DCT
1:23-cv-00190
Parus Holdings Inc v. Amazon.com Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Parus Holdings Inc. (Delaware)
- Defendant: Amazon.com, Inc. (Delaware)
- Plaintiff’s Counsel: Farnan LLP; McKool Smith, P.C.
- Case Identification: 1:23-cv-00190, D. Del., 02/17/2023
- Venue Allegations: Plaintiff alleges venue is proper in the District of Delaware because Amazon has transacted business and committed acts of direct patent infringement in the district.
- Core Dispute: Plaintiff alleges that Defendant’s Alexa devices and associated cloud services infringe three expired patents related to voice-based systems for retrieving user-specified information from online sources.
- Technical Context: The technology at issue falls within the domain of voice-activated digital assistants, a significant market segment involving the interaction between user-facing hardware and cloud-based data processing and information retrieval.
- Key Procedural History: The complaint states that all three patents-in-suit have expired. This procedural posture limits Plaintiff's potential remedy to monetary damages for past infringement occurring within the six-year statutory period preceding the complaint's filing.
Case Timeline
| Date | Event |
|---|---|
| 2000-02-04 | Priority Date for ’190, ’992, and ’981 Patents |
| 2009-04-07 | U.S. Patent No. 7,516,190 Issues |
| 2016-06-28 | U.S. Patent No. 9,377,992 Issues |
| 2017-01-01 | Alleged Infringing Activity by Amazon Alexa Begins (approx.) |
| 2019-06-11 | U.S. Patent No. 10,320,981 Issues |
| 2021-02-06 | U.S. Patent No. 10,320,981 Expires |
| 2022-05-10 | U.S. Patent No. 7,516,190 Expires |
| 2022-11-16 | U.S. Patent No. 9,377,992 Expires |
| 2023-02-17 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 7,516,190 - “Personal Voice-Based Information Retrieval System”
- Patent Identification: U.S. Patent No. 7,516,190, “Personal Voice-Based Information Retrieval System,” issued April 7, 2009.
The Invention Explained
- Problem Addressed: The patent’s background section identifies the process of accessing specific information on the Internet as "time-consuming and complicated," often requiring a user with a computer to "wade through vast amounts of information" on a website. The patent sought to solve the need for a simplified method to access frequently needed online information using only a telephone and simple speech commands (’190 Patent, col. 1:18-47).
- The Patented Solution: The invention describes a two-stage process. First, a user operates a "clipping client" on a computer to select a specific portion of a website (e.g., weather data) and assigns a custom voice command, or "recognition grammar," to it. Second, the user can later call into a "voice browsing system" from a telephone, speak the pre-defined command, and have the system retrieve the current information from that specific part of the website and read it back using a speech synthesizer (’190 Patent, col. 2:5-27; Abstract). The system architecture, showing the relationship between the user, clipping client, and voice browsing system, is illustrated in the patent’s Figure 1 (’190 Patent, Fig. 1).
- Technical Importance: This technology describes a method for creating personalized voice shortcuts to dynamic web content, abstracting away the complexity of graphical web browsing for users seeking specific, recurring pieces of information (’190 Patent, col. 2:5-14).
Key Claims at a Glance
- The complaint asserts independent Claim 1 (Compl. ¶19).
- The essential elements of Claim 1, a method claim, include:
- Providing a computer with a speech processor connected to the internet and a phone.
- Providing a URL to the computer for a pre-selected web site.
- Using the computer to designate a pre-defined portion of that web site.
- Identifying a "named object" associated with the information.
- Generating a "regular expression" based on the pre-defined portion and named object.
- Providing a speech command corresponding to the regular expression.
- Converting the speech command to a digital command and assigning the regular expression to it.
- Upon receiving a subsequent audio speech command, retrieving the information from the pre-defined portion of the website corresponding to the regular expression.
- Converting the retrieved information into an audio message and forwarding it to a user.
- The complaint notes that infringement of other claims may be disclosed later in the proceedings (Compl. ¶20).
U.S. Patent No. 9,377,992 - “Personal Voice-Based Information Retrieval System”
- Patent Identification: U.S. Patent No. 9,377,992, “Personal Voice-Based Information Retrieval System,” issued June 28, 2016.
The Invention Explained
- Problem Addressed: As a continuation of the same patent family, the ’992 Patent addresses the same problem of simplifying and personalizing voice access to specific information on computer networks like the Internet (’992 Patent, col. 1:22-50).
- The Patented Solution: The ’992 Patent claims a method for retrieving information where a speech command is received at a speech recognition engine. The system selects an "information source retrieval instruction" corresponding to the command's "recognition grammar." A key part of the solution involves accessing a portion of an information source (e.g., a webpage) "using a clipping client to separate the portion of the information from other information," where the clipping client generates a "content descriptor file" indicating the data's location within the source. The system then retrieves only that portion, converts it to an audio message, and transmits it to the user's device (’992 Patent, Abstract; col. 2:1-26).
- Technical Importance: This patent further details the technical mechanism for user-defined voice retrieval by explicitly claiming the use of a "clipping client" and a "content descriptor file" to precisely target and extract information from sources that are "periodically updated" (’992 Patent, Claim 1).
Key Claims at a Glance
- The complaint asserts independent Claim 1 (Compl. ¶59).
- The essential elements of Claim 1, a method claim, include:
- Receiving a speech command at a speech recognition engine, where the information source is periodically updated.
- Selecting an information source retrieval instruction corresponding to the recognition grammar for the speech command.
- Accessing a portion of the information source using a "clipping client" to separate it from other information.
- The clipping client generates a "content descriptor file" describing the content and its location.
- Retrieving only the designated portion of information.
- Converting the retrieved information into an audio message via a speech synthesis engine.
- Transmitting the audio message to the user's electronic communication device.
- The complaint states that additional claims may be asserted later (Compl. ¶60).
U.S. Patent No. 10,320,981 - “Personal Voice-Based Information Retrieval System”
- Multi-Patent Capsule: U.S. Patent No. 10,320,981
- Patent Identification: U.S. Patent No. 10,320,981, “Personal Voice-Based Information Retrieval System,” issued June 11, 2019 (Compl. ¶12).
- Technology Synopsis: This patent, from the same family, claims a method where a speech command is received from a voice-enabled device over a network. The system uses a "user-defined search" and selects a corresponding "speech-recognition grammar" to access a website. A "content extractor" within a web-browsing server uses a "content-descriptor file" to separate a relevant portion of information from the webpage, which is then converted to audio and transmitted to the user (’981 Patent, Abstract).
- Asserted Claims: The complaint asserts independent Claim 1 (Compl. ¶81).
- Accused Features: The complaint alleges that the Amazon Alexa system, including the Alexa Voice Service, Alexa Skills, and AWS Lambda, performs the claimed method by receiving voice commands from Echo devices, using skills to execute user-defined searches of web sources, extracting specific information, and returning an audio response (Compl. ¶84-88).
III. The Accused Instrumentality
Product Identification
- The accused instrumentalities are Amazon's Alexa-enabled smart speaker and display products (e.g., Echo, Echo Dot, Echo Show, Echo Auto) and the associated cloud-based "Alexa" voice service (Compl. ¶4, ¶14).
Functionality and Market Context
- The complaint describes the accused system as a cloud-based service where a user's speech command is captured by a device and streamed to the Alexa Voice Service (AVS) for processing (Compl. ¶22, ¶24). An "Alexa Architecture Overview" diagram provided in the complaint illustrates how AVS identifies a user's intent, sends a request to a corresponding "Skill" (often hosted on AWS Lambda), receives a response from the Skill, converts the returned text to speech, and streams the audio back to the user's device (Compl. p. 9). The complaint alleges that third-party and Amazon-developed "Skills" allow Alexa to retrieve specific information from pre-selected web sites or data feeds, such as tide information from noaa.gov (Compl. ¶26, ¶48).
IV. Analysis of Infringement Allegations
7,516,190 Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| (c) using said computer to designate a pre-defined portion of the pre-selected web site which contains the information to be retrieved | A skill developer provides code and instructions to the Amazon Alexa system that designates a pre-defined portion of a web site to be accessed by the skill, such as the Tide Pool skill accessing tide data from http://tidesandcurrents.noaa.gov/. | ¶27, ¶28 | col. 3:11-18 |
| (e) using said computer to generate a regular expression based on said pre-defined portion...said regular expression is a text string used for describing a search pattern | On information and belief, a skill developer provides code to the Alexa system that generates and uses a regular expression associated with the content. The complaint alleges that "sample utterances" provided by a developer correspond to the regular expression. | ¶31, ¶32, ¶34 | col. 4:44-52 |
| (g) said speech processor converting said speech command to a digital-form command | Amazon's speech processor, AVS, converts a user's speech command (e.g., "Alexa, get high tide for Seattle") into a digital-form command, which the complaint alleges is an "intent" provided to the skill. | ¶35, ¶36 | col. 5:28-35 |
| (m) said computer retrieving the information from the pre-defined portion of the pre-selected web site...when the requested information is found | The Amazon Alexa system, via the Tide Pool skill, accesses and retrieves information from a pre-defined portion of the noaa.gov web page that corresponds to the user's request. A screenshot illustrates the user query and the system's specific, retrieved response (Compl. p. 22). | ¶46, ¶47, ¶48 | col. 6:28-36 |
| (p) said speech processor converting said retrieved information into an audio message | The Alexa Voice Service includes a speech-synthesis engine that converts the returned text from the skill into speech and streams it to the user's device. An architectural diagram shows this text-to-speech conversion step (Compl. p. 25). | ¶52, ¶53 | col. 2:24-27 |
Identified Points of Contention
- Scope Questions: A primary question may be whether the process of a developer creating an Alexa "Skill" using the Alexa Skills Kit meets the claim limitation of a "user" operating a "clipping client" to "designate a pre-defined portion" of a website. The patent specification appears to describe an end-user performing this setup directly (’190 Patent, col. 3:2-10).
- Technical Questions: The complaint alleges on "information and belief" that the Alexa system generates and uses a "regular expression" as claimed (Compl. ¶32). A potential point of contention is whether Alexa's system of mapping user speech to "intents" and "slots"—a natural language understanding (NLU) model—is technically equivalent to the claimed "regular expression... used for describing a search pattern" (’190 Patent, Claim 1(e)).
9,377,992 Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| receiving a speech command from the user via the electronic communication device at a speech recognition engine coupled to a media server...wherein the information source is periodically updated | Amazon's AVS (speech recognition engine) and AWS Lambda (media server) receive a speech command from an Echo device. The information source, such as the noaa.gov tide website, is periodically updated with new tide information. | ¶63, ¶64, ¶68 | col. 5:28-40 |
| selecting, by the media server, at least one appropriate information source retrieval instruction corresponding to the recognition grammar established for the speech command | The Alexa Voice Service selects a retrieval instruction (an API call via a Skill) corresponding to the recognized user command (the recognition grammar). The complaint points to "sample utterances" as defining the grammar for a skill (Compl. p. 35). | ¶71, ¶72 | col. 4:53-60 |
| accessing, by a web browsing server...by using a clipping client to separate the portion of the information...wherein the clipping client generates a content descriptor file | The complaint alleges that an Alexa Skill, such as the Tide Pool skill, accesses a specific portion of the noaa.gov website, and that this functionality is performed by the equivalent of a "clipping client" that generates a "content descriptor file." | ¶73, ¶74 | col. 3:2-48 |
| converting the information retrieved...into an audio message by a speech synthesis engine...and transmitting said audio message to the electronic communication device | AVS includes a speech-synthesis engine that converts text retrieved by the Skill into an audio message (speech) and transmits it to the user's Echo device. A "User Interaction Flow" diagram illustrates this process (Compl. p. 47). | ¶75, ¶76, ¶77, ¶78 | col. 5:58-65 |
Identified Points of Contention
- Scope Questions: It raises the question of whether the combination of an Alexa Skill and its backend service (e.g., AWS Lambda) can be properly characterized as a "web browsing server" as that term is used in the patent.
- Technical Questions: The complaint does not provide direct evidence of a "clipping client" or a "content descriptor file" existing as distinct components within the Alexa architecture. The infringement theory appears to rest on an argument of functional equivalence, which will likely be a central point of technical dispute.
V. Key Claim Terms for Construction
The Term: "clipping client" (’992 Patent, Claim 1)
Context and Importance: This term is critical because the patent specification describes a specific tool for a user to define the information to be retrieved. The complaint's infringement theory maps this developer-facing "Alexa Skills Kit" and the resulting Skill to the patent's "clipping client." The construction of this term may determine whether a developer tool for creating an application is equivalent to the user-facing tool for creating a data shortcut as described in the patent.
Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification states the clipping client "allows a user... to create a set of instructions" (’992 Patent, col. 3:5-7), language that could be argued to encompass any software tool, including a developer kit, used to define information retrieval rules.
- Evidence for a Narrower Interpretation: The specification describes the client's operation in detail, where a user "selects the information from the displayed web page" by "highlighting and copying the URL address." (’992 Patent, col. 3:21-26). This suggests a specific graphical user interface for an end-user to visually select content, which may be narrower than a developer writing code or configuring an interaction model in a console.
The Term: "regular expression" (’190 Patent, Claim 1(e))
Context and Importance: Practitioners may focus on this term because the infringement allegation equates Alexa's complex NLU system of "intents" and "sample utterances" with a "regular expression" (Compl. ¶32, ¶34). The technical definition of a regular expression is a well-established term of art in computer science, and the viability of the infringement case may hinge on whether Alexa's technology falls within its scope.
Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The claim itself provides a definition: "wherein said regular expression is a text string used for describing a search pattern" (’190 Patent, col. 6:54-56). Plaintiff may argue that a developer-provided list of sample utterances constitutes a "text string" that describes the "search pattern" for invoking the correct skill.
- Evidence for a Narrower Interpretation: The patent specification includes an example "content descriptor file" containing a literal regular expression string:
WEB SERVICES: (.+) Forecast FOUR-DAY FORECAST...(’190 Patent, col. 4, Table 1). This embodiment could be used to argue that the term was intended to have its specific, technical meaning in the art, rather than covering broader NLU models.
VI. Other Allegations
- Indirect Infringement: The complaint makes a general allegation of indirect infringement under 35 U.S.C. § 271 (Compl. ¶15) but does not provide sufficient detail for analysis of the specific elements of inducement or contributory infringement, such as allegations of specific instructions to users or the provision of components with no substantial non-infringing use.
- Willful Infringement: The complaint alleges that "Amazon has had knowledge of the patents-in-suit since prior to the date of this Complaint" (Compl. ¶16). This is a conclusory allegation of pre-suit knowledge that does not specify the basis for such knowledge (e.g., a prior notice letter, citation in other litigation, or industry awareness).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: can the patent's concept of an end-user operating a "clipping client" to visually select and save a portion of a webpage be construed to cover the modern architecture of a software developer using a development kit (the Alexa Skills Kit) to program a voice application ("Skill") that interacts with a web-based API?
- A key evidentiary question will be one of technical equivalence: does the accused Alexa platform's natural language understanding (NLU) system, which maps spoken phrases to developer-defined "intents," function in the same way to achieve the same result as the claimed method of generating and matching a "regular expression"?
- As all patents-in-suit are expired, the case will also center on a question of damages apportionment: what is the reasonable royalty for the alleged use of the patented technology, and how can its incremental contribution to the value of the vast and multifaceted Alexa ecosystem during the relevant time period be calculated and proven?