DCT
1:22-cv-11467
Constructorio Corp v. Hitel Tech LLC
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: The Life is Good Company (Massachusetts) and Constructor.io Corporation (Delaware)
- Defendant: Hitel Technologies LLC (Texas)
- Plaintiff’s Counsel: Wilmer Cutler Pickering Hale and Dorr LLP
- Case Identification: 1:22-cv-11467, D. Mass., 09/12/2022
- Venue Allegations: Plaintiffs assert that venue is proper in the District of Massachusetts because Defendant Hitel directed patent assertion and licensing activities, which form the basis of the suit, at Plaintiff The Life is Good Company, headquartered in the district.
- Core Dispute: Plaintiffs seek a declaratory judgment that their e-commerce website search functionality, provided by Constructor.io, does not infringe Defendant's patent related to dynamic learning in search and navigation systems.
- Technical Context: The technology at issue involves methods for search systems to learn from user behavior, particularly in response to misspelled or otherwise "unknown" search queries, to improve future results in an e-commerce setting.
- Key Procedural History: The complaint states this action follows pre-suit correspondence in which Hitel, a patent assertion entity, accused The Life is Good Company of infringing the patent-in-suit, provided a claim chart, and demanded a license fee. The complaint also notes that Hitel has filed numerous lawsuits asserting the same patent against other companies.
Case Timeline
| Date | Event |
|---|---|
| 2005-02-25 | ’617 Patent Priority Date |
| 2010-03-30 | ’617 Patent Issue Date |
| 2022-06-29 | Hitel sends notice letter and claim chart to LIG |
| 2022-07-13 | Hitel sends follow-up email to LIG |
| 2022-08-10 | Teleconference between Hitel and LIG representatives |
| 2022-08-10 | Hitel provides draft license agreement to LIG |
| 2022-09-12 | Complaint for Declaratory Judgment filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 7,689,617 - "Dynamic Learning for Navigation Systems"
- Patent Identification: U.S. Patent No. 7,689,617, "Dynamic Learning for Navigation Systems," issued March 30, 2010. (’617 Patent).
The Invention Explained
- Problem Addressed: The patent describes a limitation in conventional search and navigation systems that rely on pre-defined keywords. These systems may fail when a user's query contains a word that is not a known keyword or synonym, such as searching for "vintage car" when the system only recognizes "antique automobile." (’617 Patent, col. 1:65-2:7). The patent also notes that prior methods for learning new word associations had limitations. (’617 Patent, col. 2:16-20).
- The Patented Solution: The invention proposes a method for a system to learn from user interaction. When the system receives an "unknown word," it allows the user to navigate to a relevant node or document. The system then creates an association between the initial "unknown word" and keywords found within the document selected by the user. (’617 Patent, Abstract; col. 3:1-14). This learned association is then used to provide more relevant results for future queries containing that same word.
- Technical Importance: This approach was designed to make search systems more adaptive and intelligent by enabling them to learn new terms and user intent directly from user navigational behavior, rather than being restricted to a static dictionary of keywords. (’617 Patent, col. 2:56-3:14).
Key Claims at a Glance
- The complaint identifies independent claims 1, 11, and 19 as having been asserted by Hitel (Compl. ¶18).
- Independent Claim 1 requires:
- In a system with user navigable nodes, documents, and keywords, a method comprising:
- receiving an input from a user;
- determining that the input contains an unknown word;
- subsequent to the determining, navigating through the system to a current node based upon at least one additional input from the user that was provided while at the user navigable node;
- presenting at least one response to the user based upon the documents attached to the current node; and
- following presentation of the at least one response, learning one or more associations between the unknown word and one or more keywords from the documents attached to the current node.
- The complaint states that Hitel has asserted "multiple claims" and notes that the remaining claims are also not infringed (Compl. ¶¶ 7, 44).
III. The Accused Instrumentality
Product Identification
- The search and product discovery functionality on The Life is Good Company's website (www.lifeisgood.com), which is powered by software from Constructor.io Corporation (Compl. ¶¶ 3, 8).
Functionality and Market Context
- The accused instrumentality is the e-commerce search engine that customers use to find products on the LIG website (Compl. ¶8). The complaint alleges that this technology uses "advanced natural language processing and machine learning algorithms" and "sophisticated spelling correction software" to interpret user queries (Compl. ¶¶ 3, 35). Plaintiffs contend this technology automatically corrects misspellings and provides relevant results based on the correctly spelled term, without performing the user-navigation-based learning steps described in the ’617 patent (Compl. ¶37).
IV. Analysis of Infringement Allegations
The complaint seeks a declaratory judgment of non-infringement, outlining its defense against Hitel's allegations. The following chart summarizes Hitel's alleged infringement theory for claim 1 as described and disputed in the complaint.
’617 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality (as described and disputed in the Complaint) | Complaint Citation | Patent Citation |
|---|---|---|---|
| receiving an input from a user; | A user enters a search query, such as the misspelled term "tablewaer," into the search field on the LIG Website. This is depicted in a screenshot from Hitel's claim chart (Compl. p. 12). | ¶33 | col. 13:51-52 |
| determining that the input contains an unknown word; | The system allegedly identifies the misspelled search term "tablewaer" as an "unknown word." | ¶33 | col. 13:53-54 |
| subsequent to the determining, navigating through the system to a current node from the user navigable node based upon at least one additional input from the user... | Hitel allegedly contends that a user's selection of a product from the initial search results, such as a "USA Home Plate" t-shirt, constitutes the required navigation based on an additional user input. | ¶35 | col. 13:55-59 |
| presenting at least one response to the user based upon the documents attached to the current node...; and | The system displays the product page for the t-shirt selected by the user. | ¶35 | col. 13:60-62 |
| following presentation of the at least one response, learning one or more associations between the unknown word and one or more keywords from the documents attached to the current node... | Hitel allegedly contends that the system learns to associate the "unknown word" "tablewaer" with a keyword from the selected t-shirt product. Plaintiffs dispute this by showing a later search did not return the previously selected shirt (Compl. p. 15). | ¶35, 41 | col. 13:62-67 |
- Identified Points of Contention:
- Scope Questions: A primary dispute concerns whether a misspelled term like "tablewaer," which the complaint alleges is recognized and corrected by modern spelling-correction software, qualifies as an "unknown word" under the claim (Compl. ¶35). Plaintiffs argue their system recognizes it as a misspelling of "tableware," and thus it is not "unknown" in the manner contemplated by the patent (Compl. ¶34).
- Technical Questions: A key factual question is whether the accused system performs the claimed "learning" step at all. The complaint directly challenges this by presenting evidence from a subsequent search for "tablewaer" that allegedly shows no association was learned from a prior user selection, as the previously selected product was not returned in the results (Compl. ¶35). The complaint includes a screenshot from August 21, 2022, showing that a search for "tablewaer" yielded tableware products but not the "USA Home Plate" t-shirt allegedly selected in Hitel's infringement example (Compl. p. 15). This raises the question of whether there is evidence the accused system creates new associations based on user navigation, or if it simply relies on pre-existing NLP and spell-check models.
V. Key Claim Terms for Construction
The Term: "unknown word"
- Context and Importance: The definition of this term is central, as Hitel's infringement theory appears to depend on a misspelled search query being classified as an "unknown word" (Compl. ¶33). Practitioners may focus on this term because its construction could determine whether modern spell-check and NLP systems, which recognize and correct misspellings, fall within the scope of the claims.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification defines an "unknown word" as "a word that is neither a keyword or an existing synonym" (’617 Patent, col. 2:65-67). A party could argue that a misspelled word literally fits this definition, as it is not found in a pre-defined list of correct keywords or synonyms.
- Evidence for a Narrower Interpretation: The patent's overall context describes a system learning the meaning of a word it genuinely does not comprehend. Plaintiffs argue that their system, by recognizing "tablewaer" as a misspelling of "tableware," does not treat the term as "unknown" but rather as a correctable error, and therefore operates differently from the patented method (Compl. ¶34-35).
The Term: "learning one or more associations between the unknown word and one or more keywords"
- Context and Importance: This is the core functional step of the invention. The dispute centers on whether the accused system actually performs this specific type of learning. The plaintiffs allege their system does not create new associations post-navigation but instead uses its own independently developed technologies (Compl. ¶¶ 37, 41).
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: A party might argue that any dynamic adjustment to search algorithms based on user-selection data, even updating statistical weights in an ML model, could constitute "learning... associations."
- Evidence for a Narrower Interpretation: The claim language recites learning associations "between the unknown word and one or more keywords from the documents attached to the current node" (’617 Patent, col. 13:62-67). This may be construed to require a specific mechanism where a direct link is formed between the verbatim unknown word and specific keywords from a user-selected document, a process the plaintiffs allege their system does not perform (Compl. ¶37).
VI. Other Allegations
- Indirect Infringement: The complaint states that Hitel's pre-suit notice letter accused LIG of infringement by, among other things, "instructing [Life is Good's] customers to use, [LIG's Website]" (Compl. ¶17). This provides the factual basis for an allegation of induced infringement, which Hitel would need to prove by showing LIG specifically intended its customers to infringe the patent.
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: Can the term "unknown word", as used in the patent, be construed to cover a misspelled search term that a modern NLP system recognizes and automatically corrects? The resolution of this claim construction issue may significantly impact the infringement analysis.
- A key evidentiary question will be one of technical operation: Does the accused Constructor.io search technology actually perform the specific, multi-step "learning" process recited in the claims—creating new associations between a verbatim "unknown word" and keywords from a user-navigated-to document—or does it, as plaintiffs allege, rely on a fundamentally different architecture of advanced spell-correction and NLP that operates without the claimed learning mechanism?