Linling Zhang

Title: Linling Zhang: Innovator in Deep Neural Networks and Data Compression

Introduction

Linling Zhang is a prominent inventor based in Watford, GB. He has made significant contributions to the fields of deep neural networks and data compression, holding a total of 11 patents. His innovative approaches have advanced the capabilities of hardware implementations and data handling techniques.

Latest Patents

Among his latest patents is the "Hierarchical mantissa bit length selection for hardware implementation of deep neural network." This patent describes hierarchical methods for selecting fixed-point number formats with reduced mantissa bit lengths for representing values input to, and/or output from, the layers of a deep neural network (DNN). The methods begin with one or more initial fixed-point number formats for each layer. The layers are divided into subsets, and the mantissa bit lengths of the fixed-point number formats are iteratively reduced from the initial formats on a per-subset basis. If a reduction causes the output error of the DNN to exceed an error threshold, then the reduction is discarded, and no further reductions are made to the layers of the subset. Otherwise, a further reduction is made to the fixed-point number formats for the layers in that subset. This process continues until no further reductions can be made to any of the subsets, after which the method is repeated for continually increasing numbers of subsets until a predetermined number of layers per subset is achieved.

Another notable patent is for a "Lossy data compression" method. This method involves compressing data, such as image data, using wrap-around wavelet compression. Each data value is divided into two parts, with the first parts comprising the most significant bits from the data values being compressed using wrap-around wavelet compression. Depending on the target compression ratio and the achieved compression ratio, additional bits from the second parts may be appended to the compressed first parts. This method can be either lossy or lossless, and a corresponding decompression method is also described.

Career Highlights

Linling Zhang is currently employed at Imagination Technologies Limited, where he continues to innovate and develop cutting-edge technologies. His work has had a profound impact on the efficiency and effectiveness of data processing and neural network implementations.

Collaborations

Linling has collaborated with notable colleagues, including Simon James Fenney and James Imber, contributing to a dynamic and innovative work environment.

Conclusion