, while commuting on a bus,
my teammate caught the title of a book his fellow passenger reading "Antiques Chop." A
spark of inspiration struck at that moment.
We realized that the design of the chair could exist in antique furniture, like they show in
historical documentaries and shows.
Overcoming the bias

As we delved into online museums and art/antique collections, the quality of the results
astonished us. Unbelievably, we found substantial prior art dating back as far as 200
years.

Reflecting on this experience, we couldn't help but wonder why this simple explanation
hadn't occurred earlier. By remaining open-minded and willing to explore unconventional
sources, we can reveal hidden treasures and break free from the confines of traditional
approaches. Research the historical context of technological advancements to
understand how ideas developed over time. This can help in assessing the true
significance of references.
Look for new sources that hoard old content to break this side of the
box.
CRACK!
The box is broken.
In retrospect, what emerged as a common thread in all the instances shared above was
our willingness to break free from the confines of established search methods. We
transitioned from the influence of different biases about how we perceive information,
and make decisions, to make more well-rounded and rational decisions.

Adapting to the demands of the moment became our guiding principle.

One of the best remedies that helped us in finding killer prior-art was keeping the
search dynamic. We found that relating a patent concept to tangible real-world systems
yielded more fruitful results than chasing elusive abstractions. This proved to be the
key to striking gold.

As we draw the final curtain on our exploration, the important lesson is loud and clear:
liberation from biases and the willingness to break the confines of a box hold the power
to unlock full potential.
The homework for the reader would be to relate these strategies to real life situations.
If that sounds like a difficult one, reach out to us, we might end up writing another book
on it. (wink)

You might also like