Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating a sequence of data elements that includes a respective data element at each position in a sequence of positions. In one aspect, a method includes: for each position after a first position in the sequence of positions: obtaining a current sequence of data element embeddings that includes a respective data element embedding of each data element at a position that precedes the current position, obtaining a sequence of latent embeddings, and processing: (i) the current sequence of data element embeddings, and (ii) the sequence of latent embeddings, using a neural network to generate the data element at the current position. The neural network includes a sequence of neural network blocks including: (i) a cross-attention block, (ii) one or more self-attention blocks, and (iii) an output block.

Abstract: Systems, methods, and computer-readable storage media for facilitating local recognition of audio content at a user device. In some embodiments, the method includes capturing, using a user device, audio data, at least some of which is processable to recognize the audio data. Thereafter, an audio fingerprint that uniquely represents perceptual information associated with the audio data is generated, and a local data store within the user device is referenced. Such a local data store can include reference audio fingerprints. Upon referencing the local data store, a determination can be made as to whether the generated audio fingerprint matches a reference audio fingerprint at least to an extent.

Abstract: Systems, methods, and computer-readable storage media for facilitating recognition of real-time content are provided. In embodiments, a new audio fingerprint associated with live audio being presented is received. In accordance with the received audio fingerprint, at least one previously received fingerprint associated with the live audio from a real-time index is removed. Thereafter, the real-time index is updated to include the new audio fingerprint associated with the live audio being presented. Such a real-time index having the new audio fingerprint can be used to recognize the live audio being presented and, thereafter, an indication of the recognized live audio can be provided to the user device.