Zhiyun Lu

Title: Zhiyun Lu: Innovator in Speech Recognition Technology

Introduction

Zhiyun Lu is a notable inventor based in Brooklyn, NY, recognized for his contributions to the field of speech recognition technology. With a focus on enhancing the efficiency and accuracy of speech recognition systems, Lu has made significant strides in this innovative domain.

Latest Patents

Zhiyun Lu holds a patent for the invention titled "Fusion of acoustic and text representations in RNN-T." This speech recognition model comprises an encoder network, a prediction network, and a joint network. The encoder network is designed to receive a sequence of acoustic frames that characterize an input utterance and generate a higher-order feature representation for each corresponding acoustic frame. The prediction network processes a sequence of non-blank symbols output by a final Softmax layer, generating a dense representation at each output step. The joint network combines the higher-order feature representation and the dense representation to produce a probability distribution over possible speech recognition hypotheses. This innovative approach includes a stack of gating and bilinear pooling to effectively fuse the two representations.

Career Highlights

Zhiyun Lu is currently employed at Google Inc., where he continues to develop and refine advanced technologies in speech recognition. His work at Google has positioned him as a key player in the ongoing evolution of artificial intelligence and machine learning applications.

Collaborations

Lu collaborates with talented individuals in his field, including Chao Zhang and Bo Li. Their combined expertise contributes to the advancement of innovative solutions in speech recognition technology.

Conclusion

Zhiyun Lu's work in speech recognition exemplifies the impact of innovative thinking in technology. His patent and contributions at Google Inc. highlight his role as a significant inventor in this rapidly evolving field.