匿名使用者
匿名使用者 發問時間: 社會與文化語言 · 1 0 年前

有關演算法網路訓練的段落..請幫我翻譯一下唄..

有哪位好心人士,請幫我翻譯一下: ^O^ ~感恩感恩~

1.

由文獻(literature)研究中發現,有學者採用A-Net為架構並輔以B和C等單一方法(approach)來完成其網路的訓練與學習。而在藉A-Net去逼近(approximation)函數的過程中,吾發現可嘗試對前述的單一方法做些巧思來結合並發展出一混合式演算法,以對其網路相關參數值做適恰地訓練與調適,來獲得較單一方法更精確的學習績效.

2.

由Table V之數據可知:training set和validation set的結果皆呈現出一致性的低,其所得結果不僅適用於所訓練和驗證的dataset,也能普遍化(generalization)的適用於其它unseen dataset. 由此可知,the proposed algorithm於所進行驗證的實驗(experiment)中並無over-fitting和over-training的problem.

2 個解答

評分
  • Anny
    Lv 6
    1 0 年前
    最佳解答

    1.

    From the literature (literature) study found that some scholars have used A-Net as the framework and supplemented with B and C, and so on a single method (approach) to complete its network of training and learning. The A-Net to use approximation (approximation) function in the process, I found the above may try to do a single method for combining and ingenuity to develop a hybrid algorithms to its network-related parameters to do fitness And training and adjustment, to obtain more accurate than a single method of learning achievement.

    2.

    Data from Table V know: training set and the validation set Jiecheng show consistency of the results of the low, the findings apply not only to training and certification by the dataset, can be generalized (generalization) of the applicable to other unseen dataset. By Know this, the proposed algorithm to verify the experiment conducted by the (experiment) there is no over-fitting and over-training of the problem.

    參考資料: 自己+電視
  • piglet
    Lv 6
    1 0 年前

    1.

    From the literature (literature) study found that some scholars have used A-Net as the framework and supplemented with B and C, and so on a single method (approach) to complete its network of training and learning. The A-Net to use approximation (approximation) function in the process, I found the above may try to do a single method for combining and ingenuity to develop a hybrid algorithms to its network-related parameters to do fitness And training and adjustment, to obtain more accurate than a single method of learning achievement.

    2.

    Data from Table V know: training set and the validation set Jiecheng show consistency of the results of the low, the findings apply not only to training and certification by the dataset, can be generalized (generalization) of the applicable to other unseen dataset. By Know this, the proposed algorithm to verify the experiment conducted by the (experiment) there is no over-fitting and over-training of the problem.

還有問題?馬上發問,尋求解答。