Timeline for A challenging question of ANN
Current License: CC BY-SA 4.0
29 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 7, 2021 at 4:10 | history | edited | user307505 | CC BY-SA 4.0 |
deleted 221 characters in body
|
| Jan 10, 2021 at 11:27 | comment | added | Javier TG | In the image added with the final update of the question, the data can be separated in different regions by using 3 lines. But note that this is not the case of the first image (the one associated with ($D$) solution). In this first image the data can be separated in different regions just by using 2 lines, and not 3 $\to$ In order to solve the first classification problem, we need at least 5 neurons if we are using bipolar/ step activation functions. | |
| Jan 8, 2021 at 12:18 | history | edited | user307505 |
edited tags
|
|
| Jan 8, 2021 at 3:00 | history | tweeted | twitter.com/StackStats/status/1347377636641792005 | ||
| Jan 7, 2021 at 15:31 | answer | added | Aksakal | timeline score: 1 | |
| Jan 7, 2021 at 15:15 | comment | added | Aksakal | @kevin307505 I think D is correct. on first neuron layer all you need is to split into two linear areas, like a cross sign. no point in having more than two neurons. | |
| Jan 7, 2021 at 15:13 | comment | added | Aksakal | @DaviedZuhraph i mean the first layer of neurons as 1st layer, not the inputs themselves. | |
| Jan 7, 2021 at 14:54 | comment | added | Aksakal | the bipolar node allows you to split the space in three linear areas. so ideally each node in first layer can separate out the L pieces, then the second layer combines the two L areas. when the node is 0/1 then 1st layer splits up into halves that contain both L and not L pieces, and the second layer further splits halves into L and not L, finally the third layer picks up two L areas into one category. i would also edit your question title to make it more descriptive of the actual problem. "difficult" is subjective, I dont find this problem difficult | |
| Jan 7, 2021 at 14:36 | history | edited | user307505 |
edited tags
|
|
| Jan 7, 2021 at 11:09 | answer | added | Igor F. | timeline score: 6 | |
| Jan 7, 2021 at 10:55 | history | edited | user307505 |
edited tags
|
|
| Jan 7, 2021 at 10:38 | history | edited | user307505 | CC BY-SA 4.0 |
added 60 characters in body
|
| Jan 7, 2021 at 10:15 | history | edited | user307505 | CC BY-SA 4.0 |
deleted 47 characters in body
|
| Jan 7, 2021 at 9:57 | history | edited | user307505 | CC BY-SA 4.0 |
added 99 characters in body
|
| Jan 7, 2021 at 9:46 | comment | added | Igor F. | @kevin307505 Then I believe E is wrong. It could be right if the output neuron could multiply its inputs. Then you would have $1 \cdot 1 = 1$ and $-1 \cdot -1 = 1$ for one class, and $1 \cdot -1 = -1$ and $-1 \cdot 1 = -1$ for the other. But for additive neurons I don't see a solution. | |
| Jan 7, 2021 at 9:26 | comment | added | Thomas | I would be interested to know where is the catch :D | |
| Jan 7, 2021 at 9:20 | comment | added | Thomas | Maybe one should focus on why E does not work for the binary step function. If we call n_1,n_2 the hidden neurons, what the first layer is doing is dividing the space into 4, and assigning a different tuple (0,1),(1,0), (1,1), (0,0) to (n_1,n_2) in each of the 4 regions. The last output layer should just map these tuples to the desired output. It looks that this is impossible, but I do not see at the moment any difference between the bipolar neuron and the standard one since in (n1,n_2) space the points (0,1),(1,0), (1,1), (0,0) and (-1,1),(1,-1), (1,1), (-1,-1) are arranged similarly. | |
| Jan 7, 2021 at 7:54 | comment | added | Igor F. | Are there any other assumptions we need to make about the neurons? By "the input" you simply mean the sum of all inputs? And each individual input is simply the output of the previous neuron multiplied by the weight? Do the neurons have a bias term? | |
| Jan 7, 2021 at 7:51 | history | edited | Igor F. |
Removed python tag. The question is not software-specific.
|
|
| Jan 7, 2021 at 7:39 | history | edited | user307505 |
edited tags
|
|
| Jan 7, 2021 at 6:36 | history | edited | user307505 |
edited tags
|
|
| S Jan 6, 2021 at 22:49 | history | suggested | Davied Zuhraph | CC BY-SA 4.0 |
some clarification was added
|
| Jan 6, 2021 at 22:48 | review | Suggested edits | |||
| S Jan 6, 2021 at 22:49 | |||||
| Jan 6, 2021 at 22:39 | history | edited | user307505 | CC BY-SA 4.0 |
added 196 characters in body
|
| Jan 6, 2021 at 22:28 | comment | added | gunes | What does Bipolar mean here? | |
| Jan 6, 2021 at 22:27 | review | First posts | |||
| Jan 7, 2021 at 2:31 | |||||
| S Jan 6, 2021 at 22:25 | history | suggested | Betty Andersson | CC BY-SA 4.0 |
some modification was done
|
| Jan 6, 2021 at 22:24 | review | Suggested edits | |||
| S Jan 6, 2021 at 22:25 | |||||
| Jan 6, 2021 at 22:22 | history | asked | user307505 | CC BY-SA 4.0 |