Storing entries in and retrieving information from an episodic object memory

Patent metadata

Field	Value
Publication number	US-2024256958-A1
Application number	US-202318204725-A
Country	US
Kind code	A1
Filing date	Jun 1, 2023
Priority date	Jan 30, 2023
Publication date	Aug 1, 2024
Grant date	—

Abstract

A method for identifying and storing a landmark memory in an episodic object memory is provided that includes receiving one or more content items. The content items each have one or more content data. The one or more content data associated with the one or more content items may be provided and integrated into one or more embedding models that represent a growing set of episodic memories. Episodic memory relates to the ability to recall content from one's personal past, such as in the form of landmark memories, which may be filtered from a plurality of memories based on a degree of salience. The one or more landmark memories and references to related content items are inserted into an episodic object memory for later recall and use in the course of context and task at hand.

First claim

What is claimed is:

1 . A method for storing a landmark memory embedding in an episodic object memory, the method comprising:
receiving one or more content items, the content items each having one or more content data; providing the one or more content data associated with the one or more content items to one or more embedding models, wherein the one or more embedding models generate one or more embeddings; receiving, from one or more of the embedding models, a collection of embeddings, wherein each embedding of the collection of embeddings corresponds to at least one content data from a respective content item; determining that one or more embeddings of the collection of embeddings are landmark memory embeddings; inserting the one or more landmark memory embeddings into the episodic object memory, wherein the one or more landmark memory embeddings are associated with a reference to related data associated with the one or more landmark memory embeddings; and providing the episodic object memory.

2 . The method of claim 1 , wherein the insertion triggers a spatial storage operation to store a vector representation of the landmark memory embeddings, and wherein the vector representation is stored in at least one of an approximate nearest neighbor (ANN) tree, a k-d tree, or a multidimensional tree.

3 . The method of claim 1 , wherein the determining comprises ranking the one or more of the embeddings based on a dissimilarity to at least one of the other embeddings of the collection of embeddings, and wherein the inserting comprises storing an indication of the rankings corresponding to the landmark memory embeddings.

4 . The method of claim 1 , wherein the determining comprises:
obtaining a machine-learning model that was previously-trained to identify landmark memories; providing the collection of embeddings to the machine-learning model; receiving a score from the machine-learning model corresponding to a degree to which embeddings are landmark memory embeddings; comparing the score to a threshold to identify that one or more embeddings of the collection of embeddings are landmark memory embeddings.

5 . The method of claim 4 , further comprising, prior to the comparing:
receiving user-input corresponding to the threshold, wherein the threshold is a threshold of episodic memory.

6 . The method of claim 5 , wherein the user-input is received from a slider of a graphical user-interface.

7 . The method of claim 1 , wherein the content data are one or more of audio content data, visual content data, gaze content data, weather content data, news content data, calendar content data, email content data, or location content data.

8 . The method of claim 1 , wherein the episodic object memory is stored at a location that is different than a location of source data corresponding to the content items.

9 . The method of claim 1 , wherein the one or more landmark memory embeddings comprise a set of properties that define a schema.

10 . The method of claim 9 , wherein the set of properties comprise a summary of the landmark memory embeddings and the reference to related data.

11 . A method for retrieving landmark memories from an episodic object memory, the method comprising:
generating a user-interface; receiving, via the user-interface, an input, wherein the input corresponds to a degree of episodic memory, receiving, from the episodic memory, an indication of one or more landmark memories, based on the degree of episodic memory, wherein the episodic object memory is generated using content data that is embedded and ranked, based on semantic context.

12 . The method of claim 11 , wherein the embeddings generated based on the content data are ranked based on a dissimilarity of an embedding to at least one other embedding.

13 . The method of claim 11 , wherein the content data are one or more of audio content data, visual content data, gaze content data, weather content data, news content data, calendar content data, email content data, or location content data.

14 . The method of claim 11 , wherein the episodic object memory is stored at a location that is different than a location of source data corresponding to the content items.

15 . The method of claim 11 , wherein the input is a first input, wherein the degree is a first degree, wherein the indication is a first indication, and wherein the method further comprises:
receiving a second input, wherein the second input corresponds to a second degree of episodic memory; and receiving, from the episodic object memory, a second indication of one or more landmark memories corresponding to the second input, based on the second degree of episodic memory.

16 . The method of claim 15 , wherein the user-interface includes a slider and wherein the first and second inputs correspond to movements of the slider.

17 . A method for storing landmark memories in and retrieving landmark memories from an episodic object memory, the method comprising:
receiving one or more content items, the content items each having one or more content data; providing the one or more content data associated with the one or more content items to one or more embedding models, wherein the one or more embedding models generate one or more embeddings; receiving, from one or more of the embedding models, a collection of embeddings, wherein each embedding of the collection of embeddings corresponds to at least one content data from a respective content item; determining that one or more embeddings of the collection of embeddings are landmark memory embeddings; inserting the one or more landmark memory embeddings into the episodic object memory; and receiving an input, wherein the input corresponds to a degree of episodic memory, receiving, from the episodic object memory, an indication of one or more landmark memories, based on the degree of episodic memory.

18 . The method of claim 17 , wherein the insertion triggers a spatial storage operation to store a vector representation of the landmark memory embeddings, and wherein the vector representation is stored in at least one of an approximate nearest neighbor (ANN) tree, a k-d tree, or a multidimensional tree.

19 . The method of claim 17 , wherein the determining comprises weighting the one or more embeddings based on a dissimilarity to at least one of the other embeddings of the collection of embeddings.

20 . The method of claim 17 , wherein the determining comprises:
obtaining a machine-learning model that was previously-trained to identify landmark memories; providing the collection of embeddings to the machine-learning model; receiving a score from the machine-learning model corresponding to a degree to which embeddings are landmark memory embeddings; comparing the score to a threshold to identify that one or more embeddings of the collection embeddings are landmark memory embeddings.

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Horvitz Eric Joel
• Becker Nicholas

Classifications

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N20/00 Primary

  Machine learning · CPC title