Wisdom-Driven Knowledge Augmented Generation at Scale - Chin Keong Lam, Patho AI

Channel: aiDotEngineer

Published at: 2025-08-22

YouTube video id: 9AQOvT8LnMI

Source: https://www.youtube.com/watch?v=9AQOvT8LnMI

[Music]
So, hi. Hi everybody. Uh, my name is
Ching Kyong Lamb. Um, I'm the founder
and CEO of PO.AI.
Uh, a bit background about my company.
uh PTO AI started two years ago with a
invitation from National Science
Foundation from the SBIR grant funding
investigating LLM. We did a LMB driven
drug discovery application. Uh since
then we branch out to leverage what we
learned about building AI system for
large corporation. We are currently
building expert AI system for several
clients. Currently the system we build
goes beyond rack system. Um many of our
client is asking for AI system that
perform task like research and advisory
role based on their area of interest. Uh
today the talk is about sharing with our
fellow AI engineer what we learned so
far building this kind of system. Okay.
Uh what is knowledge? Okay. Generally
philosophically I say knowledge is the
understanding and awareness gained
through experience, education and the
comprehension of facts and principle.
And that lead to the next question is
what is knowledge graph? Right? So
knowledge graph is a systematic method
of preserving wisdom by connecting them
and creating a network or interconnect
relationship. That's important. The
graph represent the thought process
and comprehensive tonomy of a specific
domain of expertise. That's why this is
is very important for people moving
forward is about AI system then think a
lot and return uh advice instead of just
retrieve you know data from your
database right so that comes to the
development of this uh K a okay what is
K a kag stand for knowledge augment
generations and it's different from rack
okay it is enhanced language model by
integrating structure knowledge graph
for more accurate and insightful respond
making needs smarter more structural
approach than a simple rack. K a doesn't
just retrieve remember it understand
this is different
okay after in interviewing a lot of my
client okay so or we also expert in a
certain area of scale I found that there
are common ways of their thinking
decision making process the way that
make them expert in their area knowledge
graph seems to be a perfect fit so here
is the graph or state diagram if you're
a computer engineering grad like So um
it shows a wisdom the the wisdom note as
you can see is the is a core right it's
wisdom it just isn't static it actively
guide decision and fused by other
element
the output from the wisdom actually goes
to decision making in the blue right
wisdom isn't passive it guide decision
helping us choose wisely Okay. And then
the decision making analyze the
situation given in the circle in the uh
green and decision aren't make you know
in a vacuum. Okay. They analyze real
world situation. That's the difference.
Okay. So look at the wisdom input. Okay.
Look at the relationship feedback from
the knowledge to wisdom in gold color.
Example of that is knowledge to wisdom
like all your books smart and
encyclopedia Wikipedia whatever you
store plus once that data get absorbed
by whatever model you use up there it
need to regurgitate that and understand
that's why it's very important that
wisdom is able to synthesize the data
after you ingested knowledge okay that's
a kind of abstract but I I'll come come
to that later what I'm talking about
okay from Insight example of that is
wisdom derive pattern from chaos like
some of my client has a lot of social
media they their product how do they you
know track their product sediment from
from social media right so it's okay
chaotic and from ax tweet right so so
from that you can see some pattern of
their competitor versus uh current what
my product is that that's like an
example of that and I will go to that
later okay when all these connected
notes matter together why do they matter
matter all the notes relate to one
another to ever inre um enriching wisdom
storing system. Okay, this talk is about
storing wisdom, right? So knowledge
tells you what it is, right? And
experience tell you what worked before.
Insight invent what to try next, right?
Like a pizza knowledge is recipe.
Experience is knowing your oven burn
crust inside is like hey it is adding
you know honey to the crust you make
caramelize perfectly. Right? So the most
important part of the knowledge graph is
feedback loop. Okay, feedback isn't
oneway street. It learn from itself.
Look at the feedback from the uh going
back to all the note from insight to
wisdom. Okay. Um situation inform future
wisdom. Experience deepen it insight
sharpen it. Like a tree growing roots,
the more effect the stronger it get. Now
I want to ask you a question in general.
Where do you see this circle in your
life? Maybe a tough decision that you
know taught you something.
So one practical application for
leadership is wisdom. Avoid knee jack
reaction by learning from feedback. As
for personal growth, ever notice how
past mistake make you wiser? That's the
loop in the action. All this. So the
take away from the slide in this is
wisdom isn't a trophy you learn earn it
is a muscle you exercise the more you
feed knowledge experience inside the
more that guide you now I will show you
how it being mapped to my current client
you know all this is like very abstract
right so how I one of my clients
actually doing a competitive analysis uh
they used to have a marketing department
doing that but they want AI to do Yeah,
right. They they ask me to build the
system. This is exactly what I did with
the same taxonomy of storing all this.
So this taxonomy will be later on I talk
about how multi- aent is going to handle
all that. Here is one of the chatbot
that I build for my client to do you
know not just some uh we not just some
chatbot okay it's our wisdom graph power
AI designed to turn data into strategy
right dominant. So what kind of question
I talk about talk about how do I win my
competitor in this market space that's
kind of very sophisticated question
right so without uh if you do simply
just right by first speaker talk about
right right so it's not going to cut it
they're not going to able to answer that
kind of question okay what I did is this
uh we retain the same tonomy and uh the
wisdom is then mapped the same engine
here the wisdom engine wisdom engine is
like a orchestration agent that does a
lot of decision making including
advising what the ARM is able to see bas
on the current situation what to do next
right so um what I did is uh for the uh
decision making I map it to a strategy
generator so these customers are talking
about a competitive analysis right so um
I map the knowledge in term of knowledge
what do they have they have market data
right so I map this experience to HP is
one of a kind past campaign so they have
a lot of campaign doing a lot of
marketing and then um the insight is
actually mapped to uh in industrial
insight they have a database doing
storing that and then of course the most
important is is the the situation the
situation is how how am I doing how my
product selling right so so that that is
like a situation and then I map that to
a competitor weakness that means they
say if you make the aware of that you
probably get a very good answer and then
the chatbot will probably be doing the
right thing advising so from here very
high level you know state diagram or
there how do I map it to a system that
drive well here comes the trick so
anybody here heard of n
all right all right it's all good so so
I I first encounter similar situation
when my past IoT project which is not
red developed by uh IBM right so it's
the same kind of thing it's like no code
but but underneath the hood there's a
bunch of code okay it's all nodejs code
okay so uh but but for the for for for
proving your concept and all that it's
very very very flexible and I I highly
recommend that and and and here here you
can take a look at the the workflow the
work from I enable the implementation of
this complicated state diagram with um
what I say is there is a different
community note one of the very powerful
node is AI agent well previously N is
just a workflow automation tool I'm not
selling for N here I'm just telling you
I'm using it uh for pro prototyping
further down the road maybe the client
say too like I I really need to you know
go lightweight maybe we will switch over
to some other lang chain or whatever but
uh we actually uses like I mapped the
previous uh state diagram from the
wisdom engine I actually map that to our
uh wisdom agent okay wisdom agent is now
have the option to drive uh different
model like openai model entropic model
and even onrem model and then that the
key in making the state m the state
machine work is that my wisdom agent is
now overseeing like a supervisory agent
or all these other agent that do uh
whatever I say on the state diagram like
um for example the uh state of uh going
into a note of insight inside agent will
test to do go to the social media look
for all the settlement of all your
product and then collect that and then
pump that you can see that at the dot
bottom that we I connected to a
a centralized uh graph left the central
graph will be able to get updated by
different agent uh inside agent will
update the their perspective like part
of that graph for the uh as I say for
this particular uh inside note. So, so
all the unified knowledge graph will
contain the taxonomy that eventually
just think like the marketing strategies
the way that here they will probably if
you are doing manually they probably
would think in your shareepoint will all
this you know folder will store the same
kind of uh you know wisdom I call it to
make decision based on that. So the the
final decision is LM also depend on the
model that you use. Uh but I I I I
pretty much think that not really the
way that I think the final decision come
when you make a right decision from the
advisor output is basically depend on
all the tonomy the graph structure
that's very important. So come to that I
I want to go deep down how I implement
one of the node uh just to go a bit
technical on this competitive node. How
do I implement that? Okay, before I do
that, okay, competitive analysis, right?
Why why you can actually just use rag?
Why do you want to use a knowledge graph
like new forj? Well, if you ever being
asked that question, tell them these
five uh five reason. Okay, first reason
is knowledge graph you know u system
excel at capturing and representing
complex relationship between entities
that is covered by the first speaker but
I'll just reiterate that this lead to a
deeper contextual understanding which is
crucial for comparative analysis where
this in this case the insight can be
significant different okay you want to
find the gap in your computer winners
this is very important the second is
improve accuracy by leveraging
structured data and semantics
relationship knowledge graph can provide
more accurate and relevant information
compared to traditional vector racks. Um
this ensure generated content is not
only relevant but also precise and
reduce the noise and improve decision
making making this in this case the
board is supposed to help the guy that
is marketing department make decisions.
So, so you better make this work improve
accuracy. Any inaccurate data, you will
be out of the contract, out of the door,
right? So, very important. Okay, you're
talking about contract work like me, I
have to make the rack as accurate as
possible. So, the third is scalability
and flexibility. There graphic you know
graph are incredibly scalable and can
integrate to new data source and
relationship. The flexibility allow the
continuous improvement. As I say, if
your taxonomy is correct, you will
continues to improve and and reach,
right? So, so that is important and also
rich query capability. Knowledge craft
support complex query traverse to
multiple relationship entity provide
richer and more detailed insight. This
is particularly advantage for
competitive analysis when multifacet
query like like what the first speaker
say it is super a notoriously good in
answering things that normal rack will
fail. It's like multihop question. Okay,
this is very important. And then the
final one is the enhanced data
integration. Uh knowledge graph can
seamlessly integrate diverse data source
pictures, graphics, videos. Uh however
it is now that LM is so powerful we have
OCR capability can do that as long as
you have a right structure of the graph
semistructure unstructured the holistic
approach ensure compressive view of the
competitive landscape enable more
informed strategic decision making.
Okay. So one of the this is I'm going to
just very briefly go through this just a
uh example of that some of the thing
like um problem of a vectors rack you
know vector rack is really really bad in
answering limited numerical resing
vector store excel you know at sematic
sim similarity but struggle with complex
numerical calculation this is why uh for
marketing analysis uh that I'm building
the chatbot for uh they actually rely on
number instead of just you know
returning
example like this kind of if you ask
like what is the Apple uh revenue uh
between uh two uh you know what's the
revenue in 2022 they probably will give
you a bunch of this kind of a passage
right retrieve a graph instead of uh
this kind of a very very precise thing
like uh the answer is uh you know uh got
knowledge cross able because the the
data is already there in structure form
the data source assume a knowledge gra
name this particular in this particular
case Apple financial data the query will
be able the query engine will be able to
select the the revenue figure from 2021
to 2022 and and then do a function call
the function call will eventually give
come out with 15.23 23 which is exactly
what the marketing guy was looking for a
very quantitative stuff that most of the
decision were based on that because you
have the evidence not just some passage
that you retrieve from the data it's
basically evidence based decision making
is very important for this kind of uh
complicated rack system that you know so
um there's a jungle out there right now
you can use different kind of a uh uh uh
thing to build your uh you know uh this
is just a snapshot of that you know you
can actually use lang chain plus chroma
to to build your own rack and then you
also can combine that with your
knowledge graph depend on on on your
user case. Okay, if if this this slide
show that the rack and the k a can be bu
with money. Okay, I adopt that wisdom
graph in red color. Normally you will
see if client is just asking for a
simple rack that perform product
information query, you can just use a
simple chroma DB with LM agent. And if
you start to ask so complicated
questions like how can I beat my
competition based on my current market
share. Well, this will be able the the
the the thing that I will probably be
adopting is knowledge graph here with uh
graph DB plus cipher query and then qua
and also train my rack to perform
several loop of uh we call multihop
query and this probably will give you a
very good answer.
So uh and then it come to the another
question when I was trying to extract my
uh oh I think my time is uh is almost
up. Okay. So anyway this is like to say
uh the first speaker talk about the
extraction right there's a very simple
way to extract on the right side is like
automated totally automated LM graph
transformer on the left is like manual I
would probably re recommend to send the
hybrid model which is like after you use
the LM to extract your graph you ask the
interview the the expert that you're
going to build uh to to build your
taxonomy right to prone the graph we
call it proning your graph remove a lot
of relationship that then that that will
be okay and um I will try to just
highlight like this this is the result
of benchmark that we did okay anybody
ask you you know why you want to use
graph right or k a okay first is
accuracy I had achieved 91% because it's
really good in extract structure second
is flexibility 85% third is repubity
reproducibility deterministic and then
the fourth one traceability and finally
uh most important is scalability so in
conclusion
by leveraging structure nature of
western knowledge We can significantly
enhance the quantitative capability of K
a system and a more accurate and
insightful respond to complex query by
using wisdom-driven system as
highlighted. Together we can build
smarter AI system that can scale and
store wisdom with the right framing
potentially surpass the intelligent of
the initial expert that we meant to
serve. So uh talk to Jesus. You know
what do they just do? Talk to Jesus.
He's in our in in a not booth. this my
good friend and uh anybody that want to
build graph we have a good uh so-called
LLM graph rack stack on GitHub that is
sponsored by new forj and out of the box
just spin up your docker the next thing
you know your text is going to be
converted to your graph and you can
start happy pruning your graph thank you
thank you so
[Applause]
[Music]