SYSTEMS AND METHODS FOR GENERATING MUSIC RECOMMENDATIONS
Systems, methods, and non-transitory computer-readable media can be configured to determine a video embedding for a video content item based at least in part on a first machine learning model. A set of music embeddings can be determined for a set of music content items based at least in part on a second machine learning model. The set of music content items can be ranked based at least in part on the video embedding and the set of music embeddings.
1 . A computer-implemented method comprising:
generating, by a computing system, a video embedding for a video content item based at least in part on a first machine learning model;
generating, by the computing system, a set of music embeddings for a set of music content items based at least in part on a second machine learning model, wherein a music embedding of the set of music embeddings is generated based at least in part on a combination of music feature embeddings associated with a corresponding music content item of the set of music content items and one or more values are removed from the combination of music feature embeddings based at least in part on the second machine learning model; and
ranking, by the computing system, the set of music content items based at least in part on the video embedding and the set of music embeddings.
2 . The computer-implemented method of claim 1 , further comprising:
generating one or more video feature embeddings based at least in part on one or more video features associated with the video content item; and
wherein the video embedding is generated based at least in part on the one or more video feature embeddings.
3 . The computer-implemented method of claim 2 , wherein the one or more video features associated with the video content item includes at least one of: a concept, an object, or a visual characteristic identified in the video content item.
4 . The computer-implemented method of claim 1 , further comprising:
generating the music feature embeddings for the corresponding music content item based at least in part on music features associated with the corresponding music content item; and
wherein the combination of the music feature embeddings is based at least in part on a concatenation of the music feature embeddings.
5 . The computer-implemented method of claim 4 , wherein the music features associated with the corresponding music content item include at least one of: a title, an artist, a lyric, a genre, or a spectrogram associated with the corresponding music content item.
6 . The computer-implemented method of claim 1 , wherein the ranking the set of music content items comprises:
generating a subset of music embeddings based at least in part on a proximity between the video embedding and the set of music embeddings.
7 . The computer-implemented method of claim 6 , wherein the ranking the set of music content items further comprises:
ranking a subset of the set of music content items associated with the subset of music embeddings based at least in part on a measure of similarity between the video embedding and the subset of music embeddings.
8 . The computer-implemented method of claim 1 , wherein the video embedding and the set of music embeddings are mapped in a vector space.
9 . The computer-implemented method of claim 1 , wherein the first machine learning model and the second machine learning model are trained based at least in part on training sets of data that include training video content items and training music content items included in the training video content items.
10 . The computer-implemented method of claim 1 , further comprising:
providing one or more music recommendations based at least in part on the ranking.
11 . A system comprising:
at least one processor; and
a memory storing instructions that, when executed by the at least one processor, cause the system to perform a method comprising:
generating a video embedding for a video content item based at least in part on a first machine learning model;
generating a set of music embeddings for a set of music content items based at least in part on a second machine learning model, wherein a music embedding of the set of music embeddings is generated based at least in part on a combination of music feature embeddings associated with a corresponding music content item of the set of music content items and one or more values are removed from the combination of music feature embeddings based at least in part on the second machine learning model; and
ranking the set of music content items based at least in part on the video embedding and the set of music embeddings.
12 . The system of claim 11 , further comprising:
generating one or more video feature embeddings based at least in part on one or more video features associated with the video content item; and
wherein the video embedding is generated based at least in part on the one or more video feature embeddings.
13 . The system of claim 12 , wherein the one or more video features associated with the video content item includes at least one of: a concept, an object, or a visual characteristic identified in the video content item.
14 . The system of claim 11 , further comprising:
generating the music feature embeddings for the corresponding music content item based at least in part on music features associated with the corresponding music content item; and
wherein the combination of the music feature embeddings is based at least in part on a concatenation of the music feature embeddings.
15 . The system of claim 14 , wherein the one or more music features associated with the corresponding music content item include at least one of: a title, an artist, a lyric, a genre, or a spectrogram associated with the corresponding music content item.
16 . A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:
generating a video embedding for a video content item based at least in part on a first machine learning model;
generating a set of music embeddings for a set of music content items based at least in part on a second machine learning model, wherein a music embedding of the set of music embeddings is generated based at least in part on a combination of music feature embeddings associated with a corresponding music content item of the set of music content items and one or more values are removed from the combination of music feature embeddings based at least in part on the second machine learning model; and
ranking the set of music content items based at least in part on the video embedding and the set of music embeddings.
17 . The non-transitory computer-readable storage medium of claim 16 , further comprising:
generating one or more video feature embeddings based at least in part on one or more video features associated with the video content item; and
wherein the video embedding is generated based at least in part on the one or more video feature embeddings.
18 . The non-transitory computer-readable storage medium of claim 17 , wherein the one or more video features associated with the video content item includes at least one of: a concept, an object, or a visual characteristic identified in the video content item.
19 . The non-transitory computer-readable storage medium of claim 16 , further comprising:
generating the music feature embeddings for the corresponding music content item based at least in part on music features associated with the corresponding music content item; and
wherein the combination of the music feature embeddings is based at least in part on a concatenation of the music feature embeddings.
20 . The non-transitory computer-readable storage medium of claim 19 , wherein the music features associated with the corresponding music content item include at least one of: a title, an artist, a lyric, a genre, or a spectrogram associated with the corresponding music content items.