US 11,694,165 B2
Key-value memory network for predicting time-series metrics of target entities
Ayush Chauhan, Bangalore (IN); Shiv Kumar Saini, Bangalore (IN); Parth Gupta, Roorkee (IN); Archiki Prasad, Mumbai (IN); Amireddy Prashanth Reddy, Nalgonda (IN); and Ritwick Chaudhry, Chandigarh (IN)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Oct. 5, 2022, as Appl. No. 17/960,585.
Application 17/960,585 is a division of application No. 16/868,942, filed on May 7, 2020, granted, now 11,501,107.
Prior Publication US 2023/0031050 A1, Feb. 2, 2023
Int. Cl. G06F 18/214 (2023.01); G06N 3/063 (2023.01); G06F 18/24 (2023.01); G11C 16/14 (2006.01); G06Q 10/109 (2023.01); G06F 7/544 (2006.01)
CPC G06F 18/214 (2023.01) [G06F 18/24 (2023.01); G06N 3/063 (2013.01); G06F 7/5443 (2013.01); G06Q 10/109 (2013.01); G11C 16/14 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
processing hardware; and
a non-transitory computer-readable medium communicatively coupled to the processing hardware and implementing a key-value memory network comprising:
a key matrix with key vectors that are learned from training static feature data and training time-series feature data;
a value matrix with value vectors representing time-series trends;
an input layer configured to receive, for a target entity, input data comprising a concatenation of static feature data of the target entity, time-specific feature data, and time-series feature data for the target entity;
an entity-embedding layer configured to generate an input vector from the input data;
a key-addressing layer configured to generate a weight vector indicating similarities between the key vectors and the input vector;
a value-reading layer configured to compute a context vector from the weight vector and the value vectors; and
an output layer configured to generate predicted time-series data for a target metric of the target entity by at least applying a continuous activation function to the context vector and the input vector.