Muvox LLC v. IBM Corp

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Muvox LLC (NM)
  • Defendant: IBM Corporation (NY)
  • Plaintiff’s Counsel: Rabicoff Law LLC
• Case Identification: 2:24-cv-00693, E.D. Tex., 08/21/2024
• Venue Allegations: Venue is alleged to be proper based on Defendant maintaining an established place of business within the Eastern District of Texas.
• Core Dispute: Plaintiff alleges that Defendant infringes a patent related to systems for automatically categorizing streaming music and creating personalized playlists based on computer-derived acoustic attributes.
• Technical Context: The technology concerns automated music analysis and playlist generation, a central feature in modern digital music streaming services that allows for scalable content curation.
• Key Procedural History: The asserted patent claims priority to a 2014 provisional application and is the result of a long chain of continuation applications. The complaint references, but does not include, claim charts that allegedly detail the basis for infringement.

Case Timeline

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 11,899,713 - Music streaming, playlist creation and streaming architecture

• Patent Identification: U.S. Patent No. 11,899,713, "Music streaming, playlist creation and streaming architecture," issued February 13, 2024.

The Invention Explained

• Problem Addressed: The patent’s background section states that music publishers, such as radio stations or record labels, often lack a practical way to stream their specific music catalogs directly to users and receive brand recognition, as large, third-party streaming services typically control the available content and user experience (ʼ713 Patent, col. 1:40-61).
• The Patented Solution: The invention proposes a system architecture where music is objectively categorized using computer-derived "rhythm, texture and pitch (RTP) scores" (’713 Patent, Abstract). These scores and associated metadata are stored in a universal database, which prevents the need to re-analyze a track if multiple publishers use it (’713 Patent, col. 2:16-24). Publisher-sponsored end-user applications can then access these scores via an API to create personalized, mood-based playlists and initiate streaming from a separate server, as illustrated in the system diagram of Figure 2 (’713 Patent, col. 2:24-34; FIG. 2).
• Technical Importance: This system aims to decentralize music curation by empowering individual publishers with an automated, scalable method for playlist generation, allowing them to deliver a branded experience to end-users (’713 Patent, col. 1:55-61).

Key Claims at a Glance

• The complaint does not specify which claims are asserted, referring only to "Exemplary '713 Patent Claims" in an unprovided exhibit (Compl. ¶11, ¶13). Claim 1 is the first independent method claim.
• Independent Claim 1 recites a method for selecting a song with the following essential elements:
  • selecting the song based on a computer-derived comparison between a representation of the song and known similarities in representations of other songs;
  • wherein the known similarities are based on a "human-trained machine" that was trained using representations of other songs;
  • wherein the song's representation is based on its isolated and identified "frequency characteristics";
  • wherein the representations of the other songs are based on "human listening" to identify their frequency characteristics;
  • wherein the frequency characteristics of the songs correspond to one or more "moods";
  • and wherein the selection is based on the similarity between the moods of the songs.
• The complaint does not explicitly reserve the right to assert dependent claims, but alleges infringement of "one or more claims" (Compl. ¶11).

III. The Accused Instrumentality

Product Identification

The complaint does not name any specific accused products, referring to them only as "Exemplary Defendant Products" (Compl. ¶11). It states that these products are identified in "Exhibit 2," which was not filed with the complaint (Compl. ¶13-14).

Functionality and Market Context

The complaint does not provide sufficient detail for analysis of the accused instrumentality's functionality. It offers only the conclusory allegation that the "Exemplary Defendant Products practice the technology claimed by the '713 Patent" (Compl. ¶13).

IV. Analysis of Infringement Allegations

The complaint references claim charts in an unprovided "Exhibit 2" (Compl. ¶13, ¶14). As the exhibit is not available, a claim chart summary cannot be constructed. The infringement theory is stated in general terms, alleging that Defendant's unidentified products "satisfy all elements of the Exemplary '713 Patent Claims" (Compl. ¶13).

No probative visual evidence provided in complaint.

Identified Points of Contention

• Technical Questions: A central evidentiary question will be what specific functionality in the accused IBM products performs the song selection method of Claim 1. The plaintiff will need to present evidence that the accused system utilizes a "human-trained machine" that was trained based on "human listening" to categorize songs by "moods" derived from "frequency characteristics," as the claim language requires (’713 Patent, col. 18:31-50).
• Scope Questions: Claim 1 requires a selection based on a comparison to "known similarities in representations of other songs" (’713 Patent, col. 18:33-35). A potential dispute may arise over whether the accused products perform a comparison against a pre-existing, human-curated and machine-learned database of song attributes, or if they perform a fundamentally different type of real-time analysis or categorization that falls outside the claim's scope.

V. Key Claim Terms for Construction

"human-trained machine"

• Context and Importance: This term appears in independent Claim 1 and is fundamental to how the patented system learns to categorize music. The infringement analysis will depend on whether the accused IBM system's architecture can be characterized as a "human-trained machine." Practitioners may focus on this term because its construction will define the scope of machine learning systems covered by the patent (’713 Patent, col. 18:37).
• Intrinsic Evidence for Interpretation:
  • Evidence for a Broader Interpretation: The specification suggests this could cover various machine learning models, referencing a "neural network" that may be "trained based on an ontology of audio event classes and a large-scale collection of human-labeled sound clips" (’713 Patent, col. 4:53-57). This language could support an interpretation covering a range of supervised learning models.
  • Evidence for a Narrower Interpretation: The patent also describes a more specific process where humans first listen to tracks to develop "RTP scores," and a neural network is then trained on those pre-determined scores (’713 Patent, col. 4:35-54; FIG. 1). A defendant may argue the term should be limited to this specific multi-step process of human scoring followed by machine training, rather than any system trained more directly on human-labeled data.

"moods"

• Context and Importance: The song selection process in Claim 1 is explicitly based on a similarity of "moods" (’713 Patent, col. 18:43-50). The definition of this term is critical, as it will determine whether an accused system that uses different categorical labels (e.g., genre, energy level, situational context) infringes.
• Intrinsic Evidence for Interpretation:
  • Evidence for a Broader Interpretation: The specification provides examples such as "happy, excited, manic, peaceful, sad and cautious" but also states that categories could be "colors, locations, etc., including completely arbitrary or made-up words," suggesting the term is not limited to conventional emotional states (’713 Patent, col. 6:18-22, col. 6:31-33).
  • Evidence for a Narrower Interpretation: The patent repeatedly links the derived RTP scores to specific affective states mapped on an intensity spectrum (e.g., low-intensity sad vs. high-intensity sad) (’713 Patent, col. 6:40-53). A defendant could argue that "moods" should be construed to require an emotional or affective classification, not merely an objective one like "high-tempo" or "acoustic."

VI. Other Allegations

• Indirect Infringement: The complaint does not contain factual allegations to support claims for either induced or contributory infringement. The sole count is for "Direct Infringement" (Compl. ¶11).
• Willful Infringement: The complaint does not plead any facts to support a claim of willfulness, such as alleging that Defendant had pre-suit knowledge of the ’713 Patent. The prayer for relief includes a request that the case be declared "exceptional" under 35 U.S.C. § 285, but the factual basis for this request is absent from the pleadings (Compl. p. 4, ¶E.i).

VII. Analyst’s Conclusion: Key Questions for the Case

• A primary evidentiary question for the plaintiff will be one of technical mapping: given the complaint's lack of detail, what specific features of the unidentified IBM products perform the multi-step method of Claim 1, and what evidence demonstrates that IBM employs a "human-trained machine" to select songs based on "moods" derived from frequency analysis?
• A core legal issue will be one of claim scope: can the term "human-trained machine," which the patent links to a specific process of training a neural network on human-derived RTP scores, be construed broadly enough to read on the potentially different machine learning architecture used in IBM’s accused products?