review expectations of incremental training in various objects - was UMAP incremental training doesn't work

[tremblap]

this patch illustrates how 100x 1 iteration (actually 1000x neither, I tested) is not getting anything beyond the first result, compare to the first result of 100 iteration in one go.

I reckon there is a state that gets lost or not passed between iterations?

----------begin_max5_patcher----------
2744.3oc6aszjihiD9rqeEDN1idcfjP7XNM69CXurGmYBGBP1EciAVPTcU6D
y7ae0CdHrAL1.UUSuccvksDBx7KSkYpLS98m1r0O8UZwViex3WL1r42eZyF4
PhA1T86MaOSdMHlTHursAomOSSXa2oliQekIGGXDwn4DVTZhAgYPLXQmo0WU
R44njXJSdK.sClVxpG0rZznP4sK0+K+cjc85yHrfmiRNcHmFvTTqEBu2bmA.
4J+G1T7OH+SieSrl+3omDeratbkoYKeMStAOB2.Lg1RFvyZIXmyzhBxI5UrS
VZTBqH5+RKBHwTCy83d4IXu7D3ZdB5Vud00wdKipXnsaM9sQXWfsjcwVJomo
jqgvGjcSneiSMWwsGiKiB2mjlelDy44+7Njevq4UvH75tavudV.ACZgj+yCs
FbaNmrnFgkmytCYZe7I3w4S91EnTthkhWauYwnCoEmSORyM9avc2K21iF7Cp
.WKPskJttNqm5aHgQJnrYp7hmstKFH0ZcMWOVs7LI6OM9YNyDx8cjTvM4VX.
M94FCvEBEr4ADvIBDAowo4pYL2aAQXtkY.vw1xwzdmvvI.4.wVdl.OaSGWtM
LogLArLFTJQOjsz0Ez1cU1dbLhwxIIEG4F9LdlFmIA18f8jDR7aEQEyaGC5A
2w3IQHDT411EOKd+KEkQixF9o4gzbcWweofjepT3kunVthaH3aaUr04cN4Lk
qQdflP7io5OhanScLJlulyTMU9r3TF+dMZjAlxMd1U9HcLa92hp0DDSI4ySy
.X8XpFnpsE.oNAZlFSKO6Symoqd3CKs8IImFkYsLq17KYS7pXNMHsLgM.Hf6
EDr5ADLGTZFIhTtR6dm5miJe8TrrJTcavrXYV5oSw2yAK5QQ06gDu2hMqjrU
mHAZU+4RJY+ObpMO0.hMm2VU6Aks2TAFoLcCbjroy7hbWvX9kLV6gqdVBy+j
PypxpYZBqwp4+HOhDa7OSiCqmlG1aRHUxY1hMSinWDcheV.ZPbTvWGVcAoA4
cBB.wc1C20yWzc5qAwVSQIqQv1b0wzSjf257zqIT+SZDjbWj4dGWWdvIhuwi
LA651kb52gztIqK6YuGqNF5LNL52wxXL5tDxvOWBYOO8bN7Co7PRYH9tjxfO
WRYnqqJvamEXu7jNpZmCcDNyXwPfOxSrYqRgQ0I1.NNerm9clm88C8vuHjNT
BcvqATlY7qaqR3WgDz1YHOhEmTLHIgFeMjkSo+514E+nE9AOniBBrLqR.nVv
+5KRct0MMbt5jiuPyE5DZjzlsjrLsg2nsDAb8EkPzcWyPQIpgPMCkSeI5xaK
+9lyIcFmtKyUVqe0tN.EwsIkez5D9AvEqQMHWvUQRRQjvJeQFIPsXgjrd5Vv
vUkoaUJQcPxeYgwsfgvJYbZvWUtDLqGLMilDkjkSK3GkWkWd8oCoGIkwrCBm
MhLcKIR39dm+XEE16jc8T0v.mxiBSSDDQGIgX35GG2htxbQGlQdEIjrdVLWm
hCKCLYAmIKK7I4BAUkOGX8jrzz3tS0rtX5QV0zYQIIWfhrzrgmLO5zyirVe9
FpzyicukyTbnLQM6AtNA6PA4ktnMiDGWsOt6s+URRzYBiJKgjfcMalT4284h
f7z33N7qZlW5YlPtNd.8aQgrmkOHckA9kGkUqDssQJGFchVv5NFibpn6HEr2
Tft1Pk9U6gOvnmyh4bQ2Kfu6HpfU7b52JptvZEMc.nslb56o0MJ1Y7QywyDK
AzfmlbXahCcjmaXZbv3QUYHDi5slPapcRTYu4AwlNNN5qBJ5oO8AfFX+PCXb
nY2DfGfSa5S77VEvoaYP6hN+qJ2pTC1ybTRKlEN6jUxL7oAjR9zQLCtwYtYz
BinDtWWCRteDGRyeyfiqm3qMMWdOdgDWRK1q+zhiRnxrWokTpAcTObcW6Ote
bU7JVHY7K6l3PZw0LT0iF1guxIGPF6M.iUYx88Skdz.ySFSCGbuZ3VvEPEWE
Rj.nD3kG3iwDfVYEWfs9VK1d+JSi1du2vBSBIF9Kn1h4nfhRBbQ1HGGbfpdI
.i+dxsAxcozcf.yUzwwmCSPH7hgVpRVrVVfFK.sAKeaX98.d2cvanYF8lB3.
lJELazGid1Uc6yRnVgVJ0pJGadqC3HYwAYay6NZdzhnPHywLxr9yNqpNdS8S
HIohjP5qZoWXQfGEGrcdAUpiOVamjSakojOCHPFIgFq+T7IAe8jLu68f.8CM
sMVA7dALPm.gT2ndRCtGDfscsgNVXSO2cFlWjbwgfa2pJVIS1sJtaQh15rLQ
pqtjB4gXTkUEm1TCbUR5M2KqTP6GWPO4oYo4MYiZOtGwVatfDGyYfS3KYLw7
8KOKRKyCp2BVcvRitjRHsfEkzjYreo07twTs2buDgvB4znB20jJrlJUfVSpf
GDhAXBTw0T64nPU5Yp1P34J7VTk0XfoYyu5rp1Nc5Cle.2GCAsuA+.VG9wch
ZIhinrZZISUU051ZIU5EdlZfp2zywvcS4lScWF9VjthXgNfVJW9i0hxA2Inu
NTwTsUBFjHpFrtnKaEYVO7fpJBGH7iOD4yORZwkUQ5tx18o3TeR7Eontujg+
TKw8d0.qSo.ySueVMevh4o1vU0PqPu40IeC8Bn7uYjblg+aFwojPQULKROSM
DmldzRX5W52sdP298SAZOZWnZ2o6cMmUysEFEv1+RD8ay6MpAZMBEC6TpUUW
yB7dTwS2Na4wpZL7w5g3a1PdJNsVObdsc36eeU68Q0X00EiGgWj9ptaCI7Bo
0loXmamcrc5VglYmaSIffOlcLndaY3B++4VR.YoXeUGtZaI+mknoe9QOI7id
R3G8jvb5IgQqtziTXoAxZ43UgbR0Sx0qIXAksv22DY6Wd7HMuakRJh4ZKEhp
tyCTjbHHMOqr3fZz8ei7xxkr6lv.6G.Ox8Wwlbk4PPjxGaUiF69Wg7d6NeMH
fUss1uOS8M3VM9.zUWv+Y.BtYAbyoh.wFqx+nAgIz.6klPKQLkhGAsj5Sh2S
NGQ4ibd2KJoDbtv5y4iAA6+RgnGaevFlXfMfdyqtSp2rNnsmLeR.7mnBdewa
f6BXtdAp3sBuPlt6Ax1cp5sBYHDaUphAbh4DyaESL2TyKLbEoAuIRC3UjFlZ
99gqIPHu4SgJ.t2LKy7.4vsu3AZ+5F52xPoGPAu4jXbqB1Zmlr8.a8ynjf.t
KuN0UTVhSWjGDHe4JbL8bMstthmW+JC09gVJW515iGhRDmFmV+rL2o8wEYZe
mlvJWjNmqNIg.a4bVjdwl5SX2AbtH5rwgmddySjueVxW2DYoWcrvf2KR+JOh
iS79mNFEG2vBat5JFtVy7YJYomxIgQUQD0Yxt0gV9d2HgkJUEw23CAtRooZk
fFUMa9eN658a8uTX8Rae6zt9aWszKqFNxSe1ZGTaa32lI0DkKtV706TueMhK
Cj9NzlkzqEt9UNzhK53vxRoI+IujPR0itALnRp5k9EpRs509C5Xl+1UL6pr2
N4GzTdNtKvCBLENZQdPf2InqSfCC8bVhGzTPN7B7bPS34.WhmyTT5fVKg18T
zEPKAOYgmhVm2RvSSA8rVD8AyoHmVBzyaBOH6E34XOEgTWtta6F83VklBVBf
uS7HZ.dDLOdz5wzWTdKunRZhGxEUP6hpmcckyFtpYWVwLYfCCUorm9im9erA
zXoG
-----------end_max5_patcher-----------

[weefuzzy]

The object wasn't designed to support incremental training from what I can see. Where did the expectation that it does come from?

[tremblap]

Like all other iterative process we have. it is even in the learn material. I'll leave that open as this is the only one that doesn't do it in our toolset as far as I know, or at least we have some that do and some that don't which is far from a good idea - this comes from a user that was unprompted and following our interface design expected that to be the case...

[weefuzzy]

Apologies, closed by mistake.

Apart from the MLP, which iterative processes did you have in mind? KMeans doesn't support this: incremental fitting has to be done by hand. NMF only kind of does, to the extent that its complete state can be represented in a couple of buffers. As far as I can see, MLP is the odd one out here.

Looking at the code, I'm not even sure what incremental fitting would entail for UMAP. There are a series of convergent processes, and the iteration count only pertains to the last of these (optimising the layout), once the embedding has already been learnt (for which we use library code). So fit can't be incremental.

If you call transform, it is using the learned embedding (but not changing it?) on a whole new input dataset, assumed to be high dimensional, applying the embedding and then optimising the layout. Making the layout optimisation incremental would entail some other method that took a low dimensional input and updated that with n iterations of layout optimisation.

So, my view is that this not a bug, but a feature request that I'm not sure is actually possible (or at least simple) with the code we have.

[tremblap]

thanks for the explanation. We should definitely make it clearer in the doc if this feature request is not possible.

But (s)kmeans support(s) it. There is even a tab showing how it works :) so with MLP there is something of an (wrong) assuption that it should happen in everything iterative, which is obviously wrong, so when I'm rested post holidays I'll review iterative assumptions and expectations. I'm changing the ticket name now and will assigned to me (and maybe @gerard so we can discuss that)

thanks again for your patience