Kristoffer Sjöö

Title: Kristoffer Sjöö: Innovator in Facial Position Data Generation

Introduction

Kristoffer Sjöö is a notable inventor based in Stockholm, Sweden. He has made significant contributions to the field of machine learning and audio data processing. With a total of 3 patents, his work focuses on generating facial position data through innovative methods.

Latest Patents

One of Kristoffer's latest patents involves a computer-implemented method for generating a machine-learned model to produce facial position data based on audio data. This method includes training a conditional variational autoencoder that consists of an encoder and decoder. The training process involves receiving a set of training data items, each containing a facial position descriptor and an audio descriptor. The encoder processes these training data items to obtain distribution parameters, from which a latent vector is sampled. This latent vector, along with the audio descriptor, is then processed by the decoder to generate a facial position output. The method also includes calculating a loss value based on the comparison of the output and the facial position descriptor, followed by updating the parameters of the conditional variational autoencoder based on the calculated loss value.

Career Highlights

Kristoffer currently works at Electronic Arts Inc., where he applies his expertise in machine learning and audio data processing. His innovative approaches have contributed to advancements in the gaming and technology sectors.

Collaborations

Some of Kristoffer's coworkers include Jorge Del Val Santos and Martin Singh-Blom. Their collaborative efforts enhance the innovative environment at Electronic Arts Inc.

Conclusion

Kristoffer Sjöö is a prominent inventor whose work in generating facial position data based on audio data showcases his expertise and innovative spirit. His contributions continue to influence the fields of machine learning and audio processing.