Lossy Compression
A form of data compression that preserves the semantic core of information within limited space but discards details deemed secondary, making full restoration of the original text impossible. In AI conversations it is used for summarization and context compression.
Timeline (last 90 days)
Hierarchical discrete codes are essentially lossy compression of text, and for scenarios requiring precise characters such as code and mathematical formulas, latent representations may be inadequate.
Video compression (such as H.264/H.265) encodes in blocks and leaves faint "halos" or blocky artifacts around high-contrast edges (for example, white digits on a dark background), and these artifacts get framed into the cleanup region along with the digits.
When compressing information, models usually tend to retain topic, intent and emotional tone while treating numbers, dates, negations and precise conditions as information that can be inferred or ignored.
Lossless compression merely re-encodes the data and can be fully restored after decompression, making it suitable for text and vector graphics; lossy compression discards some information, usually reducing resolution or increasing the compression rate for images, with a marked reduction in size but a loss of detail.
All Facts (5)
JPEG压缩中quality参数本质上控制的是量化步长，值越低丢弃的高频信息越多，文件越小但细节损失越大
95%UnverifiedHierarchical discrete codes are essentially lossy compression of text, and for scenarios requiring precise characters such as code and mathematical formulas, latent representations may be inadequate.
50%UnverifiedVideo compression (such as H.264/H.265) encodes in blocks and leaves faint "halos" or blocky artifacts around high-contrast edges (for example, white digits on a dark background), and these artifacts get framed into the cleanup region along with the digits.
50%UnverifiedWhen compressing information, models usually tend to retain topic, intent and emotional tone while treating numbers, dates, negations and precise conditions as information that can be inferred or ignored.
50%UnverifiedLossless compression merely re-encodes the data and can be fully restored after decompression, making it suitable for text and vector graphics; lossy compression discards some information, usually reducing resolution or increasing the compression rate for images, with a marked reduction in size but a loss of detail.
50%