RestingpotentialandNernstpotential:
1. Whatdoeseachreferto?
The Nernst potential isdefined as the equilibrium potential of any ion that is present on both
sides of a membranepermeable to that ion. Quantitatively it corresponds tothe potential we
should ap

CNS 187: Homework 0
Fall 2015
This homework set is meant to help you assess your knowledge of prerequisite concepts. This
course requires both a basic ability to code (section 1) as well as a background in a variety of
mathematical areas (section 2). If y

CNS 187: Homework 2 Answer Key
Fall 2015
1
Integrate and Fire Toilet
1. Consider the main elements of the flush toilet (http:/en.wikipedia.org/wiki/
Flush_toilet). Briefly, water enters the tank through the tank fill tube. The float
shuts off the incoming

Yang Liu
CNS 187 -Fall 2014 -Homework 3
Answer Keys
1. Poisson neurons and diffuse-to-bounds decision making: Derive the sequential
probability ratio test, and explain how it could be implemented (or approximated) by a
neuron or group of neurons.
SPRT sta

CNS 187: Homework 2
Fall 2015
Released: Tuesday, October 13, 2015
Due: Wednesday, October 21, 2015, 11:59pm
Upload a PDF file with your answers and names of collaborators.
Each question is weighted equally for grading purposes.
1
Integrate and Fire To

Predicting response time and error rates in visual
search
Bo Chen
Caltech
bchen3@caltech.edu
Vidhya Navalpakkam
Yahoo! Research
nvidhya@yahoo-inc.com
Pietro Perona
Caltech
perona@caltech.edu
Abstract
A model of human visual search is proposed. It predicts

CNS 187: Homework 1
Fall 2015
Released: Tuesday, October 6, 2015
Due: Wednesday, October 14, 2015, 11:59pm
Upload a PDF file with your answers and names of collaborators.
Each question is weighted equally for grading purposes.
1
Potentials
1. Define t

1
Neural Networks and the Backpropagation Algorithm
1.1
Overview
Neural Networks (NN) are a machine learning computational model used to estimate or approximate functions depending on a large number of inputs. We represent neural networks as a set of
inte

Lecture 4 - Decision
CNS 187 - Fall 2016
P. Perona - Caltech
Two alternative forced choices
Press alarm?
Two alternative forced choices
Time for a
beer?
Two types of error
Response
Stimulus
False alarm
Miss
One-neuron model
One-neuron model
input x
output

CNS 187: Homework 4
Fall 2015
Released: Tuesday, October 27, 2015
Due: Wednesday, November 4, 2015, 11:59pm
Upload a compressed folder containing a PDF file with your answers and names of
collaborators as well as your code.
Each question is weighted e

CNS 187: Homework 5
Fall 2015
Released: Tuesday, November 3, 2015
Due: Wednesday, November 11, 2015, 11:59pm
Upload a compressed folder containing a PDF file with your answers and names of
collaborators as well as your code.
Each question is weighted

CNS 187: Homework 7 Answer Key
Fall 2015
1
Recurrent Neural Networks
1. Explain how to use a Hopfield network as content addressable memory, storing a single
pattern of +1s and -1s. Use equations as necessary in your explanation.
Solution: To store a sing

CNS 187: Homework 6
Fall 2015
Released: Tuesday, November 10, 2015
Due: Wednesday, November 18, 2015, 11:59pm
Upload a compressed folder containing a PDF file with your answers and names of
collaborators as well as your code.
Each question is weighted

CNS 187: Homework 7
Fall 2015
Released: Tuesday, November 17, 2015
Due: Wednesday, November 25, 2015, 11:59pm
Upload a compressed folder containing a PDF file with your answers and names of
collaborators as well as your code.
Questions 1 and 2 are eac

Lecture 4 - Decision
CNS 187 - Fall 2015
P. Perona - Caltech
Two alternative forced choices
Press alarm?
Two alternative forced choices
Time for a
beer?
Two types of error
Response
Stimulus
False alarm
Miss
One-neuron model
One-neuron model
input x
output

Integrate and fire
model
Lecture 3 - CNS187 - Fall 2011
P. Perona
Figures from:
S. Seungs lecture notes http:/hebb.mit.edu/courses/9.641/lectures/index.html
Kochs biophysics book `Biophysics of Computation: Information Processing in Single Neurons
Wikiped