M-theory (learning framework)

<p>In <a href="/facts/Machine_learning/e0w0XJTu">machine learning</a> and <a href="/facts/Computer_vision/Tl2Yyk66">computer vision</a>, M-theory is a learning framework inspired by feed-forward processing in the <a href="/facts/Two-streams_hypothesis/rrlVjuLm">ventral stream</a> of <a href="/facts/Visual_cortex/qmF8F5Sb">visual cortex</a> and originally developed for recognition and classification of objects in visual scenes. M-theory was later applied to other areas, such as <a href="/facts/Speech_recognition/z7S7Pgk6">speech recognition</a>. On certain image recognition tasks, algorithms based on a specific instantiation of M-theory, HMAX, achieved human-level performance.
</p><p>The core principle of M-theory is extracting representations invariant under various transformations of images (translation, scale, 2D and 3D rotation and others). In contrast with other approaches using invariant representations, in M-theory they are not hardcoded into the algorithms, but learned. M-theory also shares some principles with <a href="/facts/Compressed_sensing/aR5DQm3n">compressed sensing</a>. The theory proposes multilayered hierarchical learning architecture, similar to that of visual cortex.
</p>

M-theory (learning framework) open-in-new

M-theory (learning framework)