I/P Engine, Inc. v. AOL, Inc. et al

Filing 241

NOTICE by AOL Inc., Gannett Company, Inc., Google Inc., IAC Search & Media, Inc., Target Corporation re 238 Memorandum in Support, of Filing of Additional Exhibits to the Declaration of Howard Chen in Support of Defendants' Motion for Summary Judgment (Attachments: # 1 Exhibit 17, # 2 Exhibit 18, # 3 Exhibit 19, # 4 Exhibit 20, # 5 Exhibit 21, # 6 Exhibit 22, # 7 Exhibit 23, # 8 Exhibit 24, # 9 Exhibit 25, # 10 Exhibit 26, # 11 Exhibit 27, # 12 Exhibit 28, # 13 Exhibit 29, # 14 Exhibit 30, # 15 Exhibit 31, # 16 Exhibit 32, # 17 Exhibit 33)(Noona, Stephen)

EXHIBIT 32

Page 1
IN THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF VIRGINIA
Norfolk Division

I/P ENGINE, INC.,
Plaintiff,
V.
AOL, INC., et als.,
Defendants

)
)
)
)
)
)
)
)
)

CIVIL ACTION
2:11cv512

TRANSCRIPT OF PROCEEDINGS
Norfolk, Virginia
June 5, 2012

(MARKMAN HEARING)

Before:

THE HONORABLE RAYMOND A. JACKSON
United States District Judge

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 5
1

But, as I think the Court is probably very well

2

aware, we filed a motion on this, and, frankly, where we

3

started in this discussion with the defendants is they

4

started with a much, much larger number of terms.

5

THE COURT:

Well, you know where the Court

6

started.

7

fully expected the parties to come together and to come

8

up with just ten terms to be construed.

9

If you read my order, I said ten terms, and I

MR. SHERWOOD:

Yes, your Honor, I do

10

understand.

11

agreed to by the parties.

We proposed four, two of which have been

12

THE COURT:

13

MR. SHERWOOD:

14
15
16

So that leaves two?
So that leaves only two from our

point of view, yes, your Honor, that's right.
THE COURT:

So these other 16 or 17 I'm looking

at, are these all, you are saying, from the defendant?

17

MR. SHERWOOD:

18

THE COURT:

Yes, your Honor, that's correct.

Well, I'm sure that's not the case.

19

Not that you are wrong, but I'm sure that they have

20

narrowed that.

21

you are only asking the Court to construe two and the

22

others that are mostly in dispute are from the defendant,

23

I'm sure the defendant understands the Court means just

24

what it said, it will construe ten.

25

I will be waiting to hear that because if

MR. SHERWOOD:

That leaves eight.

Well, your Honor, if we look at

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 6
1

Exhibit C, the joint claim construction statement, which

2

is their list of terms -- this is their document, your

3

Honor, not ours.

4

THE COURT:

Hold up one second.

5

All right.

Go on.

6

MR. SHERWOOD:

If you look at this document, we

7

can see that -- if you just look at the item numbered 1,

8

there are two terms there.

9

together, but there are no words in common between those

They have grouped them

10

two terms, there are no record citations in common

11

between those two terms, and the Court is going to have

12

to do a different analysis with respect to each of them.

13

So my point is that if you start at the top of this list

14

and you work your way down, you will go well past ten.

15

As I said to the Court a minute ago, we only

16

proposed four, two of which have been agreed to.

17

Scanning network and combining are the two that we

18

propose are still outstanding for the Court to decide.

19

THE COURT:

20

MR. SHERWOOD:

21

THE COURT:

22

25

Combining.

Okay.

I think I expressed some

concern on the phone about combining.

23
24

Scanning network and what?

MR. SHERWOOD:

I understand you did, your Honor,

yes.
THE COURT:

Okay.

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 7
1

MR. SHERWOOD:

So my proposal is that we don't

2

do any more than the first ten that appear here, your

3

Honor.

4
5

THE COURT:

decide which ten it's going to be.

6
7

MR. SHERWOOD:

I appreciate that, your Honor.

Thank you.

8
9

Well, the Court is going to have to

THE COURT:

You are proposing only the first ten

and ignoring the numbers on the page?

10

MR. SHERWOOD:

Right.

In other words, these

11

have subparts, but we know subparts count as separate

12

issues.

13
14
15

THE COURT:

Well, that would mean we would go

probably no farther than No. 7 on this page.
MR. SHERWOOD:

Your Honor, actually if we went

16

to No. 7, we would still have 12, and let me point out to

17

you why that's the case.

18

If you look at No. 4, this has addressed two

19

different claims in two different patents.

20

separateness of systems in the '420 Claim 10 and then

21

another claim in another patent, claim in the '664.

22

it's actually two issues, not one, and the same with item

23

No. 6, your Honor.

24

the Court cannot construe Item 6 at all.

25

So it's the

So

So I would propose to the Court that

The term "user" has been agreed to between the

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 15
1

they filed for the patent that's one of the

2

patents-in-suit, Lycos.com was the seventh most visited

3

web site.

4

shopping spree purchasing more than two dozen web

5

companies on the Internet.

In 1998 to 2002, they sort of went on a

6

Then in 2000 because of its filter technology,

7

among other things, for producing search results, Lycos

8

was acquired by Terra Networks for $12 billion.

9

Okay.

Patents-in-suit, the first one is the

10

'420, entitled collaborative/adaptive search engine.

11

second one is the '664 patent.

12

filter system, and method for integrated content-based

13

and collaborative/adaptive feedback queries.

14

a mouthful.

15

The

It's entitled Information

It's quite

These are foundational search engine patents,

16

your Honor.

17

search results.

18

results?

19

At a high level, they involved improving
So what do I mean by improving search

In search engines that we use today, you usually

20

have a box where you put in your search request.

21

called a query.

22

are provided back with certain links.

23

search results.

It's

When you put in your search request, you
Those are the

24

What happens in between the search query and the

25

results is the methodology that produces for you the most

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 16
1

relevant results to your query.

2

engine is to come up with a methodology that will find

3

the best sites that are of interest to the person who put

4

in the user query.

5

The goal of the search

A lot of times, almost all the time, the only

6

information that you have about what the user is looking

7

for is the query itself.

8

use the word "grill" in a search engine because I'm

9

looking to buy a new barbecue, I will get links,

So, again, in my example if I

10

hopefully, that are related to what I am looking for,

11

barbecue grills, maybe grills on sale, maybe how to

12

grill.

13

The claims in the '420 and the '664 patent

14

relate to combining two specific measures in that

15

methodology that happens in the search engine, two

16

specific measures to improve search results.

17

specific measures are content and collaborative data.

18

Those

Here's sort of an abstraction, your Honor, to

19

try and illustrate the core essence of what Mr. Lang and

20

Mr. Kosak invented.

21

On the left you have content.

Generally

22

speaking, in a search engine environment this is how well

23

a piece of information matches the search query.

24

again, if I'm looking for grill, what I mean by content

25

data, is that the information that I am looking for, the

So,

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 17
1
2
3
4
5
6
7
8
9

web page, for example, is a content matching my query.
What do I mean?

Does it have the word "grill" in it?

Key words are a popular form of content data.
has the word "grill" in it 15 times?

What if it

When you are

determining how well something matches the query with
content-based data only, you can judge how well it
matches by the number of times the word you are looking
for appears.

If it appears once, maybe you have a low

relevance level for content.

If it appears 15 or 20

10

times, then you have a high level of content data.

11

you if have some type of threshold in deciding with the

12

server, you can see that in a content-based filtering

13
14

And

system only, you would provide the user with the one that
has 15 hits rather than the one that has one hit.

15

Okay.

16

collaborative feedback.

17

technique.

18

received from other users with similar interests or

19

needs.

20

On the other side here, we have
That's another filtering

Collaborative analysis evaluates feedback

What's that mean in the search engine world?

So

21

I'm about to search and look for the word "grill" and see

22

what I get back.

23

search for a grill.

24

have a similar need.

25

the web site results show up, in the past it was ten

Ten people before me might have done a
They have similar interests.

They

They are looking for a grill.

When

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 26
1

that's a 7, the rank rating is a 7.

Fig. 6 also shows

2

the collaborative data input at 415.

The same things

3

happens there, it moves over and you get a rating

4

predictor.

5

Let's say again for simplicity 5.

So now we know that we have a document that

6

content matches the 7, collaborative matches the 5.

7

does the patent say to do?

8
9

What

It says that these rating predictors are
combining for some folk combination function.

The

10

combination function is described in the specification as

11

anything from a simple, weighted, additive function to a

12

far more complex neural network function.

13

simple, just average.

14

We will stay

6 is the overall complete rating predictor.

15

you see, by combining these two pieces of data, you

16

change the value and the ranking of the document or

17

So,

article.

18

Let me back up a second and put this in sort of

19

practical terms.

20

producing the document to the user for providing it to

21

the user is 7.

22

this document will make it.

23

goes.

24

Content-wise it's like the example I gave earlier where

25

it's patio furniture, but I know people looking for

Here let's assume the threshold for

Under content analysis only, your Honor,
It's a 7.

A 7 or better, it

But maybe this document is not so good.

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 27
1

grills might purchase my patio furniture so I stick grill

2

all over the web site.

3

you see, it's low.

4

much.

So from the collaborative what

People don't really like this that

It's not one of the top choices.

5

So when you use Mr. Lang and Mr. Kosak's

6

invention it changes from 7 in a content-only world to a

7

predictor of 6.

8

This document will be excluded.

9

content-only world it would be served up, here it gets

10

We are now under the threshold of 7.
Where in the

excluded.

11

Let me flip that around and show you the other

12

way.

13

above.

14

seem to like this link.

15

feedback.

16

the scale so that the collaborative feedback help balance

17

out the low content.

18

high content and high collaborative, but this helps

19

people get the best results based on the information

20

that's out there.

21

invention came up with an improved way to filter search

22

results combining the content analysis and collaborative

23

feedback to provide superior results.

24
25

Let's assume here that the threshold is 6, a 6 or
Here the content is not so good but people really
It's got a lot of collaborative

When you do the combination, it pushes me up

I mean, obviously, you have to have

So Lang and Kosak through their

In this litigation, I/P Engine accuses each of
the defendants of creating and using infringing apparatus

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 28
1

and using infringing methods to present relevant

2

advertisements to users of the search engines by

3

combining content data and collaborative data, and this

4

is how some of the defendants generate, essentially, all

5

of their income.

6
7

Okay.

That concludes my tutorial on opening

statement, your Honor.

8

THE COURT:

9

MR. PERLSON:

Thank you very much.

10

THE COURT:

11

MR. PERLSON:

Good morning, your Honor.

Good morning.
I will try to -- plaintiff went

12

over some of the similar concepts that we were going to

13

go over and I will try to avoid repetition as much as I

14

can, although there probably will be at least a little

15

bit.

16

Here, just as an overview, we have the two

17

patents at issue filed in December of 1998.

18

would just note is that these patents share the same

19

specification and so I think you will probably hear from

20

both of us just referring to one of them, not both of

21

them.

One thing I

22

As plaintiff had indicated, the patents concern

23

a search engine system that combines collaborative-based

24

filtering with content-based filtering and the patent

25

does this with either a demand or a wire search.

And as

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 33
1

If I may have the Court's indulgence, I have a

2

board here that just contains the claim language itself

3

that I think might help to illustrate one of my points,

4

and that is that the term that is being construed here as

5

it resides in this claim is collaborative feedback data.

6

I think we can all agree that data is

7

information.

8

information, I think that that's a neutral dispute.

9

plaintiff made a point about that in their briefs.

Whether you want to call it data or
The
We

10

thought data information was a little more user friendly

11

term, but I don't think there's a big deal there.

12

what the claim language says is that that data comes from

13

somewhere.

But

It comes from system users.

14

Now, when you look back at my slide, what you

15

can see is that they are writing additional limitation

16

into this claim.

17

users with similar interests or needs.

18

two source limitations here, your Honor.

19

from users with similar needs and interests, according to

20

the defendant; and then according to the claim language

21

that comes after the term that's being construed, it

22

would be from system users.

23

They are saying that it comes from
So, we would have
It would be

Now, I would submit to the Court that it's

24

either nonsensical, nobody would write that way, or it's

25

leaving certain language out of the claim because we

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 34
1
2

would only need to have a one-source limitation, not two.
IPE's construction, on the other hand, your

3

Honor, does not propose a second source limitation.

4

instead, what it does is it proposes to explain

5

collaborative feedback data is the information concerning

6

what informons users with similar interests or needs

7

found to be relevant.

8

just talking about data or information.

9

talking about where it's coming from.

But

So the point here is we are still
We are not

This fits

10

harmoniously and appropriately within the claim language

11

itself, your Honor.

12

I should point out, we have the same -- it's

13

exactly the same with respect to Claim 25, also

14

collaborative feedback data as the term to be construed

15

with a separate source limitation that resides outside

16

the claim term, your Honor.

17

to the '420 as well.

18

And this term only applies

It's not applicable to the '664.

So, your Honor, interestingly, both parties

19

point to the same language in the specification to

20

support their constructions, and what I would point out

21

to the Court with respect to this specification language

22

which appears here at the bottom of the slide is that it

23

is referring to the same thing that I'm talking about

24

here, which is the informons that the other users with

25

similar interests or needs have found to be relevant.

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 35
1

Now, this is a demand search environment, as I

2

think the Court just heard, and I think maybe I should

3

explain the importance of that here just briefly.

4

The patents actually have two different systems

5

that are in them, and we are only asserting the demand

6

search claims.

7

We are not asserting wire search claims.

So when we want to know what somebody else

8

thinks is relevant, we don't have a profile, as

9

Mr. Perlson alluded to in his presentation, where we

10

might know something about the user.

11

one-time searches that somebody just goes to the web and

12

they randomly decide they want to search for something.

13

Nobody knows anything about them.

14

query that they entered.

15

look to see who has similar needs or interests, what we

16

are looking at is who else made that same search?

17

else made the same query?

18

asked for Jaguar?

19

they find relevant to their query?

20

collaborative feedback data that is described in these

21

two claims here, your Honor.

22

These are all

All they know is the

So the point is that when we

Who asked for grills?

And what did they click on?

Who
Who

What did

That is the

As I say, there's no source limitation with

23

respect to the information.

24

to be making clicks, doing queries and clicking on things

25

that are results of their queries, and that is going to

Every system user is going

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 36
1

be the source of the data, but it's the source of the

2

data because that's what the claim says, it comes from

3

the system users.

4

So, your Honor, this next slide we have tried to

5

illustrate what the difference is between the two

6

parties' constructions.

7

parties rely upon appears in the left under the blue

8

heading, and the key part we have put in brackets at

9

capital [A], the language, that's really what's being

The specification language both

10

construed.

11

interests or needs found to be relevant."

12

"What informons other users with similar

And you will see the plaintiff's proposal tracks

13

that language very closely.

14

of data.

15

Court feels that data is a better term to use, we have no

16

problem with that.

17

We used information instead

As I said, we are agnostic about that.

If the

What we propose is that this term be construed

18

to mean information concerning what informons other users

19

with similar interests or needs found to be relevant.

20

Now, the defendants' proposal imports some of

21

that into their claim construction, but as I have already

22

pointed out by highlighting this claim language, they put

23

in additional source limitation, your Honor, which is not

24

appropriate.

25

said, nonsensical or superfluous.

It either renders the claim language, as I

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 37
1

Your Honor, I have a couple of things I wanted

2

to point out from the plaintiff's slides which I just saw

3

this morning, so if the Court would bear with me for one

4

second here.

5

The defendants make the argument, and you will

6

hear this when they get up to present their materials,

7

repeatedly that IPE's construction does not include the

8

collaborative element.

9

the Court is the collaborative element is that which you

What I just want to point out to

10

collect from the other system users who made the same

11

query as to what they found relevant to that query, and

12

our claim construction contains all of that without

13

muddying up the claim language with additional source

14

limitation.

15

If I may now, your Honor, I would like to turn

16

to the next claim term, unless the Court has any

17

questions?

18
19

THE COURT:

No, that's fine.

The Court

understands it.

20

MR. SHERWOOD:

21

And we see a similar kind of issue here with

Thank you.

22

respect to the '664 and the two claims that are at issue

23

here.

24

I have learned over doing these cases, like to express

25

the same concepts in different ways.

The language is different because patent lawyers,

We could probably

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 38
1

cut down the number of patents we have in our system if

2

they didn't do that, but that's what they do.

3

And so here what we are talking about, again, is

4

a feedback system for receiving information found to be

5

relevant to the query by other users.

6

with respect to this term, actually, your Honor, is that

7

we don't think the Court needs to construe it.

8

has already declined to construe the term "relevant."

9

have an agreement with respect to the meaning of the term

Our first position

The Court
We

10

"relevance," which I suspect is going to inform the

11

parties with respect to the meaning of the term

12

"relevant."

13

term "users," and we have an agreement with respect to

14

the term "query."

15

whole lot left for the jury to have to figure out, just

16

some plain words that reside in between those words that

17

we already have an agreement for.

18

And we have an agreement with respect to the

So it seems to me there's not really a

But, in any event, if we look at the

19

defendant -- actually, your Honor, let me point one other

20

thing out to you.

21

at the bottom of this slide is from Fig. 9 of the patent

22

and it illustrates the system of receiving what I'm going

23

to call feedback or collaborative data.

24

see is that down at the bottom left there's a box that

25

says other user and it shows an arrow going up to the

This patent drawing that we have down

And what you can

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 39
1

feedback processor that Mr. Cimino talked about earlier.

2

All that this claim term here is describing is

3

the receipt into that processor of the informons that the

4

other users found to be relevant to the query.

5

nothing more than that, your Honor.

6

pointed out, there's a separate source limitation here,

7

too, just as there is in the '420 patent for this

8

information.

9

limitations here, any more than there is in the '420

10
11

It's

And as I have

There's no need to have two source

patent.
The defense wants to layer on here by saying

12

that the information can only come from certain users.

13

It can only come from other users, which is what we see

14

in the patent, but they want to add that it can also only

15

come from users with similar interests or needs.

16

already know it's coming from users with similar

17

interests or needs because, in fact, they are the ones

18

who clicked on the search results.

19

analyzing, and the patent and the claim construction that

20

the plaintiff has proposed are very clear with respect to

21

that, I think, your Honor.

22

But we

That's what we are

There's some additional issues with respect to

23

the defendants' construction.

24

"determining," as the Court can see.

25

is receiving.

They put in the word
The claim language

Receiving and determining are not

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad

Page 40
1

synonymous, your Honor.

2

patent that they are the same, and I think we can tell,

3

again, from just plain English usage that they are, in

4

fact, very different things.

5

There's no suggestion in the

In addition, as I have already alluded to, they

6

would equate other users with users with similar

7

interests or needs.

8

would suggest to the Court, in fact, that it's redundant

9

because we know that these are users with similar

Those two are not the same, and I

10

interests or needs because of the fact that they have

11

clicked -- they have entered similar queries and they

12

have clicked on informons that the system is going to

13

determine are relevant to the query.

14

It's a noninfringement position which defendants

15

are pretty honest about, which is they are saying in our

16

system we don't keep track of information about the

17

users.

18

or their needs are similar.

19

Those are other claims in the patents which we are not

20

asserting.

21

We don't know whether their interests are similar
That's the profile system.

The only way to know whether people have similar

22

interests or needs, just as is explained in the patent,

23

is to look and see what they click on, and that's exactly

24

what this claim construction would entail.

25

Your Honor, I would reserve my remarks on that

SHARON B. BORDEN, OFFICIAL COURT REPORTER

de5c7d49-b7af-4561-b6e2-4c87d99142ad