vault backup: 2023-05-31 22:21:56

Affected files:
.obsidian/global-search.json
.obsidian/workspace.json
Health/Alexithymia.md
Health/BWS.md
STEM/AI/Neural Networks/Activation Functions.md
STEM/AI/Neural Networks/Architectures.md
STEM/AI/Neural Networks/CNN/CNN.md
STEM/AI/Neural Networks/MLP/Back-Propagation.md
STEM/AI/Neural Networks/Transformers/Attention.md
STEM/CS/Calling Conventions.md
STEM/CS/Languages/Assembly.md

AI/Neural Networks/Activation Functions.md

@@ -11,7 +11,7 @@
 - Bipolar
 	- -1 <-> +1
 
-![[threshold-activation.png]]
+![threshold-activation](../../img/threshold-activation.png)
 
 # Sigmoid
 -   Logistic function
@@ -26,7 +26,8 @@ $$\frac d {dx} \sigma(x)=
 \right]
 =\sigma(x)\cdot(1-\sigma(x))$$
 
-![[sigmoid.png]]
+![sigmoid](../../img/sigmoid.png)
+
 ### Derivative
 
 $$y_j(n)=\varphi_j(v_j(n))=
@@ -58,7 +59,7 @@ Rectilinear
 	- Sometimes small scalar for negative
 		- Leaky ReLu
 
-![[relu.png]]
+![relu](../../img/relu.png)
 
 # SoftMax
 - Output is per-class vector of likelihoods

AI/Neural Networks/Architectures.md

@@ -2,7 +2,7 @@
 - *Acyclic*
 - Count output layer, no computation at input
 
-![[feedforward.png]]
+![feedforward](../../img/feedforward.png)
 
 # Multilayer Feedforward
 - Hidden layers
@@ -12,12 +12,12 @@
 - Fully connected
 	- Every neuron is connected to every neuron adjacent layers
 - Below is a 10-4-2 network
-![[multilayerfeedforward.png]]
+![multilayerfeedforward](../../img/multilayerfeedforward.png)
 
 # Recurrent
 - At least one feedback loop
 - Below has no self-feedback
-![[recurrent.png]]
-![[recurrentwithhn.png]]
+![recurrent](../../img/recurrent.png)
+![recurrentwithhn](../../img/recurrentwithhn.png)
 
 - Above has hidden neurons

AI/Neural Networks/CNN/CNN.md

@@ -5,13 +5,13 @@
 - Niche
 	- No-one cared/knew about CNNs
 ## After
-- [[Datasets#ImageNet|ImageNet]]
+- [ImageNet](../CV/Datasets.md#ImageNet)
 	- 16m images, 1000 classes
 - GPUs
 	- General processing GPUs
 	- CUDA
 - NIPS/ECCV 2012
-	- Double digit % gain on [[Datasets#ImageNet|ImageNet]] accuracy
+	- Double digit % gain on [ImageNet](../CV/Datasets.md#ImageNet) accuracy
 
 # Full Connected
 [[MLP|Dense]]

AI/Neural Networks/MLP/Back-Propagation.md

@@ -79,7 +79,7 @@ $$\Delta w_{ji}(n)=\eta\cdot\delta_j(n)\cdot y_i(n)$$
 	2.  Error WRT output $y$
 	3.  Output $y$ WRT Pre-activation function sum
 	4.  Pre-activation function sum WRT weight
-		-   Other [[Weight Init|weights]] constant, goes to zero
+		-   Other [weights](../Weight%20Init.md) constant, goes to zero
 		-   Leaves just $y_i$
 	-   Collect 3 boxed terms as delta $j$
 		-   Local gradient

AI/Neural Networks/Transformers/Attention.md

@@ -10,16 +10,16 @@
 	- [LSTM](../RNN/LSTM.md) tends to poorly preserve far back [knowledge](../Neural%20Networks.md#Knowledge)
 - Attention layer access all previous states and weighs according to learned measure of relevance
 	- Allows referring arbitrarily far back to relevant tokens
-- Can be addd to [[RNN]]s
-- In 2016, a new type of highly parallelisable _decomposable attention_ was successfully combined with a [[Architectures|feedforward]] network
-	- Attention useful in of itself, not just with [[RNN]]s
-- [[Transformers]] use attention without recurrent connections
+- Can be addd to [RNNs](../RNN/RNN.md)
+- In 2016, a new type of highly parallelisable _decomposable attention_ was successfully combined with a [feedforward](../Architectures.md) network
+	- Attention useful in of itself, not just with [RNNs](../RNN/RNN.md)
+- [Transformers](Transformers.md) use attention without recurrent connections
 	- Process all tokens simultaneously
 	- Calculate attention weights in successive layers
 
 # Scaled Dot-Product
 - Calculate attention weights between all tokens at once
-- Learn 3 [[Weight Init|weight]] matrices
+- Learn 3 [weight](../Weight%20Init.md) matrices
 	- Query
 		- $W_Q$
 	- Key

CS/Calling Conventions.md

@@ -5,15 +5,15 @@
 	- Also known as: callee-saved registers or non-volatile registers
 - How the task of preparing the stack for, and restoring after, a function call is divided between the caller and the callee
 
-Subtle differences between [[compilers]], can be difficult to interface codes from different [[compilers]]
+Subtle differences between [Compilers](Compilers.md), can be difficult to interface codes from different [compilers](Compilers.md)
 
-Calling conventions, type representations, and name mangling are all part of what is known as an [application binary interface](https://en.wikipedia.org/wiki/Application_binary_interface) ([[ABI]])
+Calling conventions, type representations, and name mangling are all part of what is known as an [application binary interface](https://en.wikipedia.org/wiki/Application_binary_interface) ([ABI](ABI.md))
 
 # cdecl
 C declaration
 
-- Originally from Microsoft's C [[compilers|compiler]]
-	- Used by many C [[compilers]] for x86
+- Originally from Microsoft's C [compiler](Compilers.md)
+	- Used by many C [compilers](Compilers.md) for x86
 - Subroutine arguments passed on the stack
 - Function arguments pushed right-to-left
 	- Last pushed first

CS/Languages/Assembly.md

@@ -1,11 +1,11 @@
 [Uni of Virginia - x86 Assembly Guide](https://www.cs.virginia.edu/~evans/cs216/guides/x86.html)
 
 ## x86 32-bit
-![[x86registers.png]]
+![x86registers](../../img/x86registers.png)
 
 ## Stack
 - push, pop, call, ret
 
-![[stack.png]]
+![stack](../../img/stack.png)
 - Growing upwards