[數撚圍爐區] Scientia Manu et Mente (67)

1001 回覆
2 Like 1 Dislike
2022-07-01 00:06:32
2022-07-01 00:10:49
2022-07-01 00:11:26
2022-07-01 00:15:47
咪話做research assistant囉, 都係工
2022-07-01 02:14:36
2022-07-01 02:43:42
唔一定要講
2022-07-01 02:44:28
2022-07-01 07:06:43
2022-07-01 09:12:21
2022-07-01 10:05:08
e.g. 呢段係 generalized power method 度搵出嚟
Ax=A'*x;
for i=1:m,
     Ax(:,i)=Ax(:,i)*mu(i);
end
tresh=max(Ax.^2-repmat(RHO,n,1),0);
f(iter)=sum(sum(tresh)); %cost function
if f(iter)==0, %the sparsity parameter is too high: all entries of the loading vector are zero.
    break
else
grad=zeros(p,m);
for i=1:m,
    pattern=find(tresh(:,i)'>0);
    grad(:,i)=A(:,pattern)*Ax(pattern,i)*mu(i); %gradient 
end                      
[U,S,V]=svd(grad,'econ');
x=U*V';
end
2022-07-01 10:27:03
2022-07-01 21:32:27
2022-07-01 21:57:58
2022-07-01 21:58:31
2022-07-01 22:05:52
下, 醫療好先進wor
早排先有單豬心移植
你盧寵物識唔識
2022-07-01 22:19:07
吹水台自選台熱 門最 新手機台時事台政事台World體育台娛樂台動漫台Apps台遊戲台影視台講故台健康台感情台家庭台潮流台美容台上班台財經台房屋台飲食台旅遊台學術台校園台汽車台音樂台創意台硬件台電器台攝影台玩具台寵物台軟件台活動台電訊台直播台站務台黑 洞