《莊子‧外篇 》
《達生》

仲尼適楚，出於林中，見痀僂者承蜩，猶掇之也。仲尼曰：「子巧乎？有道邪？」曰：「我有道也。五六月累丸，二而不墜，則失者錙銖；累三而不墜，則失者十一；累五而不墜，猶掇之也。吾處身也若厥株拘，吾執臂也若槁木之枝，雖天地之大，萬物之多，而唯蜩翼之知。吾不反不側，不以萬物易蜩之翼，何為而不得！」孔子顧謂弟子曰：「用志不分，乃凝於神，其痀僂丈人之謂乎！」

顏淵問仲尼曰：「吾嘗濟乎觴深之淵，津人操舟若神。吾問焉，曰：『操舟可學邪？』曰：『可。善游者數能。若乃夫沒人，則未嘗見舟而便操之也。』吾問焉而不吾告，敢問何謂也？」仲尼曰：「善游者數能，忘水也。若乃夫沒人之未嘗見舟而便操之也，彼視淵若陵，視舟之覆猶其車卻也。覆卻萬方陳乎前而不得入其舍，惡往而不暇！以瓦注者巧，以鉤注者憚，以黃金注者殙。其巧一也，而有所矜，則重外也。凡外重者內拙。」

梓慶削木為鐻，鐻成，見者驚猶鬼神。魯侯見而問焉，曰：「子何術以為焉？」對曰：「臣工人，何術之有！雖然，有一焉。臣將為鐻，未嘗敢以耗氣也，必齊以靜心。齊三日，而不敢懷慶賞爵祿；齊五日，不敢懷非譽巧拙；齊七日，輒然忘吾有四枝形體也。當是時也，無公朝，其巧專而外骨消；然後入山林，觀天性；形軀至矣 ，然後成見鐻，然後加手焉；不然則已。則以天合天，器之所以疑神者，其是與？」

 

痀僂志一處凝神、津人忘水無所矜、梓慶齋心以天合皆是『技』可通『道』者也。方法萬千，主題則一。其神只能神會，其巧只能自得，其器只能己鑄。故無言可說，但請讀 Michael Nielsen 先生之文自然能了︰

Other techniques for regularization

There are many regularization techniques other than L2 regularization. In fact, so many techniques have been developed that I can’t possibly summarize them all. In this section I briefly describe three other approaches to reducing overfitting: L1 regularization, dropout, and artificially increasing the training set size. We won’t go into nearly as much depth studying these techniques as we did earlier. Instead, the purpose is to get familiar with the main ideas, and to appreciate something of the diversity of regularization techniques available.

……

Summing up: We’ve now completed our dive into overfitting and regularization. Of course, we’ll return again to the issue. As I’ve mentioned several times, overfitting is a major problem in neural networks, especially as computers get more powerful, and we have the ability to train larger networks. As a result there’s a pressing need to develop powerful regularization techniques to reduce overfitting, and this is an extremely active area of current work.

───