Oracle Patent US12608586B2 - PSDP Auto Composing
Summary
Oracle International Corporation has been granted US Patent 12,608,586 B2 for a parameter sharing decoder pair (PSDP) architecture for auto composing using transformer-based language models. The patent, filed October 30, 2020, claims 14 claims and reduces model parameters while maintaining the ability to generate understandable compositions. The CPC classifications span neural networks (G06N 3/04–3/084) and natural language processing (G06F 40/56).
“Techniques for auto composing using a transformer-based language model having a parameter sharing decoder pair (PSDP) that reduces a number of parameters of the model and maintains the capability of generating understandable and reasonable compositions.”
About this source
GovPing monitors USPTO Patent Grants - AI & Computing (G06N) for new telecom & technology regulatory changes. Every update since tracking began is archived, classified, and available as free RSS or email alerts — 32 changes logged to date.
What changed
Oracle International Corporation has received US Patent 12,608,586 B2, titled 'Parameter sharing decoder pair for auto composing.' The invention relates to transformer-based language models using a PSDP architecture where a first decoder's parameters are shared across all N layers and a second decoder's parameters are shared across all N layers, with the two decoders having different parameters. The model predicts sequence elements from a full encoder sequence to generate an output sequence distinct from the input.
Affected parties include technology companies developing transformer-based language models and auto-composition systems. The patent grants Oracle exclusive rights to the PSDP technique, which may be relevant for competitors designing similar parameter-efficient transformer architectures. Organizations engaged in language model development should review this patent to assess whether their products or research implicate the claimed methods.
Archived snapshot
Apr 23, 2026GovPing captured this document from the original source. If the source has since changed or been removed, this is the text as it existed at that time.
Parameter sharing decoder pair for auto composing
Grant US12608586B2 Kind: B2 Apr 21, 2026
Assignee
Oracle International Corporation
Inventors
Xu Zhao
Abstract
Techniques for auto composing using a transformer-based language model having a parameter sharing decoder pair (PSDP) that reduces a number of parameters of the model and maintains the capability of generating understandable and reasonable compositions. In one particular aspect, a method is provided that includes obtaining a full encoder sequence, and inputting the full encoder sequence into a transformer model having a PSDP. The PSDP includes: a first decoder having parameters that are shared across all N layers of the first decoder; and a second decoder having parameters that are shared across all N layers of the first decoder. The parameters of the first decoder are different from the parameters of the second decoder. The method further includes using the transformer model to predict sequence elements based on the full encoder sequence, generate an output sequence comprising the sequence elements, and output an output sequence different from the full encoder sequence.
CPC Classifications
G06N 3/04 G06N 3/063 G06N 3/044 G06N 3/045 G06N 3/084 G06F 40/56
Filing Date
2020-10-30
Application No.
17086271
Claims
14
Mentioned entities
Parties
Related changes
Get daily alerts for USPTO Patent Grants - AI & Computing (G06N)
Daily digest delivered to your inbox.
Free. Unsubscribe anytime.
Source
About this page
Every important government, regulator, and court update from around the world. One place. Real-time. Free. Our mission
Source document text, dates, docket IDs, and authority are extracted directly from USPTO.
The summary, classification, recommended actions, deadlines, and penalty information are AI-generated from the original text and may contain errors. Always verify against the source document.
Classification
Who this affects
Taxonomy
Browse Categories
Get alerts for this source
We'll email you when USPTO Patent Grants - AI & Computing (G06N) publishes new changes.
Subscribed!
Optional. Filters your digest to exactly the updates that matter to you.