Techniques to embed a data object into a multidimensional frame
Various embodiments are generally directed to techniques for embedding a data object into a multidimensional frame, such as for training an autoencoder to generate latent space representations of the data object based on the multidimensional frame, for instance. Additionally, in one or more embodiments latent space representations of data objects may be classified, such as with a machine learning algorithm. Some embodiments are particularly directed to embedding a data object comprising a plurality of object entries into a three-dimensional (3D) frame.
1. An apparatus, comprising:
a processor; and
memory comprising instructions that when executed by the processor cause the processor to:
determine a multidimensional frame based on a data object, the multidimensional frame comprising a row dimension, a column dimension, and a contents dimension;
add padding to one or more dimensions of the multidimensional frame based on metadata included in the data object that indicates one or more characteristics of the multidimensional frame;
utilize a trained encoder to embed the multidimensional frame into a latent space representation, wherein the trained encoder is trained with a machine learning model to generate latent space representations of multidimensional frames;
compute a set of correlation scores for the latent space representation of the multidimensional frame, wherein each correlation score in the set of correlation scores corresponds to a different latent space representation of a different multidimensional frame generated based on a different data object;
compare the set of correlation scores to a correlation threshold to identify a set of correlated latent space representations;
identify a set of correlated data objects to the data object based on the set of correlated latent space representations; and
provide the set of correlated data objects as output.
2. The apparatus of claim 1 , the memory comprising instruction that when executed by the processor to cause the processor to generate the trained encoder with the machine learning model, wherein the trained encoder comprises a neural network trained on a set of training multidimensional frames.
3. The apparatus of claim 2 , wherein the machine learning model comprises an autoencoder.
4. The apparatus of claim 2 , wherein the trained encoder to identify patterns in the set of training multidimensional frames, the identified patterns in the set of training multidimensional frames to enable the trained encoder to identify the latent space representations of multidimensional frames.
5. The apparatus of claim 2 , wherein the latent space representation comprises a latent variable value set for each training frame of the set of training multidimensional frames.
6. The apparatus of claim 5 , wherein each latent variable value set comprises a one-dimensional array of values, and each value of each one dimensional array of values to correspond to a dimension in a training frame.
7. The apparatus of claim 1 , the memory comprising instructions that when executed by the processor to cause the processor to utilize a second machine learning model to train a classifier to compute the set of correlation scores, wherein the classifier is trained on a set of training latent space representations generated with the trained encoder based on a set of training 3D frames.
8. A computer-implemented method, comprising:
determining a multidimensional frame based on a data object, the multidimensional frame comprising a row dimension, a column dimension, and a contents dimension;
adding padding to one or more dimensions of the multidimensional frame based on metadata included in the data object that indicates one or more characteristics of the multidimensional frame;
utilizing a trained encoder to embed the multidimensional frame into a latent space representation, wherein the trained encoder is trained with a machine learning model to generate latent space representations of multidimensional frames;
computing a set of correlation scores for the latent space representation of the multidimensional frame, wherein each correlation score in the set of correlation scores corresponds to a different latent space representation of a different multidimensional frame generated based on a different data object;
comparing the set of correlation scores to a correlation threshold to identify a set of correlated latent space representations;
identifying a set of correlated data objects to the data object based on the set of correlated latent space representations; and
providing the set of correlated data objects as output.
9. The computer-implemented method of claim 8 , comprising generating the trained encoder with the machine learning model, wherein the trained encoder comprises a neural network trained on a set of training multidimensional frames.
10. The computer-implemented method of claim 9 , wherein the machine learning model comprises an autoencoder.
11. The computer-implemented method of claim 9 , wherein the trained encoder to identify patterns in the set of training multidimensional frames, the identified patterns in the set of training multidimensional frames to enable the trained encoder to identify the latent space representations of multidimensional frames.
12. The computer-implemented method of claim 9 , wherein the latent space representation comprises a latent variable value set for each training frame of the set of training multidimensional frames.
13. The computer-implemented method of claim 12 , wherein each latent variable value set comprises a one-dimensional array of values, and each value of each one dimensional array of values to correspond to a dimension in a training frame.
14. The computer-implemented method of claim 8 , comprising utilizing a second machine learning model to train a classifier to compute the set of correlation scores, wherein the classifier is trained on a set of training latent space representations generated with the trained encoder based on a set of training 3D frames.
15. At least one non-transitory computer-readable medium comprising a set of instructions that, in response to being executed by a processor circuit, cause the processor circuit to:
determine a multidimensional frame based on a data object, the multidimensional frame comprising a row dimension, a column dimension, and a contents dimension;
add padding to one or more dimensions of the multidimensional frame based on metadata included in the data object that indicates one or more characteristics of the multidimensional frame;
utilize a trained encoder to embed the multidimensional frame into a latent space representation, wherein the trained encoder is trained with a machine learning model to generate latent space representations of multidimensional frames;
compute a set of correlation scores for the latent space representation of the multidimensional frame, wherein each correlation score in the set of correlation scores corresponds to a different latent space representation of a different multidimensional frame generated based on a different data object;
compare the set of correlation scores to a correlation threshold to identify a set of correlated latent space representations;
identify a set of correlated data objects to the data object based on the set of correlated latent space representations; and
provide the set of correlated data objects as output.
16. The at least one non-transitory computer-readable medium of claim 15 , comprising instructions that, in response to being executed by the processor circuit, cause the processor circuit to generate the trained encoder with the machine learning model, wherein the trained encoder comprises a neural network trained on a set of training multidimensional frames, wherein the machine learning model comprises an autoencoder.
17. The at least one non-transitory computer-readable medium of claim 16 , wherein the trained encoder to identify patterns in the set of training multidimensional frames, the identified patterns in the set of training multidimensional frames to enable the trained encoder to identify the latent space representations of multidimensional frames.
18. The at least one non-transitory computer-readable medium of claim 16 , wherein the latent space representation comprises a latent variable value set for each training frame of the set of training multidimensional frames.
19. The at least one non-transitory computer-readable medium of claim 18 , wherein each latent variable value set comprises a one-dimensional array of values, and each value of each one dimensional array of values to correspond to a dimension in a training frame.
20. The at least one non-transitory computer-readable medium of claim 10 , comprising instructions that, in response to being executed by the processor circuit, cause the processor circuit to utilize a second machine learning model to train a classifier to compute the set of correlation scores, wherein the classifier is trained on a set of training latent space representations generated with the trained encoder based on a set of training 3D frames.