[00:01:13] Speaker 05: Okay. [00:01:14] Speaker 05: The next order of case is number 191178, Phillips against Microsoft. [00:01:19] Speaker 05: Mr. Oliver. [00:01:38] Speaker 00: Good morning. [00:01:38] Speaker 00: May it please the court. [00:01:40] Speaker 00: This case involves a means plus function limitation in the form of decoding means, which cause a signal to be provided in substantially unmodified form when a particular or certain identification signal is provided. [00:01:54] Speaker 00: There was no dispute below as to the corresponding structure in the 695 patent for decoding means. [00:02:00] Speaker 00: Microsoft identified the patent's decoding unit 76, which is shown in figure two of the 695 patent on appendix 35. [00:02:09] Speaker 00: That broader structure, decoding unit 76, included a number of structures therein, as defined in the patent. [00:02:16] Speaker 00: It included an input. [00:02:18] Speaker 00: It included a switch. [00:02:20] Speaker 00: It included an output 80, where the pathways came back together. [00:02:24] Speaker 00: And it included a specific decoder 77. [00:02:27] Speaker 00: So a specific decoder within a larger structure. [00:02:31] Speaker 00: And that specific decoder 77 may be an entropy decoder. [00:02:36] Speaker 00: With that construction, the corresponding structure in the prior arc cannot be limited to only an entry decoder, where the claim construction for decoding means requires more. [00:02:46] Speaker 00: But the board lost sight of the claim construction, instead getting mired in trying to figure out what Microsoft intended it to rely upon. [00:02:55] Speaker 00: Specifically, Microsoft originally relied upon the decoding unit shown in Figure 7 of Venbrooks. [00:03:01] Speaker 00: However, once it was established that that decoding unit in Venbrooks could not provide a substantially unmodified signal, in its reply brief, Microsoft said, no, no, we don't want you to look at the entirety of the decoding unit, but only a specific part, only an entropy decoder. [00:03:18] Speaker 00: And the board agreed, having lost sight of the claim construction. [00:03:21] Speaker 00: That resulted in two reversible errors. [00:03:24] Speaker 00: First, the board improperly allowed Microsoft to change its position from its petition reply. [00:03:29] Speaker 00: And second, and more critically, by relying on only an entry decoder in the prior art, where the claim constructions for decoding means requires more, there was simply a failure to provide what is required of means plus function analysis, a structure to structure comparison. [00:03:48] Speaker 00: There was simply no structure-to-structure comparison by the board, and had they performed that, they would have seen that the structure ultimately relied upon the final decision did not match. [00:03:59] Speaker 00: Now, this can be seen from the figures in the patent and Ben Brooks. [00:04:04] Speaker 00: Specifically, figure two of the patent on the appendix at 35 has this broader structure 76, and it is described in the patent as requiring or having a number of elements. [00:04:15] Speaker 00: It has the input 74. [00:04:17] Speaker 00: It has a switching unit 79, which brings these different pathways back together and brings them back together to an ultimate output 80. [00:04:25] Speaker 00: Along the lines, there's also an entry decoder 77, what may be an entry decoder, but all of those structures make up the decoding means as identified by Microsoft and the claim construction and as agreed to by Phillips. [00:04:39] Speaker 00: However, what happened is that because it was determined that Venbrooks could not provide a substantially unmodified signal, [00:04:47] Speaker 00: Going from the broad decoding unit, they went down to only an entry decoder. [00:04:52] Speaker 00: But if the claim construction requires more than an entry decoder, relying on only an entry decoder in the prior art simply does not work. [00:05:01] Speaker 00: The structures no longer match. [00:05:03] Speaker 00: Specifically, we have to look at the appendix at 1787. [00:05:09] Speaker 00: What Microsoft did in its reply brief, not in its original petition, it provided a marked up version [00:05:16] Speaker 00: of Figure 7 of Ben Brooks' decoder. [00:05:19] Speaker 00: And it said, okay, now we're going to divide this decoder up into different sections. [00:05:25] Speaker 00: And as we can see there, they boxed around part of it and said, this is the input section. [00:05:29] Speaker 00: And they boxed another part and said, this is the decode section. [00:05:32] Speaker 00: And they boxed another part and said, this is the output section. [00:05:35] Speaker 00: And they said, now, board, we want you to look at only the decode section, which Microsoft admitted in the appendix at 1490 is an entropy decoder. [00:05:44] Speaker 00: So they said, don't look at the input, don't look at the output, only look at the entry decoder. [00:05:50] Speaker 02: You're talking about the unmapper here? [00:05:53] Speaker 00: The unmapper is what they said is part of the output section. [00:05:56] Speaker 00: And the unmapper is part of the overall decoding unit. [00:06:00] Speaker 00: And it is where the pathways come together and where the ultimate decoded output is provided. [00:06:06] Speaker 00: But the claim construction for the 695 patent includes the input section. [00:06:11] Speaker 00: includes the output section and includes the entropy decoder 77. [00:06:16] Speaker 00: All of those structures are included in the claim construction for a means plus function term. [00:06:23] Speaker 04: But in Venbrooks, isn't the default portion of the entropy decoder the unmodified portion? [00:06:34] Speaker 00: It is only coming out of the entropy decoder. [00:06:37] Speaker 00: If you were to look at Venbrooks, what you see in that decode section from the default, [00:06:42] Speaker 00: The pathways only come back together in the unmapper, which is the far right. [00:06:47] Speaker 00: That unmapper always modifies the signal. [00:06:50] Speaker 00: There's not always one type of decoding that takes place in a decoder, as acknowledged by some of the evidence provided by Microsoft, including the Hoffman reference they relied upon. [00:07:01] Speaker 00: Decoders can have entry decoders, plus tables for the entry decoders, plus unmappers, plus other elements. [00:07:08] Speaker 00: The question is, what is the corresponding structure for decoding these? [00:07:12] Speaker 00: And if we look at the patent, it's not the exit of 77. [00:07:15] Speaker 00: 77 is the entropy decoder in the patent. [00:07:18] Speaker 00: It is not the unmodified signal is not what coming out of the decoder, but what ultimately comes out of output 80, where those pathways ultimately come back together. [00:07:29] Speaker 00: If we were actually to compare things, the unmapper looks more like the switch 79. [00:07:32] Speaker 00: It's where those pathways come back together. [00:07:35] Speaker 00: What they tried to do is say, OK, we're only looking at the entropy decoder. [00:07:39] Speaker 00: And your honor is correct that if you only look at entropy decoder, there's a part within that entropy decoder where there's a lack of a modification. [00:07:46] Speaker 00: But if the claim construction requires more than an entropy decoder, which it clearly does because 77 is the entropy decoder of the patent, then you can't look at only the output of the entropy decoder event books. [00:08:00] Speaker 00: There's a structural requirement here. [00:08:02] Speaker 00: Are you looking at the structures? [00:08:04] Speaker 00: And not only did the board not do that properly, it didn't do that at all. [00:08:08] Speaker 00: It simply said, we think Microsoft only wants us to look at the entropy decoder of Venbrooks. [00:08:13] Speaker 00: But this is a means plus function limitation. [00:08:16] Speaker 00: And specifically, the law requires from this court, including in the IPCOM case cited in page 40 of our blue brief, and I quote, the PTO may not disregard the structure disclosed in the specification corresponding to such means plus function language [00:08:31] Speaker 00: when rendering a patentability determination. [00:08:34] Speaker 00: The board said only this. [00:08:36] Speaker 00: We think Microsoft wants us to look over here. [00:08:39] Speaker 00: But they never did what is required in the next step to say, OK, if you only want us to look at the entry decoder, now let's look at the entry decoder. [00:08:46] Speaker 00: And does that structurally match what is required of the claim construction? [00:08:51] Speaker 00: And it simply cannot. [00:08:52] Speaker 02: And the board does a does an unmapper actually decode any singles at all? [00:08:59] Speaker 00: Yes, your honor mapper. [00:09:01] Speaker 00: The unmapper is a decoder and that's admitted in the record. [00:09:06] Speaker 00: Specifically, Microsoft's expert Dr. Castleman acknowledge that even under Microsoft technology dictionaries, where is that? [00:09:14] Speaker 00: I'll give you a few sites. [00:09:16] Speaker 00: First, in the appendix at 1710 through 1713 and 1732 through 1733, it was acknowledged that the unmapping is a type of decoding that would fall under the definition of decoding offered by Microsoft. [00:09:32] Speaker 00: He also admitted in the appendix at 231 that the decoded data are available at the output of the unmapper section. [00:09:39] Speaker 00: and that the unmapper is part of the, and I quote, decode pipeline. [00:09:43] Speaker 00: That's in the appendix at 1740 through 1741. [00:09:47] Speaker 00: What was done in the record was to show Microsoft's expert a technical dictionary from Microsoft that defined decoding. [00:09:55] Speaker 00: And he agreed that an unmapper falls under the classification of decoding. [00:10:00] Speaker 00: So what we have here is a large decoder that has a couple different types of decoding. [00:10:05] Speaker 00: It has an entropy decoder. [00:10:07] Speaker 00: It has an unmapper, which is another type of decoder. [00:10:10] Speaker 00: What the patent requires is that the overall decoding unit, not just an entropy decoder, but the overall decoding unit, operate in a certain way. [00:10:19] Speaker 00: And by limiting the analysis to only an entropy decoder, excluding the mapper, they're not doing a one-to-one comparison. [00:10:25] Speaker 00: There's not an apples-to-apples comparison. [00:10:27] Speaker 00: You're saying we're going to look at more of the entropy decoder as the claim construction, but only an entropy decoder in the prior arc. [00:10:34] Speaker 02: Where again in the record do we have that the unmapper decodes singles? [00:10:42] Speaker 00: Well, unmapping, Dr. Kasselman has shown a Microsoft technical dictionary that defined decoding. [00:10:47] Speaker 00: And he agreed that the [00:10:49] Speaker 00: definition of decoding in the Microsoft dictionary would cover unmapping. [00:10:54] Speaker 00: And that's in the appendix at 1710 through 1713 and 1732 through 1733. [00:11:00] Speaker 00: It's also stated in Ven Brooks itself that the unmapper is part of the decode pipeline, such that the final step of providing decoded data is only available after the unmapper. [00:11:15] Speaker 00: And that was admitted to [00:11:16] Speaker 00: by Dr. Kaussman in 1728 and 1729. [00:11:20] Speaker 00: So unmapping is one type of decoding in that it's putting the signals back together. [00:11:25] Speaker 00: And the problem that Microsoft ran into is that they originally pointed to this block diagram 7 and said, see, Ben Brooks has the decoding means of the invention. [00:11:35] Speaker 00: And what was established during the case is that the unmapper always substantially modifies the signal, such that there's never an output of an unmodified signal. [00:11:45] Speaker 00: That is when Microsoft in its reply said, don't look at the entirety of the decoding unit in Ben Brooks. [00:11:51] Speaker 00: Look only at the entry decoder of Ben Brooks. [00:11:54] Speaker 00: But that causes a problem where it simply no longer strictly matches. [00:11:58] Speaker 00: And there's no analysis whatsoever from the board as to a structure to structure comparison. [00:12:04] Speaker 00: Only the idea of Microsoft wants us to shade our eyes and not see anything but the entry decoder. [00:12:11] Speaker 00: But a proper analysis by the board has to say, does that entry decoder match the claim construction for decoding? [00:12:19] Speaker 00: And it simply doesn't. [00:12:21] Speaker 00: The entry decoder in the patent is 77. [00:12:25] Speaker 00: What is required is a substantially unmodified signal not exiting 77, but exiting the ultimate output 80, where these signal pathways come back together. [00:12:36] Speaker 00: Then Brooks doesn't have that. [00:12:38] Speaker 00: By the time the signal pathways come back together and there's an exit of the signal from the overall decoding unit, it is always substantially modified. [00:12:48] Speaker 00: And in fact, it was admitted by Castleman himself that the unmapper always is a substantial modification to the actual output. [00:13:11] Speaker 00: which is stated in his deposition testimony at 1787. [00:13:14] Speaker 00: Specifically, question, so is it your opinion that the mapping function in Ben Brooks would be a substantial or insubstantial modification as you interpret the term here? [00:13:24] Speaker 00: Answer, it would be a substantial modification. [00:13:28] Speaker 00: There was no dispute that the overall decoding unit in Figure 7 provided a substantial modification. [00:13:35] Speaker 00: Therefore, it simply no longer matches the claim construct. [00:13:39] Speaker 00: We have one of two problems. [00:13:41] Speaker 00: Either you look at the entire decoding unit of figure seven of the patent, of the Fenbrooks, in which case you have a substantially modified signal. [00:13:48] Speaker 00: If you limit yourself only to the entropy decoder, you no longer have a structure that matches what is claimed. [00:13:53] Speaker 00: Because the entropy decoder cannot be limited alone to what is substantially modified because the patent requires more. [00:14:01] Speaker 00: It requires the entropy decoder 77, the input, the output, the switch, the signal pathways, all of it taken together. [00:14:09] Speaker 00: And for that reason, we believe that the board erred below. [00:14:14] Speaker 00: We also believe that the board erred below by allowing Microsoft to change its position in this regard. [00:14:20] Speaker 00: In its original petition, they simply said, Microsoft simply said, figure seven, excerpted below, shows that the decoder and then Brooks uses a control signal. [00:14:32] Speaker 00: That's in the appendix of 114. [00:14:34] Speaker 00: Just to be clear, figure seven, excerpted below, shows. [00:14:38] Speaker 00: Thus, Figure 7 shows that the decoder in Ben Brooks uses a control signal. [00:14:43] Speaker 00: They never limited themselves to only a portion of Figure 7. [00:14:48] Speaker 00: And, in fact, their declarant, Dr. Castleman, not only showed the entirety of the figure, but quoted sections of Ben Brooks and bolded the language referring to the unmapper, indicating the importance of the unmapper in the overall structure. [00:15:04] Speaker 00: Therefore, what they originally relied upon was all of Figure 7, the decoding unit. [00:15:10] Speaker 00: And only once they realized their mistake, that the coding unit in Ben Brooks cannot provide the signal, did they change their position to say, oh, stop looking at the input, stop looking at the output, [00:15:22] Speaker 00: despite the claim construction requiring the initial input and the initial output, and look only at the end of the code. [00:15:28] Speaker 00: That was a change of position from the original petition to the reply. [00:15:32] Speaker 00: And that is simply improper with respect to changing of the arguments along the case. [00:15:39] Speaker 00: So that is another reason why the board's decision below should be reversed. [00:15:43] Speaker 05: Okay. [00:15:43] Speaker 05: Do you want to save the rest of your rebuttal time? [00:15:45] Speaker 00: I would do. [00:15:46] Speaker 00: Thank you, Your Honor. [00:15:47] Speaker 05: Thank you, Mr. O'Hare. [00:15:52] Speaker 01: Good morning, your honors, and may it please the court. [00:15:59] Speaker 01: Substantial evidence supported the board's finding that the Venbrex prior art reference taught the decoding means limitation of the challenge 695 patent claims, which is the only aspect of the board's decision that Phillips challenges on appeal. [00:16:11] Speaker 01: Phillips makes two arguments contesting this finding. [00:16:14] Speaker 01: The first is that it argues that the petition relied on the unmapper component as part of the decoding means structure. [00:16:21] Speaker 01: That is the only argument that Phillips presented before the board, disputing the Venbrex priori grounds. [00:16:28] Speaker 01: The record provides substantial evidence that supports the board finding that the petition did not rely on the unmapper component. [00:16:35] Speaker 01: The board recognized that the petition began its analysis of the decoding means limitation by specifically excerpting [00:16:42] Speaker 01: the relevant portions of Figure 7, which shows the entire decoder chip that has many components that are unrelated to the process of decoding or passing a signal portion through unmodified. [00:16:52] Speaker 01: The entire decoder chip includes a FIFO input block. [00:16:55] Speaker 01: It includes a serial to parallel converter block. [00:16:58] Speaker 01: It includes the portion that the petition excerpted as being relevant for purposes of the decoding means limitation. [00:17:04] Speaker 01: That excerpt was at Appendix 115. [00:17:06] Speaker 01: And it includes this final post-processing unmapper block. [00:17:10] Speaker 01: So the board recognized that the petition had specifically excerpted the relevant components of Figure 7, which were an ID decode unit and default and decoding pipelines. [00:17:20] Speaker 01: The board also recognized that in addition to excerpting the relevant components, the petition mapped these components to the claimed function and explained that these portions, these excerpted components of Venbrex, either passed a portion of the signal through unmodified, [00:17:35] Speaker 01: by routing the signal through the default path or decoded that signal portion by routing it through the K and FS split pipelines. [00:17:44] Speaker 01: And so the board recognized that the petition specifically identified structure in Venbrex and described how that maps to the claimed function. [00:17:52] Speaker 01: Nowhere in the petition does the word unmapper appear. [00:17:56] Speaker 01: So substantial evidence supports the board's finding that the petition did not rely on the unmapper component as being part of the decoding mean structure. [00:18:03] Speaker 01: The petition instead relied squarely on the ID decode unit and the default and decoding pipelines within Venbrooks. [00:18:11] Speaker 01: The second argument that Phillips makes that challenges the board's point. [00:18:14] Speaker 02: I asked counsel whether the unmapper actually decodes any singles, and he said that, yes, it did. [00:18:19] Speaker 02: Do you agree with that? [00:18:20] Speaker 01: We do not, Your Honor. [00:18:21] Speaker 01: So there is evidence and testimony from Dr. Castleman at appendix 1733 that clarifies this. [00:18:30] Speaker 01: that the confusion that Phillips is playing on is in the term decoder and in the term in decoding in the abstract in general and substantially unmodified in the abstract in general. [00:18:42] Speaker 01: And those words as they are used and as the board has construed them in these claims. [00:18:46] Speaker 01: And what Dr. Castleman stated at page appendix 1733 is that a person of skill in the art reading this reference, the Venbrex reference, would understand that the mapping and unmapping functions are different. [00:18:59] Speaker 01: from the encoding and decoding functions. [00:19:02] Speaker 01: Now, the board construed the word substantially unmodified to mean not decoded. [00:19:07] Speaker 01: Decoding in the abstract can mean any number of things. [00:19:10] Speaker 01: In the compression context, and specifically in the context of these claims, which deal with compression of data that is then transmitted across a channel, decoding has a very specific meaning. [00:19:21] Speaker 01: And that is different from mapping and unmapping, which are pre- and post-processing operations. [00:19:26] Speaker 01: They are designed to make data more adaptable to being encoded. [00:19:30] Speaker 01: They are different from encoding and decoding. [00:19:36] Speaker 01: Philip's second challenge to the board's decision is that the board allegedly failed to apply the proper test in finding that Vin Brooks taught the decoding means. [00:19:45] Speaker 01: This is also incorrect. [00:19:46] Speaker 01: The board recognized that there was an uncontested construction. [00:19:50] Speaker 01: After recognizing that there was an uncontested construction, [00:19:53] Speaker 01: the board pointed to the specific structure in Venbrex and the prior art that the petition had relied on as meeting the decoding means limitation. [00:20:02] Speaker 01: Now, just like the decoding unit 76, the agreed upon corresponding structure, these Venbrex components included a decoder element, the K and FS decoding path, and they included a second pass-through element, a separate element, that default path. [00:20:19] Speaker 01: And the board, having identified the specific structure in Venbrex that was at issue, [00:20:23] Speaker 01: proceeded to describe how these components performed the claimed function. [00:20:27] Speaker 01: Specifically, based on a control signal output from the ID decode unit, portions of a signal were routed through either the default path or through the K and FS decoding paths. [00:20:37] Speaker 01: And the board cited to the Venbrex reference itself, as well as the Castleman declaration in the original petition, the original supporting declaration, to support its findings. [00:20:47] Speaker 01: So we submit that the board did provide a sufficient analysis of the decoding means limitation, and that substantial evidence in the record exists to support this finding. [00:20:59] Speaker 01: Do your honors have further questions for Appley? [00:21:03] Speaker 01: I'll save my time for the government. [00:21:12] Speaker 05: Ms. [00:21:12] Speaker 05: Gershwin. [00:21:13] Speaker 03: Thank you. [00:21:14] Speaker 03: May it please the court? [00:21:15] Speaker 03: I'm here for the United States as intervener to address the constitutional issues. [00:21:21] Speaker 03: This court has now repeatedly held that there's no taking a retroactivity problem here in Selgi and Arthrex OSI. [00:21:30] Speaker 03: I'm happy to answer any questions. [00:21:32] Speaker 03: Otherwise, I'll rest on my briefs. [00:21:37] Speaker 05: Thank you. [00:21:38] Speaker 05: Thank you. [00:21:40] Speaker 05: Right, Mr. Oliver, you have the last word. [00:21:50] Speaker 00: Thank you, Your Honor. [00:21:53] Speaker 00: I would like to touch on something that Council said. [00:21:56] Speaker 00: They said something that's correct. [00:21:58] Speaker 00: That the Board mapped the claimed function. [00:22:02] Speaker 00: in the final written decision. [00:22:03] Speaker 00: It did do that. [00:22:04] Speaker 00: It says there is this function. [00:22:06] Speaker 00: But what the board did not do is map structure. [00:22:10] Speaker 00: Just because there might be an unmodified signal somewhere in a circuit does not mean that the board provided the analysis of what a decoder decoding means is. [00:22:21] Speaker 00: And is there a corresponding structure that provides a substantially unmodified signal? [00:22:26] Speaker 00: What Microsoft's attorney argues here [00:22:29] Speaker 00: counsel argues is that well there is some structure in there that someplace has an unmodified signal but if you look through the final decision you'll see nowhere where the board actually says okay you're pointing me to the entropy decoder in Venbrook's let me see where the corresponding structure is in the patent there is no analysis in that regard because that analysis cannot be done they point us to only an entropy decoder in Venbrook's the patent decoding means requires not only [00:22:59] Speaker 00: an overall decoder 76, but an entropy decoder therein. [00:23:04] Speaker 00: How can if we require the decoding mean structure this much with an entropy decoder inside, can we limit our analysis for the corresponding structure in the prior art to be only the entropy decoder? [00:23:15] Speaker 00: There is not a corresponding structure there, and there's no analysis from the board in that regard. [00:23:21] Speaker 00: And we believe at the very least this would be demanded for the board to actually provide that structural analysis. [00:23:27] Speaker 00: For that reason, we ask that the case be either reversed or remanded. [00:23:34] Speaker 05: Thank you. [00:23:35] Speaker 05: Thank you all. [00:23:36] Speaker 05: The case is taken under submission.