Posts

Showing posts with the label Current

A practical way to view neural networks (whiteboard view)

Image
 A whiteboard view: You can take a decision matrix view of ReLU neural networks. Instead of ReLU(Wx) you create a diagonal matrix with binary 1 or 0 entries according to the ReLU decisions (x>=0?) Then conceptually a layer is DWx. Where D is doing row selection on W. And in fact D does column selection on the weight matrix in the next layer. That is very coarse parameter selection which is linear mapping selection.

Current Mirror Radio Circuits

Image
  A Collection of mostly current mirror regenerative radio circuits A current mirror circuit is an electronic circuit that replicates a current flowing through one active device (such as a transistor) by controlling the current in another active device. This is achieved by ensuring that the two devices have the same base-emitter or gate-source voltage. As a result, the output current of the current mirror is a replica of the input current, with a high output resistance and a low input resistance. Current mirror circuits are commonly used in analog circuits to provide bias currents and active loads. Bias currents are used to set the operating point of transistors, while active loads are used to provide a stable load impedance for amplifiers. Current mirrors can also be used to create current sources, which are circuits that provide a constant current regardless of the load. Basic Current Mirror. In radio circuits, the current mirror effect can be exploited at DC because inductors ha...