公开(公告)号：US07385611B1

公开(公告)日：2008-06-10

申请号：US11295774

申请日：2005-12-07

申请人： Michael J. M. Toksvig ,  Apoorv Gupta

发明人： Michael J. M. Toksvig ,  Apoorv Gupta

IPC分类号： G09G5/00 ,  G06K9/00

CPC分类号： G06K9/00275 ,  H04N19/186 ,  H04N19/21 ,  H04N19/44

摘要： Systems and methods that decompress block compressed texture data may decompress the texture data while simplifying computations to reduce die area while maintaining the required accuracy. Reducing the die area permits more texture data to be decompressed in the same die area compared with a more accurate decompression, thereby increasing texture decompression throughput. Computations are simplified by combining denominators for linear interpolation with format conversion to decompress texture data components compressed using conventional block compression formats.

摘要翻译： 解压缩块压缩纹理数据的系统和方法可以解压缩纹理数据，同时简化计算以减少管芯面积，同时保持所需的精度。 与更准确的解压缩相比，减小管芯面积允许在相同管芯区域中更多的纹理数据被解压缩，从而增加纹理解压缩吞吐量。 通过将用于线性插值的分母与格式转换组合来解压缩使用常规块压缩格式压缩的纹理数据分量，来简化计算。

公开(公告)号：US20230097940A1

公开(公告)日：2023-03-30

申请号：US17897570

申请日：2022-08-29

申请人： David Sandai Kurokawa ,  Shayak Sen ,  Anupam Datta ,  Divya Gopinath ,  Apoorv Gupta

发明人： David Sandai Kurokawa ,  Shayak Sen ,  Anupam Datta ,  Divya Gopinath ,  Apoorv Gupta

IPC分类号： G06N20/00

摘要： In some implementations, a computing machine accesses an artificial intelligence model and a dataset for the artificial intelligence model, the dataset comprising at least one datapoint. The computing machine identifies a feature group used by the artificial intelligence model, the feature group comprising at least two features having a similarity with one another exceeding a similarity threshold, wherein the feature group comprises a subset of the features used by the artificial intelligence model. The computing machine determines an overall influence value for the feature group on an output of the artificial intelligence model applied to the dataset. The computing machine provides an output representing the overall influence value.

公开(公告)号：US08094164B1

公开(公告)日：2012-01-10

申请号：US12127637

申请日：2008-05-27

申请人： Michael J. M. Toksvig ,  Apoorv Gupta

发明人： Michael J. M. Toksvig ,  Apoorv Gupta

IPC分类号： G09G5/00

CPC分类号： G06K9/00275 ,  H04N19/186 ,  H04N19/21 ,  H04N19/44

摘要： Systems and methods that decompress block compressed texture data may decompress the texture data while simplifying computations to reduce die area while maintaining the required accuracy. Reducing the die area permits more texture data to be decompressed in the same die area compared with a more accurate decompression, thereby increasing texture decompression throughput. Computations are simplified by combining denominators for linear interpolation with format conversion to decompress texture data components compressed using conventional block compression formats.

摘要翻译： 解压缩块压缩纹理数据的系统和方法可以解压缩纹理数据，同时简化计算以减少管芯面积，同时保持所需的精度。 与更准确的解压缩相比，减小管芯面积允许在相同管芯区域中更多的纹理数据被解压缩，从而增加纹理解压缩吞吐量。 通过将用于线性插值的分母与格式转换组合来解压缩使用常规块压缩格式压缩的纹理数据分量，来简化计算。