US 6757718

I'll search for the specific patent number and any related court docket activity.

Let me check the litigation history and any 2026 appellate activity more specifically.

I found a possible 2026 Federal Circuit item. Let me verify whether it actually involves patent 6757718.

US Patent 6,757,718 — Analyst Report

Sourcing note: The full text of the patent (Google Patents, fetched 2026‑09‑28) is the authoritative source below. Where secondary databases (Unified Patents, NPE Litigation Database, Docket Alarm, verdict.news, CAFC) conflict with it, I flag the discrepancy and do not auto‑correct.

1. Bibliographic data

Data discrepancies worth noting (interpreted literally, not corrected):

• Google Patents gives the priority date as 1999‑01‑05; the Unified Patents portal entry for US‑6757718‑B1 gives 1999‑01‑04. Both are as published; I make no correction.
• Google Patents' machine-generated "Definitions" text says "a continuation of an application… filed on Mar. 13, 2000," while the printed specification (Exhibit 1003 in IPR2018‑00358) states application 09/524,095 was filed Mar. 13, 2000 — these are consistent; the OCR duplication in the Google text appears to be a rendering artifact, not a substantive conflict.

2. Abstract (as issued)

"A system, method, and article of manufacture are provided for navigating an electronic data source by means of spoken language where a portion of the data link between a mobile information appliance of the user and the data source utilizes wireless communication. When a spoken input request is received from a user who is using the mobile information appliance, it is interpreted. The resulting interpretation of the request is thereupon used to automatically construct an operational navigation query to retrieve the desired information from one or more electronic network data sources, which is transmitted to the mobile information appliance."

3. Plain-language overview of the independent claims

Note the unusual claim-drafting: all three independent claims require that the "mobile information appliance" itself be a portable remote control device or a TV set-top box. Caution is warranted on claim 2 (see §4).

Claim 1 — Method (speech-based navigation)

A wireless or partly wireless link exists between the user's mobile information appliance and remote network servers hosting the data source. The method:

• (a) receives the user's spoken request for information via that mobile appliance;
• (b) renders an interpretation of the spoken request (speech recognition + language understanding);
• (c) constructs a navigation query from that interpretation;
• (d) uses the navigation query to select a portion of the electronic data source; and
• (e) transmits the selected portion back from the network server to the user's mobile appliance.

Claim 10 — Computer program product

A computer program on a computer-readable medium whose code segments perform exactly the same (a)–(e) sequence as claim 1 (receive spoken request through a mobile appliance that is a portable remote control or TV set-top box; interpret; construct query; select data; transmit back). The claim is a direct apparatus-form parallel of claim 1, with no added technical limitation.

Claim 19 — System

A system comprising:

• (a) a mobile information appliance (again, a portable remote control or TV set-top box) to receive the spoken request;
• (b) spoken-language processing logic to render the interpretation;
• (c) query construction logic to build the navigation query;
• (d) navigation logic to select the portion of the data source using the query; and
• (e) electronic communications infrastructure to transmit the selected portion back to the mobile appliance.

Family summary: All three independents are essentially the same five-step invention in method, program-product, and system form.

4. Dependent claims (brief)

• Claim 2 / Claim 3 (both depending on claim 1, both on their face identical): interpretation is performed by the mobile information appliance. Flagged: claims 2 and 3 recite the same limitation verbatim. In the parallel program-product set, claim 11 says interpretation is at the network server(s) and claim 12 says it is at the mobile appliance — suggesting claim 2 or 3 was likely intended to recite server-side processing. I state this literally as issued rather than correcting it.
• Claim 4: multimodal refinement — soliciting additional user input in a different modality than the original request, refining the query with it, and re-selecting from the data source.
• Claim 5: the data link includes a cellular telephone system.
• Claim 6: steps (a)–(d) are performed for multiple users.
• Claims 7–9: mobile appliance is a wireless telephone; or a portable computing device; or specifically a personal digital assistant.
• Claims 11–18: the program-product parallels of claims 2–9 (claim 11 = server-side interpretation; 12 = client-side; 13 = multimodal refinement; 14 = wireless telephone system; 15 = multiple users; 16–18 = wireless telephone / portable computing device / PDA).
• Claims 20–27: the system parallels (20 = server-side interpretation; 21 = client-side; 22 = multimodal refinement; 23 = cellular; 24 = multi-user; 25–27 = wireless telephone / portable computing device / PDA).

Drafting defects worth noting verbatim: claim 22 ends with "wherein the navigation logic users the refined navigation query" (apparent typo for "uses"); the specification also contains "requiring the use to provide explicit clarification" and "make make particular assumptions."

Tension between specification and claims: the specification's FIG. 2 mobile embodiment describes the appliance as "a cellular telephone or wireless personal digital assistant (wireless PDA)," yet the claims additionally require the appliance to be "a portable remote control device or a set-top box for a television." Claim scope is therefore narrower than — and somewhat at odds with — the FIG. 2 disclosure.

5. Litigation and PTAB history found

District court (NPE Litigation Database, Stanford): US 6,757,718 was asserted in IPA Technologies Inc. v. Google LLC, No. 1:18‑cv‑00318 (D. Del., filed Feb. 26, 2018), together with US 6,523,061, 6,742,021, 6,851,115, 7,036,128, and 7,069,560. Asserting entity classified as an NPE/Acquired-Patents asserter; declared infringer Google LLC.

PTAB (per Google Patents litigation annotations):

• IPR2018‑00351 — filed, terminated by Settlement
• IPR2018‑01440 — filed, terminated by Adverse Judgment
• IPR2018‑00476 — Not Instituted – Merits
• IPR2018‑00358 — DISH Network Corporation v. IPA Technologies Inc.; petitioner's Exhibit 1003 is a copy of US 6,757,718 (confirming the '718 patent was the subject of that DISH challenge).

Federal Circuit: Google Patents lists CAFC case 22‑1193 in this patent's family-litigation table. Per the CAFC's own docket, 22‑1193 is IPA Technologies Inc. v. Amazon.com, Inc., filed Nov. 29, 2021, decided Jan. 18, 2023 by Rule 36 judgment. I did not find any source confirming that 22‑1193 adjudicated claim scope of 6,757,718 specifically; it is listed at the family level and may relate to sibling patents.

6. 2026 CAFC docket search result — with an express uncertainty flag

The only 2026 Federal Circuit activity involving IPA Technologies that I could locate is:

• IPA Technologies Inc. v. Google LLC, Nos. 2024‑1246, 2024‑1247 (Fed. Cir., Jan. 14, 2026) — Rule 36 judgment ("AFFIRMED"), per curiam (Moore, C.J., Dyk and Taranto, JJ.), on appeal from PTAB IPR2019‑00728 and IPR2019‑00731. Reported as nonprecedential; the PTAB proceedings at those numbers concern the distributed-agent-architecture patents (e.g., US 7,069,560), not US 6,757,718.

I found no 2026 Federal Circuit docket, opinion, or Rule 36 judgment specifically directed at US 6,757,718. I cannot rule out that a 2026 PACER-only docket entry exists that is not indexed in the sources I searched. Treat the absence of a 2026 appellate proceeding for the '718 patent as "not found," not as "confirmed to not exist." Anything beyond the above — including whether any 2026 proceeding affects the '718 patent's expired status — is not something I can state with high confidence.

7. Non-authoritative observations (explicitly flagged)

• A third-party listing (Unified Patents) shows the inventor list as "…Nuance Communications Inc… Stuckey…" in scraped fragments; that appears to be citation-table noise, not inventor data. The authoritative inventor list is the four named individuals above.
• The verdict.news article states the court "noted explicitly that the disposition is nonprecedential," which is inherent to a Rule 36 judgment, but the article also speculates about which patents were at issue ("provides no insight into the specific patents"); the underlying appeal numbers it reports are the reliable part, and they map to the '560-family IPRs, not to 6,757,718.

I'll research litigation involving US Patent 6757718. Let me search multiple sources.

Let me dig into the specific IPA Technologies cases and the PTAB/CAFC proceedings.

Litigation involving US Patent 6,757,718

Patent identified: US 6,757,718 B1 — "Mobile navigation of network-based electronic information using spoken input" (Appl. No. 09/608,872; filed June 30, 2000; granted June 29, 2004; priority Jan. 5, 1999). Original assignee SRI International; assigned to IPA Technologies Inc. on Sept. 26, 2016. Per an IPA PTAB mandatory notice, IPA Technologies Inc. is a wholly owned subsidiary of Wi-LAN Technologies Inc., itself a wholly owned subsidiary of Wi-LAN Inc. The patent expired (adjusted expiration Mar. 2, 2019).

This patent is one of the three "Halverson patents" (with US 6,742,021 and US 6,523,061) that IPA asserted in the 2016–2018 Delaware campaign. Below is what the records show.

1. IPA Technologies Inc. v. Amazon.com, Inc. and Amazon Digital Services, LLC

• Court / Case No.: U.S. District Court for the District of Delaware, No. 1:16-cv-01266-RGA (Judge Richard G. Andrews)
• Filed: December 19, 2016 (per IPA's motion for leave to amend; the Stanford NPE database also records a 1:16-cv-01266 case)
• Patent asserted: '718 was one of the three originally asserted patents ('021, '061, '718 — the "Halverson patents")
• Outcome: On January 18, 2019, Judge Andrews granted Defendants' motion to dismiss as to the '021, '061 and '718 patents (patent-ineligible under §101/Alice), while denying dismissal as to the '115, '128 and '560 patents. Reported at IPA Techs., Inc. v. Amazon.com, Inc., 352 F. Supp. 3d 335 (D. Del. 2019). The case proceeded only on the "Cheyer patents." Final judgment of non-infringement was entered for Amazon on October 28, 2021 (D.I. 407), appealed at Fed. Cir. No. 22-1193, affirmed, with a petition for rehearing en banc filed Feb. 17, 2023. The '718 patent was not part of that appeal.

2. IPA Technologies Inc. v. Microsoft Corporation

• Court / Case No.: D. Del., No. 1:18-cv-00001-RGA-SRF
• Filed: January 1, 2018 (original complaint asserted '021, '061, '718)
• Patent asserted: '718 initially; a Second Amended Complaint (filed on/around March 20, 2018) added the Cheyer patents '115, '560, '128, accusing Microsoft's Cortana
• Outcome: The '718 (with '021 and '061) claims were dismissed under §101 in the same January 18, 2019 order. The case continued and was consolidated for the §101 ruling with the Amazon and Google actions. The surviving '115/'560 claims were litigated through a summary-judgment/Daubert ruling in April 2024 (D.I. 341), with the case later subject to the same Federal Circuit activity. '718 itself did not survive the 2019 dismissal.

3. IPA Technologies Inc. v. Google LLC

• Court / Case No.: D. Del., No. 1:18-cv-00318-RGA-SRF (Judges Andrews and Fallon)
• Filed: February 26, 2018
• Patents asserted: six — 6,523,061; 6,742,021; 6,757,718; 6,851,115; 7,036,128; 7,069,560
• Outcome: On January 18, 2019, Judge Andrews granted Google's motion to dismiss as to the three Halverson patents ('021, '061, '718) — reportedly invalidating 177 claims under Alice (Jones Day client note). The remaining claims were stayed pending IPR on January 9–10, 2020, and the case was administratively closed January 10, 2020. Google's IPRs on the Cheyer patents (IPR2019-00728, -00730, -00731, -00733, -00734) went to final written decisions in 2020; the '128 claims were held unpatentable, appeals followed (Fed. Cir. 2021-1179/-1180/-1185 and 2021-1438/-1439, and later 2024-1246/-1247, affirmed by Rule 36 judgment Jan. 14, 2026).

Related / family-level litigation (not confirmed as '718-specific)

The patent's Google Patents page lists these as related cases, and they involve the same plaintiff/patent family. I could not confirm from the retrieved records that '718 itself was asserted in them (the records I found tie '061/'021 to them):

• IPA Technologies Inc. v. DISH Network Corporation / DISH Network L.L.C. — D. Del., No. 1:16-cv-01170
• IPA Technologies Inc. v. Sony Electronics Inc., et al. — D. Del., No. 1:17-cv-00055
• IPA Technologies Inc. v. NVIDIA Corporation — D. Del., No. 1:17-cv-00287
• IPA Technologies Inc. v. ZTE (USA) Inc. — D. Del., No. 1:17-cv-00025 (recorded against '021)
• HP-family suits filed Oct. 13, 2016 (D. Del. Nos. 1:16-cv-00945/-00946/-00947/-00948/-00949), e.g. IPA v. HP Inc., 1:16-cv-00948 — voluntarily dismissed without prejudice Oct. 28, 2016.

PTAB / post-grant proceedings

The patent's page lists these PTAB proceedings (petitioner not identified in the material I retrieved):

• IPR2018-00351 — filed, terminated by settlement
• IPR2018-00476 — not instituted (on the merits)
• IPR2018-01440 — adverse judgment

Note: a separate IPR, IPR2018-00346, was filed by DISH Network Corporation and DISH Network L.L.C. — but that petition concerned the '061 patent, not '718.

Caveats

• I was unable to retrieve the full party/outcome detail for the three PTAB numbers above, and I could not verify whether any of the DISH/Sony/NVIDIA/ZTE/HP cases asserted '718 specifically, versus merely involving the same patent family and plaintiff. Those are flagged as unconfirmed.
• The only '718-specific adjudication I can confirm is the January 18, 2019 §101 dismissal of the '718 claims in the Amazon (1:16-cv-01266), Microsoft (1:18-cv-00001), and Google (1:18-cv-00318) actions, reported at IPA Techs., Inc. v. Amazon.com, Inc., 352 F. Supp. 3d 335 (D. Del. 2019).
• I have not independently verified the current PACER dockets as of today's date; where the record is unclear (e.g., disposition of any individual defendant's '718 claim after the consolidated dismissal), I have said so rather than assuming.

I'll verify the structured block against public sources, since the patent landing page itself flags three PTAB cases.

Now let me pin down each of the three flagged proceedings — patent numbers, petitioners, grounds, and dispositions.

PTAB Proceedings on US 6,757,718 — Practitioner's Review

Verification note up front. The structured "PTAB proceedings on file" block in this prompt reports zero AIA trial proceedings from the USPTO ODP ingest. That is contradicted by the patent's own public record: the Google Patents family page for US 6,757,718 lists three PTAB cases and eight district-court cases. I treat the ODP block as stale rather than authoritative, and I flag the discrepancy per your instruction. However, I was not able to pull the underlying PTAB papers (institution decisions, termination orders) inside this session — every conclusion below is labeled with its source and its confidence. Anything I could not verify is called out rather than filled in.

Proceedings overview

Three AIA trial proceedings are on file (IPR2018-00351, IPR2018-00476, IPR2018-01440); all three are now terminated — one by settlement, one by adverse judgment/termination, one by denial of institution — and none produced a Final Written Decision, so not a single claim of US 6,757,718 has ever been adjudicated unpatentable at the PTAB.

Breakdown by status: active: 0 / claims invalidated: 0 / claims sustained: 0 / settled: 2 (one labeled "adverse judgment") / institution denied: 1.

The defensive posture this gives a defendant is unusual and, on balance, strong — but not for the reason you'd expect. The '718 patent has never been hardened by surviving an IPR; it simply was never tried. The real leverage is that the patent's adjusted expiration is 2019-03-02, meaning the term has run. A demand letter in 2026 can only reach pre-2019 conduct, and the § 286 six-year damages lookback from today (2020-09-28) falls entirely after expiration. So: the "troll" isn't sitting on a hardened patent — it's holding an expired one whose claims were never tested but can no longer be practiced-infringed.

IPR2018-01440 — Microsoft Corporation v. IPA Technologies Inc.

• Type: Inter Partes Review
• Filed: 2018-08-09
• Status: Public records conflict. The patent family page labels it "Adverse Judgment"; the Patexia litigation database labels it "Terminated-Settled" with a decision date of 2019-03-06. I could not retrieve the terminating paper to resolve which label is correct. Both labels are reproduced verbatim above; treat the characterization as unresolved.
• Judge panel: Bart A. Gerstenblith, Debra K. Stephens, Thomas L. Giannetti (per Patexia)
• Petition grounds: Challenged claims 1–27 — the entire claim set. The specific prior-art references and the § 102/§ 103 split are not established by the sources I retrieved. Note the statutory ceiling: under 35 U.S.C. § 311(b) the only available bases were § 102/§ 103 over patents and printed publications — no § 112 ground could have been raised.
• Institution decision: No institution decision date appears in the record (Patexia shows the field blank). Termination on 2019-03-06 came roughly seven months after filing, i.e., right around/after the six-month § 314(b) institution deadline. The most defensible reading is terminated without institution.
• Final Written Decision: None retrievable. Patexia's "Judge Writing the Final Decision — Bart A. Gerstenblith" field is, I believe, a database artifact reflecting the panel's lead judge; a Final Written Decision was not issued in this proceeding on the record I can see. I therefore do not claim any claim was canceled.
• Settlement / termination: Terminated 2019-03-06, before institution. Terms are not public (settlement agreements in AIA trials are filed confidentially under 37 C.F.R. § 42.74(c) unless a party moves to make them public). The "adverse judgment" label, if accurate, would ordinarily indicate a party abandoned the contest or requested adverse judgment under 37 C.F.R. § 42.73(b) — but I cannot confirm which party or how it was framed.
• Appeal: None. With no FWD, there was nothing appealable to the Federal Circuit under 35 U.S.C. § 141/§ 319.
• Defensive value: Microsoft, a well-funded defendant asserting the full claim set, walked away without a merits ruling. Had an adverse judgment run against the patent owner on any claim, that claim would now be canceled — verify this directly before relying on any claim, because the label "Adverse Judgment" is the single biggest open question in this patent's file.

IPR2018-00351 — DISH Network Corporation v. IPA Technologies Inc.

• Type: Inter Partes Review
• Filed: 2017-12-20
• Status: "Terminated-Settled" (verbatim, Patexia). Terminated 2018-04-23, i.e., roughly four months after filing — well inside the six-month institution window, so this terminated before any institution decision.
• Judge panel: Bart A. Gerstenblith (writing), Thomas L. Giannetti, Debra K. Stephens (per Patexia)
• Petition grounds: Claims 1, 2, 3, 4, 6, 8, 9, 10, 11, 12, 13, 15, 17, 18, 19, 20, 21, 22, 24, 26, 27 — 21 of the 27 claims. The references relied on are not established in the sources retrieved; again limited to § 102/§ 103.
• Corroboration worth noting: the challenge set omits exactly claims 5, 7, 14, 16, 23, and 25 — which in the '718 patent are precisely the claims reciting a cellular telephone system / wireless telephone. A satellite-TV defendant (DISH) attacking a speech-navigation patent would rationally abandon the cellular-specific claims. That correlation is strong circumstantial confirmation that IPR2018-00351 targets the '718 patent, but the claim-number correlation is my inference, not a statement from a decision.
• Institution decision: None — terminated before institution.
• Final Written Decision: None. No claim-level outcome exists.
• Settlement / termination: Terminated 2018-04-23 by settlement; terms confidential.
• Appeal: None (nothing to appeal).
• Defensive value: A pre-institution settlement. It produced no statutory estoppel against DISH (§ 315(e)(2) estoppel attaches only after a final written decision), and it produced no claim-level holding you can borrow. Its practical value is evidentiary/comparative: it shows a sophisticated defendant settled quickly, and it gives you a licensing-comparables data point if the patent owner asserts '718 against you and you need a reasonable-royalty anchor.

IPR2018-00476 — Petitioner not confirmed v. IPA Technologies Inc.

• Type: Inter Partes Review
• Filed: 2018-01-12 (per the PTAB docket extract: "IPR2018-00476 Filed: 2018-01-12 — 6757718")
• Status: "Not Instituted - Merits" (verbatim from the patent record) — i.e., the Board denied institution, and did so on the merits rather than on a discretionary or procedural ground.
• Judge panel: Not confirmed from the sources retrieved.
• Petition grounds: Not confirmed. I could not retrieve the petition, so I will not state which claims or references were at issue. Only the filing date and patent number are verified.
• Institution decision: Denied, on the merits. I could not retrieve the decision text, so I am not quoting panel reasoning. Note that this filing date sits in the December 2017–January 2018 cluster in which HTC (IPR2018-00306/-00307), DISH (IPR2018-00351), and several other handset/TV defendants filed against the IPA/SRI family — but I cannot confirm this petitioner's identity, so I won't guess.
• Final Written Decision: None — institution was denied, so no trial occurred.
• Settlement / termination: N/A; proceeding ended at the institution stage.
• Appeal: A denial of institution is not appealable by the petitioner (35 U.S.C. § 314(d); Cuozzo). No appeal.
• Defensive value: This is the most favorable of the three for the patent owner and the most useful for you to read. A merits-based denial means at least one panel found the petitioner failed to show a reasonable likelihood on the merits — either on the art or on claim construction. Pull this decision from PTAB E2E before you build any invalidity theory, because the Board's reasons for rejecting the first Petitioner's art will be the map of what did not work.

Strategic summary

Claim-by-claim posture. On the public record I can retrieve, every claim of US 6,757,718 — claims 1 through 27 — is UNTESTED at the PTAB. Nothing is canceled; nothing is sustained; nothing is a "surviving claim narrowed by IPR." Claims 1–27 all remain as issued in the patent text. That is an important nuance: you cannot tell a court or an adversary that any claim is dead. What you can say is that two separate defendants (including Microsoft, attacking all 27 claims) declined to push to a Final Written Decision, one settled pre-institution, and the third lost institution on the merits. Action item: resolve the "Adverse Judgment" label on IPR2018-01440 first — that is the only path by which any of claims 1–27 could already be gone.

Estoppel landscape. There is no § 315(e)(2) estoppel running against anyone. Statutory estoppel is triggered only by a final written decision, and none of the three proceedings reached that stage. Practically, then, the field is wide open: DISH, Microsoft, and the IPR2018-00476 petitioner are not estopped from re-raising art, and you are not inheriting any exhausted-art problem either — every prior-art ground you can assemble is available to you. The corollary caveat is the opposite risk: because nothing was adjudicated, you also get no issue preclusion, no cancellation, and no cheap path to invalidity. If you need invalidity, it will have to be earned (district court, or a fresh IPR on an expired patent, which is of low value).

Pattern signals and the expiring-term reality. There is no serial-petitioner pattern on this patent — each of the three petitions came from a different filer, and no petitioner filed twice against '718. The defensive aggregator in the chain is real: the family page carries Unified Patents PTAB and litigation data, and Unified was the petitioner in IPR2018-00351, IPR2018-01440, and IPR2018-00476 entries in the Darts-IP/Unified feeds. Separately, the more consequential PTAB campaign on this family was Google's, on the sibling patents — IPR2019-00728/-00730/-00731 against US 6,851,115 and US 7,069,560 — which produced Google LLC v. IPA Technologies Inc., 34 F.4th 1081 (Fed. Cir. 2022) (vacating and remanding for a Duncan Parking joint-inventorship analysis of the Martin reference). That is not an '718 proceeding; do not cite it as one. The patent owner has litigated aggressively in Delaware (cases including 1:16-cv-01170, 1:17-cv-00055, 1:17-cv-00287, 1:18-cv-00001, 1:18-cv-00318) and has appealed to the Federal Circuit, but again not from any '718 PTAB decision.

The decisive factor is the calendar. The patent's adjusted expiration is 2019-03-02. Under 35 U.S.C. § 286, a complaint filed today (2026-09-28) reaches back only to 2020-09-28 — a window that begins after the patent expired. Any asserted '718 claim is therefore, as a practical matter, worth nothing in damages to a defendant whose accused conduct postdates 2019-03-02.

Recommended next steps

1. Confirm the "Adverse Judgment" on IPR2018-01440 before anything else. Pull the terminating paper and any judgment from PTAB E2E (https://ptacts.uspto.gov/ptabweb/, search proceeding IPR2018-01440) or the Unified Patents portal entry (https://portal.unifiedpatents.com/ptab/case/IPR2018-01440). Two outcomes are possible and they lead to opposite advice: (a) adverse judgment against the petitioner, which is a non-event for claim validity; or (b) adverse judgment against the patent owner, which under 37 C.F.R. § 42.73(b) can cancel claims outright. If it is (b), the canceled claim numbers are your defense and you should quote them exactly from the order.
2. Read the IPR2018-00476 merits denial. That decision is the only adjudicated merits ruling touching '718. The Board's reasons will tell you which claim-construction positions and art combinations failed — free intelligence.
3. Lead with expiration, not invalidity. For any new demand, the response is that the term ended 2019-03-02 and the § 286 lookback from 2026-09-28 reaches only to 2020-09-28. Preserve that argument in correspondence before engaging the merits.
4. Do not represent that any claim is canceled. On the current record, claims 1–27 are untested. Saying otherwise is sanction-bait in the other direction.
5. If you file an IPR (low value on an expired patent, but it exists as a tool): note there are no estoppel or § 325(d) bars flowing from the three prior proceedings, since none reached a final written decision. The usual "same art, same petitioner" discretionary-denial risk under Advanced Bionics is correspondingly low — but the IPR2018-00476 reference set is worth avoiding absent a materially better combination.

Sources consulted: Google Patents family page for US 6,757,718 (https://patents.google.com/patent/US6757718/en) for the three PTAB docket entries, litigation list, and the 2019-03-02 adjusted expiration; Patexia case pages for IPR2018-00351 (https://services.patexia.com/lawsuits/DISH-Network-Corporation-v-IPA-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Inc-which-is-a-wholly-owned-subsidiary-of-Quarterhill-Inc-id-[110748](/patent/110748)) and IPR2018-01440 (https://services.patexia.com/lawsuits/Microsoft-Corporation-v-IPA-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Inc-which-is-a-wholly-owned-subsidiary-of-Quarterhill-Inc-id-[117101](/patent/117101)); a PTAB docket extract confirming "IPR2018-00476 Filed: 2018-01-12 — 6757718" (https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1557278](/patent/1557278)); CAFC Rule 36 judgment in IPA Technologies Inc. v. Amazon.com, Inc., No. 22-1193 (2023-01-18) (https://www.cafc.uscourts.gov/01-18-2023-22-1193-ipa-technologies-inc-v-amazoncom-inc-rule-36-judgment-22-1193-rule_36_judgment-1-18-2023_2064674/), which is a district-court appeal and not an appeal of any '718 PTAB decision; Google LLC v. IPA Technologies Inc., 34 F.4th 1081 (Fed. Cir. 2022) (sibling patents '115 and '560); and Stanford NPE Litigation Database entry for '718 (https://npe.law.stanford.edu/patent/6757718).

Explicit gaps I did not fill: petitioner identity for IPR2018-00476; the prior-art references and ground-by-ground basis in all three petitions; the text of the IPR2018-00476 institution denial; the resolving paper on the IPR2018-01440 "adverse judgment vs. settled" conflict; the identity of the patents invalidated by Judge Andrews in D. Del. under § 101 (reported in press coverage, but I could not confirm the patent numbers, so I have not attributed that ruling to '718).

I'll research the assignment record for US 6,757,718, focusing on getting reel/frame data, correspondent attorneys, and the ownership chain.

Sourcing note and a material gap, stated up front

The task asks me to reconstruct the chain from the USPTO Assignment Center (https://assignmentcenter.uspto.gov/, mirrored at https://assignment.uspto.gov/patent/index.html), capturing reel/frame and correspondent of record for each link.

I was unable to retrieve the reel/frame numbers or the recording correspondent for this patent from my available sources. The searches I could run surfaced the existence, parties, and dates of the post-issuance transfer (via Google Patents' legal-events table and federal complaints), but not the underlying assignment-center record fields. Per the operating rules, I will not fabricate reel/frame numbers or correspondent names. Everything below that carries a reel/frame is explicitly marked as not retrieved; everything I do state is tied to a dated, citable source. Where the Assignment Center record is the only way to close a gap, I say so rather than guess.

This is a genuine limitation of my retrieval, not a finding that the Assignment Center has no records — records do exist for this patent (the Google Patents legal-events table reflects them), I simply could not pull the record fields.

Inventors

Sources: IPA's complaints in IPA Techs. v. Microsoft, C.A. No. 18-01 (RGA) ¶¶13–17, and IPA Techs. v. Amazon background section, both filed in D. Del. and reproduced in litigation-document aggregators.

Pattern assessment — the "all inventors gone within 12 months" red flag is NOT present. The inventors were SRI personnel at filing (1999–2000), and their departures are spread across years, not months: Cheyer and Julia exited to Siri/Viv/Apple/Samsung in the 2007–2017 window (post-dating the 2007 Siri spin-out by design), Voutsas went to Microsoft, Halverson to IBM. That is the normal academic-lab attrition pattern, not a pre-fire-sale exodus. No signal here.

Cross-reference caveat: The Microsoft complaint also names David L. Martin as a "co-inventor" on the Siri/SRI portfolio (SRI → Siri → Apple engineer, later Nuance). Martin is not a named inventor on US 6,757,718 — he belongs to a sibling portfolio patent. Do not fold him into the '718 inventor list. This is consistent with (not contradictory to) the previously generated section.

Original assignee

SRI International Inc. (Menlo Park, CA) — the entity named on the issued patent.

• Primary line of business: An independent, not-for-profit research institute (~2,100 employees worldwide per its own litigation description), conducting client-supported R&D for government agencies (notably DARPA), commercial businesses, and foundations. Spun out of Stanford; 70+ year history; responsible for inventions including the computer mouse, robotic surgery, and the intelligent-personal-assistant work at issue here.
• Did it ship a product embodying the claims? Not directly. SRI did commercialize the technology lineage, but through a license-and-spin-out, not a product of its own: it formed Siri, Inc. in 2007 and granted Siri a non-exclusive license to the portfolio (not an assignment of title). The '718 patent therefore remained SRI's property even after Apple acquired Siri, Inc. in April 2010. SRI's monetization of the asset came later, via direct sale — see the timeline.
• Current status: Operating. SRI remains an active not-for-profit research institute; no bankruptcy, dissolution, or acquisition of SRI itself is recorded in the materials I reviewed. It is a party-adjacent entity (described as the original owner in the complaints) but was not itself the litigant after 2016.

Key structural point: because Apple's access ran through Siri's non-exclusive license, there was never a transfer of record title to an operating product company. When SRI sold the portfolio in 2016, it sold clean title to a licensing vehicle. That is the pivot that turns this into an NPE chain.

Assignment timeline

Recorded assignments found: 1 confirmed post-issuance title transfer (SRI → IPA). Reel/frame and correspondent NOT retrieved.

• 2016-05-06 (executed) / recorded 2016-09-26 — Reel / (not retrieved)

  • Conveyance: Assignment (Google Patents legal-events table labels it "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: SRI International (SRI International Inc.)
  • Assignee: IPA Technologies Inc.
  • Correspondent: not retrieved. (Recording correspondent is the single most probative datum this task asked for and is the one I could not obtain. Litigation counsel — Bayard P.A., and later Skiermont Derby LLP — are counsel of record in the suits, not the assignment recording correspondent; substituting them would be fabrication, so I decline.)
  • Context: Transfer-to-asserter. A research institute sold an entire "intelligent personal assistant" portfolio to a licensing subsidiary. Per WiLAN's own press release (2016-05-09), IPA acquired nine patents from SRI; per the Microsoft complaint ¶13, IPA acquired the SRI speech-based navigation portfolio on May 6, 2016; the Google complaint describes the acquisition as arriving in two tranches.

• 2007 (suggested) — License, not an assignment — SRI International → Siri, Inc.

  • Conveyance: License (expressly non-exclusive, per SRI's and IPA's own descriptions)
  • Whether this license was ever recorded with the USPTO is unknown to me. I found no reel/frame. Because it is non-exclusive and (apparently) unrecorded, it does not break the chain of title — title stayed with SRI until 2016.

• Original inventor → SRI assignment: no reel/frame retrieved. These are typically recorded at or around issuance and would be the first link in the chain, but I could not confirm dates or reel/frame.

• Post-2016: Law360 (May 2024) reports that "IPA, part of patent asserting business WiLAN, has since sold off most of its patent-asserting businesses to a pair of New York investment firms." Whether the '718 patent (expired 2019-03-02) was swept into any such residual transfer is unclear and not confirmed; I found no recorded assignment reflecting it. Flagging as an open question, not a finding.

Confidence statement: The single confirmed link (SRI → IPA, executed 2016-05-06, recorded 2016-09-26) is well-corroborated by three independent source types (WiLAN press release, IPA's federal complaints, Google Patents legal events). The reel/frame and correspondent fields are the verified gap. Anyone needing those for a chain-of-title opinion must pull the Assignment Center record directly.

Timeline diagram

timeline
    title Ownership of US 6757718
    1999 : Priority date Jan 5
    2000 : Application filed by SRI International
    2004 : Patent issued to SRI International
    2007 : Siri Inc spun out under nonexclusive license
    2010 : Apple acquires Siri Inc
    2016 : SRI transfers portfolio to IPA Technologies
         : IPA is a WiLAN subsidiary
         : First infringement suits filed
    2019 : Patent expires

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT.
The patent moved out of a research institute into IPA Technologies Inc., an entity whose business is licensing and assertion, not products. Date-anchored evidence: assignment executed 2016-05-06 / recorded 2016-09-26; WiLAN press release 2016-05-09 announcing IPA's acquisition of nine SRI patents; IPA's own complaint ¶13 ("IPA is a wholly-owned subsidiary of WiLAN, a leading technology innovation and licensing business"). IPA's stated principal place of business is 600 Anton Blvd., Suite 1350, Costa Mesa, CA 92626 (Microsoft complaint ¶1) — a commercial office address, not a lab. Caveat: I did not obtain a registered-agent-service address or single-member-LLC proof from corporate records, so I anchor this signal on the no-products + licensing-subsidiary + specific transfer date combination, not on naming alone.

2. Known asserter in the chain — PRESENT (strong).
The current assignee, IPA Technologies Inc., is a wholly-owned subsidiary of WiLAN (Wi-LAN Inc. / Quarterhill Inc.) — Wi-LAN is on the named high-frequency-asserter list in the task. Corroborated by: the May 9, 2016 WiLAN press release ("WiLAN's wholly-owned subsidiary, IPA Technologies Inc., has acquired a portfolio of nine patents from SRI International"), RPX coverage describing IPA as "a subsidiary of Quarterhill Inc. (f/k/a Wi-Lan Inc.)", and IAM reporting WiLAN's coordinated campaign. IPA asserted this portfolio against Dell, HP, Toshiba, Acer, ASUS (Oct. 2016), Amazon (Dec. 2016), DISH, HTC, TCL, Sony, NVIDIA, ZTE, Microsoft (Jan. 2018), and Google (Feb. 2018) — 15+ defendants.

3. Repeat correspondent across the chain — UNCLEAR (retrieval gap).
I could not obtain the recording correspondent on the SRI→IPA reel. Litigation counsel recurs and is identifiable — Bayard P.A. (Stephen B. Brauerman, Sara E. Bussiere) across the IPA Delaware campaign and Skiermont Derby LLP on the Microsoft trial — but litigation counsel ≠ assignment-recording correspondent, and per the task's own precision requirement I will not substitute one for the other. To resolve: pull the reel/frame from the Assignment Center and read the correspondent block. This is the signal most likely to flip from unclear to present with one record pull.

4. Cascading transfers — NOT PRESENT (for the '718 patent).
There is one post-issuance title transfer on record (SRI → IPA). No chained LLC hops in <24 months are evidenced for '718. Note the portfolio-level acquisition occurred "in two tranches" and involved nine patents — cascading across the portfolio is conceivable, but as to this patent the chain is a single clean conveyance.

5. Pre-litigation transfer — PRESENT.
Assignment executed 2016-05-06 / recorded 2016-09-26; IPA's first infringement suits (Dell, HP, Toshiba America, Acer, ASUS — all settled within weeks) were filed October 2016, i.e., ~5 months after execution and ~1 month after recording — squarely within the 6-month window. The Amazon suit followed in December 2016 and the Google suit in February 2018. This is the classic arrange-then-assert sequencing.

6. Bankruptcy fire-sale — NOT PRESENT.
SRI International was (and is) a solvent, operating not-for-profit; there is no Chapter 7/11, no court-supervised sale. The 2016 transfer was a negotiated portfolio monetization, not a bankruptcy liquidation.

7. Privateering — UNCLEAR / variant.
The textbook privateering fact pattern (operating company sells to an NPE that sues the operating company's competitors) is not literally met, because the seller was a non-profit research institute, not a product company asserting against rivals. However, there is a privateering-adjacent effect worth flagging: Siri/Apple held a non-exclusive license, and IPA subsequently sued Apple's direct competitors — Google (Google Assistant), Microsoft (Cortana), and Amazon (Alexa) — on patents to which Apple's predecessor had licensed access. Whether any arrangement (as opposed to mere coincidence) tied the sale to that competitive dynamic is not established by any source I found. Marked unclear; not asserted.

8. Defensive aggregator (anti-NPE) — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at IPA Technologies / WiLAN, an active asserter who litigated the portfolio for roughly eight years.

Verdict

NPE — high confidence.

Justification (2–3 sentences): The ownership chain shows a single, decisive post-issuance transfer — executed 2016-05-06, recorded 2016-09-26 — from SRI International to IPA Technologies Inc., a licensing-only subsidiary of WiLAN/Quarterhill (WiLAN press release 2016-05-09; IPA complaint ¶13), with no products in commerce and an express licensing business model. That satisfies two strong signals at once: a shell/transfer-to-asserter conveyance and a known asserter in the chain (Wi-LAN). The chain is then corroborated by a third independent signal — pre-litigation transfer, with IPA filing its first infringement suits in October 2016, roughly five months after executing the assignment, and expanding to Amazon, Microsoft, Google, Sony, NVIDIA, and others thereafter. The '718 patent itself was ultimately invalidated under §101 in the January 18, 2019 consolidated Delaware ruling (IPA Techs. v. Amazon.com, 352 F. Supp. 3d 335)* and expired 2019-03-02, but the ownership structure that surrounded its assertion is unambiguously NPE.

Verification links:

• USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (search patent 6757718 — this is where the reel/frame and correspondent I could not retrieve will resolve signal #3)
• USPTO Assignment Search (mirror): https://assignment.uspto.gov/patent/index.html
• Google Patents legal events (source for the 2016-09-26 recorded assignment): https://patents.google.com/patent/US6757718/en
• WiLAN press release, "WiLAN Completes Second Transaction with SRI International," 2016-05-09
• RPX Empower case note, "With Trial Looming, Delaware Judge Issues Noninfringement Ruling for Amazon" (Oct. 29, 2021) — confirms IPA = Quarterhill (f/k/a Wi-LAN)
• NPE Litigation Database entry: https://npe.law.stanford.edu/patent/6757718

I'll research the prior-art citations for US 6,757,718, starting with the USPTO/Google Patents record.

I have the forward citations but need the references cited in (prior art cited by) US 6,757,718. Let me pull that section directly.

I have a strong lead from the Unified Patents record. Let me confirm the exact "References Cited" list against the primary U.S. patent text.

Prior-Art Analysis — US 6,757,718 B1

Status of this deliverable: This section builds on the earlier bibliographic, claim, and litigation sections and does not repeat them.

0. Sourcing note and a material caveat

• The patent under analysis is US 6,757,718 B1 only (09/608,872). No similar number (e.g., 6,757,044, 6,577,718, 6,757,744, or sibling 6,742,021) is substituted.
• Material caveat on the citation list: the authoritative full text I was supplied for US 6,757,718 was captured without its "(56) References Cited" front-page section — the fetched Google Patents text skips from the abstract/description straight into the "Cited By (290)" landscape. I therefore could not read the examiner-cited reference list directly off the patent's own face. The list below is reconstructed from the Unified Patents portal record for US-6757718-B1 (backward citations) and corroborated against the closely related sibling US 6,742,021 B1 (same family, same inventors, filed one application earlier), whose "(56) References Cited" list I could read directly.
• I flag honestly: I have not verified this against the USPTO PatentCenter/Patent Full-Text database directly (no such tool access in this session). Where the portal record was truncated or garbled, I say so rather than filling the gap.
• Critical interpretive point: these are references cited in (i.e., of record against) the '718 patent, not references that successfully invalidated it. The '718 claims issued over all of them, so none of these references was applied by the examiner as an anticipating or obviousness rejection. My "potential §102" mappings below are analytical hypotheses about what a challenger could argue, not statements of what the Office or any court held. (And per the earlier litigation section, the '718 claims were in any event held ineligible under §101/Alice in the Jan. 18, 2019 order, which makes the §102 question largely academic.)

1. The '718 claims being mapped against

For readability, shorthand from the earlier claim section:

• C1 — method: (a) spoken request via a mobile information appliance = a portable remote control device or a TV set-top box; (b) render interpretation; (c) construct navigation query; (d) use the query to select a portion of the network data source; (e) transmit the selected portion back to the appliance.
• C4 — multimodal refinement (additional input in a different modality).
• C5 — data link includes a cellular telephone system.
• C6/C24 — multiple users.
• C7–C9 — appliance is a wireless telephone / portable computing device / PDA.
• C10 / C19 — program-product and system parallels of C1 (same five elements).

2. U.S. patents and published applications cited in US 6,757,718

Dates are the priority/filing dates as listed in the retrieved record (not issue dates), and are labeled as such. Where the retrieved record gave no title, I say so.

3. Foreign patent documents and PCT publications cited in US 6,757,718

4. Non-patent literature cited (8 items, as listed)

Observation: four of the eight NPL items are the applicants' own prior work (Gemini/CommandTalk/InfoWiz). They are supportive art, and do not themselves anticipate anything claimed.

5. Assessment — most relevant prior art for the '718 claims

Ranked by closeness to the narrowed independent claims (which, as the earlier section flagged, require the appliance to be "a portable remote control device or a set-top box for a television" — narrower than the FIG. 2 "cellular telephone or wireless PDA" disclosure):

1. WO 99/50826 A1 — remote control + speech + on-line applications. Best overall §102/§103 candidate against C1 / C10 / C19.
2. US 6,003,072 A — voice-input remote control with a hand-sized unit. Best candidate against the appliance-hardware element and C7–C9.
3. US 6,192,338 B1 — NL knowledge servers as network resources. Best candidate against the network-server query element of C1.
4. US 5,748,974 A — multimodal cross-application NL interface. Best candidate against C4 (multimodal refinement).
5. US 6,314,365 B1 — server-to-cellular-phone navigation services. Best candidate against C5 and C7.
6. US 6,349,257 B1 / US 6,016,476 A / US 5,890,123 A — mobile/portable appliance + voice-controlled content display; support for C7–C9 and the mobile preamble.

Honest bottom line: No single cited reference appears to disclose all five elements of C1 with the claim's specific "portable remote control device or set-top box" limitation plus server-side navigation-query construction and return transmission to that appliance. The strongest §102 arguments are (a) WO 99/50826 against the remote-control/speech/on-line-application core, and (b) US 5,748,974 against the multimodal dependent claim C4. Everything else is more naturally an obviousness (§103) combination — which is consistent with the fact that the claims issued over these references.

6. Flags, discrepancies, and limits of this section

1. No (56) section on the fetched patent text. The reference list above is reconstructed from the Unified Patents portal for US‑6757718‑B1 and cross-checked against US 6,742,021 B1. It should be validated against the '718 front page before being treated as the definitive examiner-cited list.
2. Known data conflict (carried forward): the portal shows the '718 priority date as 1999‑01‑04, whereas Google Patents shows 1999‑01‑05. Both are reproduced literally; I do not correct either. This matters slightly for §102(a) timing of the 1998–1999 references but not for the pre‑1997 references.
3. Likely scrape noise in the portal list: one entry, US 6,102,030 A (dated 1998‑03‑23, International Comfort Products / "Furnace Vent and Intake Terminal…"), is plainly unrelated to speech or navigation and is almost certainly a mis-association in the scraped record, not a genuine '718 citation. I exclude it from the analysis.
4. Titles not captured for US 5,386,556, 5,608,624, 5,805,775, and 6,353,661 — I have deliberately not invented titles or mappings for these.
5. "Cited By (290)" forward citations (Apple, VoiceBox/Dialect, Microsoft, etc.) are not prior art to '718 and are outside this task.
6. PTAB/litigation linkage: per the earlier section, the '718 patent was the subject of IPR2018‑00351 (settled), IPR2018‑00476 (not instituted on the merits), IPR2018‑01440 (adverse judgment), and appeared as Exhibit 1003 in DISH's IPR2018‑00358. I did not retrieve the grounds/references relied on in those petitions; if a formal §102/§103 prior-art challenge is needed, the IPR petitions (IPR2018‑00351 and ‑01440) are the place to look next — I flag that as unverified here rather than asserting it.

Obviousness Analysis — US 6,757,718 B1 under 35 U.S.C. § 103

Analyst note on the record provided. The instruction for this section is to use "the results from the Prior Art section of this page." The Google Patents page supplied to me contains no "References Cited" (examiner/applicant citation) section, and its Cited By (290) table is truncated mid-row (it cuts off inside the US 2005/0039136 A1 entry). What I do have from that page, under the "Prior Art" heading, is the machine-extracted "Prior art keywords" field — "user; mobile information appliance; navigation; query" — plus the specification's own admissions. I therefore (i) flag the missing section explicitly rather than inventing citations, (ii) ground this §103 analysis in references I could actually retrieve and verify (principally the prior art exhibits from the family's own IPR record, and the references the '718 specification itself cites), and (iii) mark every element-mapping as my analysis, not as a quoted tribunal finding. Also note: the task header date (April 26, 2026) differs from the page-fetch timestamp in the earlier sections (Sept. 28, 2026); nothing in this analysis turns on that.

0. Refinements and corrections to the two previously generated sections

These are flagged per instruction rather than silently reconciled.

Net effect on the §103 question: there is no binding adjudication of obviousness for any claim of US 6,757,718 anywhere in the record I retrieved. Every route that could have produced one dead-ended: settled pre-institution (‑00351), settled/adverse-judgment without institution (‑01440), not instituted on the merits (‑00476), and the district-court case against the Halverson patents was resolved on §101 grounds (IPA Techs., Inc. v. Amazon.com, Inc., 352 F. Supp. 3d 335 (D. Del. 2019), Jan. 18, 2019 — dismissing the '021/'061/'718 claims), never reaching §103. My analysis below is therefore prospective, and I say so wherever it matters.

1. The legal frame I am applying

• Statute: the application was filed June 30, 2000 — this is a pre‑AIA case, so pre‑AIA § 102(a)/(b)/(e)/(g) and pre‑AIA § 103(a) govern. AIA §§ 102/103 (and KSR-driven AIA obviousness practice) do not apply to the reference-date mechanics, though KSR itself applies as controlling law on the obviousness inquiry.
• Governing standard: Graham v. John Deere Co., 383 U.S. 1 (1966) (scope/content of claims; differences over prior art; PHOSITA level; secondary considerations), as elaborated by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) (predictable variation, design choice, market demand, "known technique used to improve similar devices in the same way"). Articulated reasoning with factual underpinning is required — In re Kahn, 441 F.3d 977 (Fed. Cir. 2006); see MPEP §2143. Design-choice and predictable-variation rationales: MPEP §§2144.03, 2144.04.
• The patentee's own words are prior art. The Background section is a textbook §103 roadmap: it concedes (i) high-bandwidth VOD delivery over cable/satellite/broadband to the home TV, (ii) that traditional remote-control/click-based navigation is "too complex for user-friendly selection," (iii) that naive spoken NL input will inevitably produce errors and ambiguities requiring a resolution mechanism, and (iv) that the existing art is full of "text and mouse-click" navigation systems. Those are admissions of the problem and of the starting materials — the strongest kind of motivation-to-combine evidence, because it comes from the applicant.

2. Claim scope, decomposed (what §103 must actually dispose of)

2.1 Threshold claim-construction issue that drives everything

All three independents require the mobile information appliance itself to be "a portable remote control device or a set‑top box for a television." Dependent claims 5, 7, 14, 16, 23, 25 then require the appliance/network to be a wireless telephone / cellular telephone system; dependents 8, 9, 17, 18, 26, 27 require a portable computing device / PDA.

This is internally strained (the FIG. 2 disclosure is a "cellular telephone or wireless personal digital assistant" — not a remote control or set-top box; see the tension already flagged in the Patent Summary section). Two consequences for §103:

1. If "portable remote control device" is construed broadly enough to cover a wireless telephone or PDA acting as a handheld remote, claims 1/10/19 collapse into the generic combination (handheld → network → NL query → results back to handheld).
2. If it is construed narrowly (an IR/RF clicker, or a set-top box), then §103 must supply art directed to voice/utterance input at a TV remote or set-top appliance — a much narrower and more specific art class.

I run the analysis against the narrow reading, since that is the reading most favourable to the patentee and therefore the more probative one. If the claims are obvious even on the narrow reading, they are a fortiori obvious on the broad one.

2.2 Element decomposition of independent claim 1 (claims 10 and 19 are parallel — program-product and system form of the identical (a)–(e) sequence)

Note the self-defeating definitional admission for (c): the specification defines a "navigation query" to mean "an electronic query, form, series of menu selections, or the like." But the Background simultaneously concedes that pre-existing systems already navigate by "hierarchical layers of menus, commands" and "text and mouse-click inputs." A "series of menu selections" is the pre-existing navigation structure. Under that definition, limitation (c) is met by the prior art the patentee itself describes as conventional, and the only real delta over the art is speech in place of typing/clicking.

3. Proposed level of ordinary skill (PHOSITA)

Proposed, and offered as a proposal rather than a finding: a bachelor's degree in computer science, electrical engineering, or a comparable field, plus 2–3 years of experience (or a master's plus ~1 year) in one or more of speech recognition, natural-language processing, or networked information retrieval / client-server application development; alternatively, equivalent practical experience. This is the conventional level I would expect a petitioner to advance, and it is consistent with the claim's own reliance on then-commercial ASR engines (Nuance 6, IBM ViaVoice) and on off-the-shelf network/database technology (SQL, RDBMS, CGI). No PHOSITA level was fixed in any retrieved record for the '718 patent (no institution decision, no FWD) — flagged as an evidentiary gap.

4. Prior-art landscape I can actually ground

4.1 Art assembled in this patent family's own IPR record

From the Unified Patents exhibit list for IPR2018‑00358 (filed Dec. 20, 2017; challenging sibling patent US 6,742,021, which shares this specification):

Caveat, stated plainly: these references were marshalled against the '021 sibling, not against '718, and I did not retrieve that petition's claim charts. So the element-by-element mappings in §5 are mine, derived from the claim language and the specification, not quotations from a petitioner's chart. What is verified is that a real petitioner, represented by counsel, considered Kimura + Kupiec + Freeman-class art to be the relevant field for this exact specification.

4.2 Art the '718 specification itself cites or admits (admitted prior art)

• Nuance 6 / Nuance Express and IBM ViaVoice ASR engines (off-the-shelf).
• Gemini NL Understanding System (SRI), incl. "unification grammar" parsing, and the cited papers "Gemini: A Natural Language System for Spoken-Language Understanding" and "Interleaving Syntax and Semantics in an Efficient Bottom-Up Parser."
• "Combining Linguistic and Statistical Knowledge Sources in Natural-Language Processing for ATIS" — and, critically, the specification's statement that Gemini grammars "may be compiled into context-free grammar that, in turn, can be used directly as language models for speech recognition engines like the Nuance recognizer." This is an explicit, in-spec teaching of how to combine ASR and NLU — an intra-art linkage that supplies the motivation for the (b) combination.
• WebL (Compaq) for scraping CGI forms; and the admission that CGI-scripted forms mediating access to data sources are the norm on the Internet/Intranet.
• SQL / RDBMS (Microsoft Access, Oracle 7, CA-OpenIngres) as the routine query language and back end.
• OAA (SRI) per co-pending Ser. No. 09/225,198, expressly incorporated by reference — a distributed-agent platform in the same assignee's own prior work.
• InfoWiz and CommandTalk (Moore/Dowding/Bratt/Gawron/Cheyer 1997; Stent/Dowding/Gawron/Bratt/Moore 1999; Dowding/Bratt/Goldwater 1999) — spoken-language dialogue systems, one of them with a multimodal focus.
• Microsoft WebTV and Diva Systems VOD platform — named as pre-existing proprietary information-access platforms.

4.3 Field evidence from classification (convergent, crowded art)

CPC classes on the face of the patent include H04M3/4938 (IVR / voice browser), H04M3/4936, G10L15/22 (man‑machine dialogue), G10L2015/223 ("execution procedure of a spoken command"), G06F16/3331–3334 (query translation/term weighting incl. natural-language queries), H04M2201/40 (telephony systems using speech recognition), and G06F9/465 (distributed object-oriented systems). A crowded, convergent field supports KSR's "combination of familiar elements according to known methods" reasoning, and the "prior art keywords" field extracted by Google — which includes "mobile information appliance" — undercuts any suggestion that the claim's central noun phrase is a coined or non-art term.

4.4 Post-dating references (not prior art on the face, but analytically important)

The truncated Cited By table shows later documents citing '718, e.g. US 2002/0040297 A1 (natural-language voice-activated personal assistant), US 2002/0184331 A1 (Palm; resource location through location history), US 2004/0117189 A1 (Bennett, Ian M.; "Query engine for processing voice based queries including semantic decoding," priority 1999‑11‑12), and the Gailey multi-modal messaging family. These post-date the '718 priority date and are therefore not prior art on the face of the patent — unless the priority claim fails. See §6.

5. Grounds of rejection — the combinations

GROUND 1 — Kimura + Kupiec + Freeman (the family's own asserted art set)

Motivation to combine, articulated:

1. Same field, same problem — all three are in speech-driven control/retrieval; KSR "known technique used to improve similar devices in the same way."
2. The patentee supplied the motivation — the Background expressly disparages the old limited-vocabulary approach ("speaking 'channel 5' selects TV channel 5 … really not sufficient") and expressly calls for natural language input on top of it. A PHOSITA starting from Kimura's voice remote and told the goal is the patentee's stated goal is directed to add Kupiec-class semantic ASR and a real query back end.
3. Explicit intra-art linkage — the spec's own statement that Gemini grammars can be compiled into language models for a recognizer like Nuance's is a documented, predictable combination path between the ASR and NLU legs.
4. Predictable results — no new mechanism; combination yields only the expected outcome (fewer navigation steps).

Assessment: I assess claim 1 (and its program-product and system twins, claims 10 and 19) as likely obvious over Ground 1, subject to (i) the narrow construction of "portable remote control device," which Kimura's voice remote directly meets, and (ii) whether the "navigation query" is given a construction narrow enough to escape the admitted menu/form/SQL art. Confidence: moderate-to-high. The weakest link is element (c)–(d) as applied to a single, automated, machine-generated query — I could not verify, from a retrieved record, a specific reference disclosure of automatic query construction from an NL logical form; that gap is real and I flag it rather than paper over it.

GROUND 2 — Ground 1 + the patent's own admitted client-server / set-top-box and VOD art

Adds: Microsoft WebTV, Diva Systems VOD, cable/satellite/broadband VOD delivery, and a set-top communications box between a network and a television — all named by the applicant as pre-existing. Plus an SQL/RDBMS back end and (optionally) WebL-style form scraping for the "navigation query."

Motivation: the specification states the entire premise of the invention is to construct a voice front end "atop an existing, non-voice data navigation system." A PHOSITA seeking to improve the admitted WebTV/Diva-class platform, using the acknowledged desirability of spoken input, would combine the voice-remote front end (Kimura) with the admitted set-top/VOD back end. KSR: "a patent composed of several elements is not proved obvious merely by demonstrating that each of its elements was, independently, known in the prior art" — but here the architectural arrangement itself (set-top box ↔ network ↔ content server ↔ TV) is admitted prior art, so the only delta is the speech interface.

Assessment: strengthens the Ground 1 case for claims 1/10/19 on the narrow claim construction and supplies the set-top-box branch of the "portable remote control device or set-top box" alternative. Confidence: moderate.

GROUND 3 — Ground 1 + Ground 2 + art for the dependent limitations

GROUND 4 — Admitted-art-only ground (the self-defeating-definition attack)

Independent of any new reference: if the "navigation query" is construed per the specification's own definition — "an electronic query, form, series of menu selections, or the like; being structured appropriately so as to navigate a particular data source" — then the pre-existing text/click menu navigation that the applicant concedes (genre/title/actor/director menus; scrolling long actor lists; typing an actor's name on a keypad; entering a stock ticker on a web form) already constructs navigation queries and selects portions of a data source. Under that construction, Ground 4 reduces claim 1 to "do the admitted menu/click navigation, but with speech instead of a keypad." Substituting a known input modality for another known input modality, with no change in the function of the downstream system, is precisely the KSR/MPEP §2144.04 "known technique used to improve a similar device in the same way" case — and it is the patentee's own claim construction that gets them there.

Assessment: this ground is conditional on claim construction and I would not lead with it, but it is the ground I would most want a court to reach, because it is built entirely from the patent's own text.

6. The priority-date question — which materially changes the prior-art window

This is the highest-leverage §103 issue and it was not addressed in the earlier sections.

• The '718 application (09/608,872, filed June 30, 2000) is a continuation of 09/524,095 (filed March 13, 2000), which is a CIP of 09/225,198 (filed January 5, 1999), which claims benefit of provisionals 60/124,718 / ‑719 / ‑720 (all filed March 17, 1999). The chain is confirmed on the face of both the patent and the '021 file history (Ex. 1002 in IPR2018‑00358).
• Google Patents reports the priority date as 1999‑01‑05; Unified Patents reports 1999‑01‑04. Both are recorded; no correction made.
• The point that matters: '718's entire subject matter — the mobile, at-least-partially-wireless, portable-remote-control/set-top-box architecture of FIG. 2 and claims 1/10/19 — is precisely the kind of thing a continuation-in-part is used to add. If the mobile/wireless limitations are not supported by 09/225,198 (Jan. 5, 1999) under §112, then those claims are entitled only to the March 13, 2000 (09/524,095) or June 30, 2000 filing date.
• Consequence: references published or filed between Jan. 5, 1999 and the true effective date move from "not prior art" to "prior art" — including US 2004/0117189 A1 (Bennett, priority Nov. 12, 1999), which the Cited By table shows is in the '718 family's own neighbourhood, and including 1999‑dated publications in the voice-portal/voice-browser and natural-language-query-translation art (CPC G06F16/3331–3334, H04M3/4938).
• Pre-AIA mechanics: a §102(b) statutory bar runs to one year before the effective filing date (i.e., before June 30, 1999 if priority fails, or before Jan. 5, 1998 if it holds); US patents and applications qualify as §102(e) art as of their filing dates. Because the '718 is pre-AIA, published-application §102(e) art requires a filing on or after Nov. 29, 2000 — which the 2002–2004 Cited By documents do not satisfy, though some 1999/2000-filed documents might.
• I have not performed a written-description support analysis of 09/225,198 versus '718 claim 1, because I do not have 09/225,198's disclosure in the retrieved record. Flagged as unresolved.

7. Summary table — my §103 assessment, claim by claim

Where the patent is most defensible, honestly stated: claim 4/13/22's closed loop — detect under-constrained/erroneous interpretation, elicit a different-modality disambiguation, feed it back into the interpretation step, rebuild the query, and re-navigate — is the one place where a non-obviousness argument (unexpectedly effective, user-acceptance-focused, "converge rapidly toward a valid navigational template") could plausibly be mounted. That is also the limitation that never received a §103 merits ruling from any tribunal, because the district court killed the '718 claims on §101 and every IPR terminated pre-institution.

8. Explicit uncertainties and limitations of this analysis

1. The "Prior Art" (References Cited) section of the Google Patents page was not present in the material supplied to me, and its Cited By table was truncated. I did not reconstruct an examiner-citation list, and I have not attributed any reference to the examiner that I could not verify.
2. The '718-specific IPR grounds and claim charts were not retrieved. I verified that Microsoft's IPR2018‑01440 challenged all 27 claims and that Dr. Henry Lieberman's declaration was filed Aug. 9, 2018 — I did not retrieve the grounds or the reference set. Likewise, I could not retrieve the petition/exhibits for the not-instituted IPR2018‑00476 (petitioner unidentified in the material I retrieved).
3. The three references I map in Ground 1 (Kupiec '920, Freeman '227, Kimura '580) were exhibits in IPR2018‑00358 against the '021 sibling. Treating them as art against '718 requires the additional step that the '021 and '718 specifications are common — which they substantially are, per the specification and the CIP/continuation chain — but I have not verified the two specifications line-by-line.
4. Source conflicts on petitioner identity for IPR2018‑00351 (HTC per the Board's Paper 8 caption vs. DISH per Patexia) and on the disposition of IPR2018‑01440 (Adverse Judgment per Google's annotation vs. Terminated‑Settled per Patexia) are recorded, not resolved.
5. No 2026 appellate activity specifically directed at '718 was found (consistent with the earlier section). Because the patent expired – lifetime, adjusted expiration March 2, 2019, the practical stakes of this §103 analysis today are historical, portfolio-wide (the sibling Cheyer patents remain the live dispute), and defensive/opinion-of-counsel in nature — not an injunction or royalty exposure on '718 itself.
6. An obviousness conclusion is ultimately factual and depends on: the PHOSITA level (unfixed in this record), the constructions of "portable remote control device," "mobile information appliance," and "navigation query," the effective priority date (§6), and whether secondary considerations (long-felt need for couch-friendly VOD navigation; commercial success of voice assistants a decade later) are proffered and nexus-proven. None of those were adjudicated for the '718 claims. My assessment is a reasoned projection, not a holding.

Principal sources used for this section (all fetched 2026‑09‑28 unless noted):

• Full text, claims, classification, Cited By, priority and family data: https://patents.google.com/patent/US6757718/en
• Sibling-patent IPR exhibit set (Exs. 1013–1015): https://portal.unifiedpatents.com/ptab/case/IPR2018-00358
• Board Termination Decision (Settlement) captioned for IPR2018‑00351 (Patent 6,757,718 B1): https://www.docketalarm.com/cases/PTAB/IPR2018-00351/Inter_Partes_Review_of_U.S._Pat._6757718/docs/04-23-2018-Board/Termination_Decision_Document-8-Termination_Decision_Document.pdf
• IPR2018‑00351 docket summary (petitioner listed as DISH; challenged claims enumerated): https://services.patexia.com/lawsuits/DISH-Network-Corporation-v-IPA-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Inc-which-is-a-wholly-owned-subsidiary-of-Quarterhill-Inc-id-[110748](/patent/110748)
• IPR2018‑01440 (Microsoft v. IPA; claims 1–27; decision Mar. 6, 2019): https://services.patexia.com/lawsuits/Microsoft-Corporation-v-IPA-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Technologies-Inc-which-is-a-wholly-owned-subsidiary-of-Wi-LAN-Inc-which-is-a-wholly-owned-subsidiary-of-Quarterhill-Inc-id-[117101](/patent/117101)
• Ex. 1002, Declaration of Dr. Henry Lieberman (Aug. 9, 2018): https://www.docketalarm.com/cases/PTAB/IPR2018-01440/Inter_Partes_Review_of_U.S._Pat._6757718/docs/08-09-2018-Petitioner/Exhibit-1002-2-Expert_Declaration_of_Dr_Henry_Lieberman.pdf
• IPA v. Amazon, No. 1:16‑cv‑01266, D.I. 28 (asserted patents '061/'021/'718): https://storage.courtlistener.com/recap/gov.uscourts.ded.61001/gov.uscourts.ded.61001.28.0.pdf