MIT introduction to EECS I video note

21:50 2013-7-14 Sunday
start 6.01 introduction to EECS I, lec1
22:09 2013-7-14
PCAP == Primitive + Combination + Abstraction + Pattern
22:13 2013-7-14
resistive networks
22:15 2013-7-14
photoresistor
23:02 2013-7-14
interpreter-based
23:05 2013-7-14
compositional
23:05 2013-7-14
the system is compositional
23:26 2013-7-14
heterogeneous list
23:26 2013-7-14
lists of lists
23:52 2013-7-14
class, attribute, method
23:56 2013-7-14
subclass
23:58 2013-7-14
ground rules
23:58 2013-7-14
binding environment
0:02 2013-7-15
global environment
0:10 2013-7-15
a binding environment associates
a name with a thing

22:18 2013-7-15 Monday
do nothing...
///
10:10 2013-7-20 Saturday
review 6.01 video 1
10:10 2013-7-20
modes of reasoning
11:05 2013-7-20
PCAP == Primitive + Combination + Abstraction + Pattern
13:02 2013-7-20
binding environment
13:12 2013-7-20
procedure name <-> procedure
13:20 2013-7-20
calling a procedure creates a new environment
13:25 2013-7-20
calling environment
13:25 2013-7-20
parent environment
13:28 2013-7-20
destroy this environment
**************************************
14:06 2013-7-20
review 6.01 video 2, PCAP
14:13 2013-7-20
functional programming
14:13 2013-7-20
eat input, generate output, without side effect
14:21 2013-7-20
imperative(procedural) programming
14:36 2013-7-20
functional programming,
pure function: have no side-effect
14:38 2013-7-20
functional programming:
treat procedures as 1st class object
14:43 2013-7-20
lists of lists
14:54 2013-7-20
recursion
15:52 2013-7-20
OOP == Object-Oriented Programming
16:00 2013-7-20
controlling processes:
we need a different kind of abstraction:
state machine(SM)
16:03 2013-7-20
modularity
16:09 2013-7-20
state-transition diagram
16:22 2013-7-20
transition table
16:31 2013-7-20
modular decomposition of the robot:
map maker + planner + mover
16:33 2013-7-20
What does modular mean?
sensor input ->map maker-> map
map ->planner->heading
16:37 2013-7-20
SM == State Machine
16:37 2013-7-20
SM class:
start(), step(), transduce()
16:40 2013-7-20
subclass
16:40 2013-7-20
generic state machine class
16:47 2013-7-20
an example of SM: bank account
input: withdraw, deposit
16:59 2013-7-20
combination of SM:
cascade, parallel, feedback
22:40 2013-7-20
start review 6.01 video 3,
signals & systems
22:43 2013-7-20
proportional controller
22:49 2013-7-20
undershoot
22:51 2013-7-20
there is a delay between sensor & actuator
22:53 2013-7-20
the signals & systems abstraction:
input signal ->system-> output signal
22:54 2013-7-20
solve "mass & spring system"
using signals & systems approach
22:57 2013-7-20
input signal: x(t), position of hand
output signal: y(t), position of mass
23:03 2013-7-20
PCAP == Primitive + Combination + Abstraction + Pattern
23:04 2013-7-20
continuous time systems
discrete time systems
23:05 2013-7-20
difference equations
23:07 2013-7-20
unit sample
23:12 2013-7-20
multiple representation of discrete-time systems:
difference equation
block diagram
23:15 2013-7-20
signal flow path
23:15 2013-7-20
We have to start the system in a state
23:15 2013-7-20
What does "initially at rest" mean?
all the Delays have zero output......
23:20 2013-7-20
the system ticks!
23:28 2013-7-20
comparison between "Difference equation" &
"Block diagram"?
Difference equation is declarative,
Block diagram is imperative,
Block diagram contains more information than Difference equation
e.g: which is input & which is output?
23:31 2013-7-20
from "samples" to "signals"
23:31 2013-7-20
signal is the single object
23:32 2013-7-20
Operators manipulates signals rather than
individual samples
23:33 2013-7-20
R == the right-shift operator
23:35 2013-7-20
operators work on signals
23:47 2013-7-20
operator algebra
23:59 2013-7-20
equivalent system
0:09 2013-7-21
feedforward
feedback(have cycles)
0:09 2013-7-21
feedback loop
0:17 2013-7-21
recipe <-> constraint
0:19 2013-7-21
the output signal persist
0:24 2013-7-21
system functional
0:26 2013-7-21
synthetic division
0:27 2013-7-21
numerator / denominator
0:33 2013-7-21
adders, gains, delays
0:33 2013-7-21
LCCDE == Linear Constant Coefficient Difference Equation
0:34 2013-7-21
imperative rule
0:37 2013-7-21
multiple representation of systems
1. difference equation
2. block diagram
3. operators
///
22:18 2013-7-21 Sunday
back from home
22:18 2013-7-21
desired temperature, actual temperature,
thermostat
22:19 2013-7-21
regulation
22:24 2013-7-21
controller, plant, sensor
22:27 2013-7-21
characterizing system performance
22:29 2013-7-21
sample approach
22:41 2013-7-21
initial condition
22:45 2013-7-21
operator equations
22:45 2013-7-21
system functional
//
19:38 2013-7-22 Monday
watching friends videos
21:23 2013-7-22
review 6.01 video 5,
characterizing system performance
21:24 2013-7-22
glucose delivery
21:30 2013-7-22
feedback & control
21:36 2013-7-22
"samples approach"
21:38 2013-7-22
operator equations
21:40 2013-7-22
reducing complexity
21:40 2013-7-22
system functional
21:59 2013-7-22
faster & oscillatory
22:04 2013-7-22
signal approach <-> sample approach
22:05 2013-7-22
power series
22:06 2013-7-22
signal flow paths
22:07 2013-7-22
persistent response
22:08 2013-7-22
feedback: cyclic signal flow paths
22:09 2013-7-22
feedback: even transient signal generates reponse forever
22:09 2013-7-22
loop back
22:10 2013-7-22
feedforward: acyclic paths
feedback: cyclic paths
22:16 2013-7-22
geometric growth
22:17 2013-7-22
pole: the base of the geometric sequence
22:42 2013-7-22
growing without bound
22:42 2013-7-22
unit sample response: h[n]
22:42 2013-7-22
2nd order system
22:44 2013-7-22
initially at rest(initial rest):
the initial output of the R is zero
// R == Right Shift operator
22:45 2013-7-22
now the clock ticks
22:46 2013-7-22
operator R means delay....
22:52 2013-7-22
equivalent systems
22:53 2013-7-22
partial fractions
22:53 2013-7-22
partial fraction expansion
22:54 2013-7-22
sum of geometric sequences
22:58 2013-7-22
2nd order systems
22:58 2013-7-22
higher-order systems
22:59 2013-7-22
LCCDE == Linear Constant-Coefficient Difference Equation,
///
Linear Difference Equation with Constant-Coefficient
22:59 2013-7-22
Rational polynomial
23:00 2013-7-22
factor theorem
23:01 2013-7-22
sum of geometrics
23:01 2013-7-22
poles: bases
23:01 2013-7-22
you do not know the coefficient, but you know
the basic shape of the response....
23:02 2013-7-22
fundamental theorem of algebra
23:12 2013-7-22
poles: roots of denominator polynomial
23:16 2013-7-22
complex poles
23:16 2013-7-22
nonzero imaginary parts
23:16 2013-7-22
ratio of polynomials
23:17 2013-7-22
fundamental mode
23:17 2013-7-22
complex poles: that base is complex
23:22 2013-7-22
modes: geometric sequence
23:23 2013-7-22
complex poles,
complex modes
23:26 2013-7-22
real coefficient -> complex poles
23:26 2013-7-22
complex poles always come with pair,
complex conjugate // symmetric about real axis
23:28 2013-7-22
real-valued coefficient
23:33 2013-7-22
break the pole into magnitude & angle
23:35 2013-7-22
I've separate the magnitude & angle
23:38 2013-7-22
poles & modes
23:39 2013-7-22
sequence <-> series
23:39 2013-7-22
population growth
23:41 2013-7-22
the signals & systems view of Fibonacci sequence
y[n] = x[n] + y[n-1] + y[n-2]
// x[n] == delta[n], unit sample
// x[n] is the input signal
23:47 2013-7-22
we found a whole new way of thinking the
Fibonacci sequence problem with poles!
23:47 2013-7-22
powerful abstraction: poles!

20:00 2013-7-23 Tuesday
designing control systems
20:05 2013-7-23
feedback -> cyclic path -> pole
20:06 2013-7-23
poles are just bases for geometric sequence
20:10 2013-7-23
black's equation
20:12 2013-7-23
forward gain
loop gain
closed-form loop gain
20:14 2013-7-23
accumulator
20:14 2013-7-23
equivalent block diagram
20:21 2013-7-23
step response // lab measure
unit sample response // analytical
20:23 2013-7-23
unit sample ->accumulator-> unit step
20:24 2013-7-23
h[n] // unit sample response
s[n] // unit step response
20:29 2013-7-23
fastest convergence.....
20:33 2013-7-23
unit sample response
20:44 2013-7-23
sensor with delay
20:47 2013-7-23
delay in a sensor path
20:57 2013-7-23
substitute 1/z for R
20:58 2013-7-23
geometric mode
21:00 2013-7-23
slow mode(pole near 1) dominates the response
21:01 2013-7-23
dominant pole
21:03 2013-7-23
complex poles -> oscillations
21:08 2013-7-23
root locus
21:34 2013-7-23
pole -> base of geometric sequence
23:02 2013-7-23
cyclic signal flow paths -> persistent reponses to transient inputs
23:02 2013-7-23
modes: response to a single pole
23:05 2013-7-23
transient signal -> persistent response
23:06 2013-7-23
inner loop
23:06 2013-7-23
accumulator
23:09 2013-7-23
system function
23:09 2013-7-23
forward gain
loop gain
closed-loop gain
23:11 2013-7-23
black's equation
23:12 2013-7-23
equivalent system
23:12 2013-7-23
black's equation:
closed-loop gain == forward gain / (1 - loop gain)
23:18 2013-7-23
mode: natrual response to the pole
23:19 2013-7-23
unit sample response, step response
23:20 2013-7-23
desired distance, actual distance
23:22 2013-7-23
unit sample response: h[n]
->accumulator->
step response: s[n]
23:24 2013-7-23
accumulator
23:27 2013-7-23
you can infer properties of step response from
properties of poles
23:28 2013-7-23
monotonic
23:28 2013-7-23
alternating
23:29 2013-7-23
converge <-> diverge
23:30 2013-7-23
best performance(fastest response)?
which kind of pole?
23:32 2013-7-23
unit-sample response
23:38 2013-7-23
desiredFront, distanceFront
23:39 2013-7-23
overshoot, undershoot
23:39 2013-7-23
adding sensory delay
23:40 2013-7-23
sensor introduce delay
23:42 2013-7-23
adding delay tends to destabilize control systems
23:45 2013-7-23
system functional: about R(delay)
system function: about 1/z == R
23:49 2013-7-23
quadratic equation
23:50 2013-7-23
adders, gains, delays
23:50 2013-7-23
partial fraction expansion
23:51 2013-7-23
geometric mode
23:51 2013-7-23
feedback & control: poles
23:52 2013-7-23
system response
23:55 2013-7-23
dominant pole
23:57 2013-7-23
transient input -> persistent response // because feedback
23:57 2013-7-23
complex poles -> oscillations
23:59 2013-7-23
poles with imaginary part
0:04 2013-7-24
root locus
0:05 2013-7-24
closed-loop poles
0:06 2013-7-24
1st order response
0:07 2013-7-24
slowest pole(closest to the unit circle)
0:10 2013-7-24
delays are bad generally
0:10 2013-7-24
adding delay to destabilize the system
0:11 2013-7-24
monotonic <-> alternating
converging <-> diverging
0:12 2013-7-24
unit sample response is the sum of geometric sequences
0:22 2013-7-24
equally dominant
0:23 2013-7-24
dominant pole: the pole that is worstly behaved
0:25 2013-7-24
the unit sample response associated with this pole
0:27 2013-7-24
periodic
0:34 2013-7-24
initial rest(initially at rest):
all outputs of Delays(R) are zero
0:35 2013-7-24
clock ticks!

22:31 2013-7-24 Wednesday
review 6.01 lec 7, circuits
22:44 2013-7-24
PCAP == Primitive + Combination + Abstraction + Pattern
22:44 2013-7-24
state machine
22:45 2013-7-24
resistive network
22:46 2013-7-24
light tracking system
22:50 2013-7-24
circuit abstraction
22:51 2013-7-24
circuit idea is very general
22:52 2013-7-24
across variable // voltage
throught variable // current
22:54 2013-7-24
circuits are also used as "models of things"
22:56 2013-7-24
biological system <-> circuits
22:57 2013-7-24
primitives
22:57 2013-7-24
system functional -> sytem function(z = 1/R)
22:58 2013-7-24
delay, gain, adder
22:58 2013-7-24
rules of combination
23:00 2013-7-24
voltage source, current source
23:06 2013-7-24
more structured way of thinking the results
23:09 2013-7-24
current source
23:14 2013-7-24
any closed path
23:15 2013-7-24
reference direction
23:28 2013-7-24
planer network
23:29 2013-7-24
inner loop
23:29 2013-7-24
linearly independent
23:30 2013-7-24
KVL == Kirchoff's Voltage Law
KCL == Kirchoff's Current Law
23:30 2013-7-24
node
23:31 2013-7-24
charge build up
23:36 2013-7-24
do not allow the buildup of electrons
23:37 2013-7-24
buildup
23:41 2013-7-24
linearly independent
23:43 2013-7-24
element current
23:47 2013-7-24
How many linearly independent KCL equations?
N-1 // N == Node
23:51 2013-7-24
basic method(2b method):
element law: B
KCL: N-1
KVL: B-(N-1) == B-N+1
23:54 2013-7-24
supernode
23:54 2013-7-24
closed surface
23:56 2013-7-24
generalization of the KCL equation
23:57 2013-7-24
super KCL
23:57 2013-7-24
closed region
23:59 2013-7-24
element law // constituent relation
0:01 2013-7-25
node analysis(node method)
0:01 2013-7-25
node voltage
0:02 2013-7-25
element voltage <-> node voltage
0:02 2013-7-25
reference voltage(gnd)
0:03 2013-7-25
gnd == ground
0:09 2013-7-25
loop current
0:20 2013-7-25
series combination
/
19:34 2013-7-25 Thursday
start review eecs I lec 8,
op-amps
19:37 2013-7-25
N == Node,
B == Branch
19:38 2013-7-25
element voltage <-> node voltage
19:52 2013-7-25
NVCC == Node Voltage with Component Current
20:06 2013-7-25
buffer
20:06 2013-7-25
op-amp
20:08 2013-7-25
dependent source
20:08 2013-7-25
CCCS == Current Controlled Current Source
20:12 2013-7-25
there's no complicated coupling between the two
20:16 2013-7-25
port, terminal
20:17 2013-7-25
input port, output port
20:18 2013-7-25
independent source
dependent source(controlled source)
20:18 2013-7-25
op-amp: VCVS(Voltage Controlled Voltage Source)
20:23 2013-7-25
non-inverting amplifier
20:24 2013-7-25
Op-Amp
20:28 2013-7-25
ideal Op-Amp
20:32 2013-7-25
open-loop gain
20:43 2013-7-25
inverting summer
21:05 2013-7-25
1. ideal Op-Amp assumption,
2. VCVS model
21:11 2013-7-25
but what is the mechanism?
21:13 2013-7-25
charge accumulation
21:13 2013-7-25
accurate representation <-> model
21:15 2013-7-25
schematic diagram
21:17 2013-7-25
dynamic analysis of Op-Amp
21:22 2013-7-25
stable feedback(negative feedback)
21:24 2013-7-25
power rails
21:24 2013-7-25
power pins
21:41 2013-7-25
other ways to achieving modularity in circuit design
21:53 2013-7-25
primitive element
21:56 2013-7-25
straight-line constraints
21:57 2013-7-25
current-voltage relationship
22:06 2013-7-25
reference direction
22:09 2013-7-25
i-v curve
22:10 2013-7-25
horizontal sum
22:12 2013-7-25
linear resistor
22:13 2013-7-25
component equations
22:14 2013-7-25
system of linear equations
22:17 2013-7-25
Thevenin Equivalents
22:17 2013-7-25
read off some critical numbers from this plot
22:21 2013-7-25
Norton equivalent
22:26 2013-7-25
linear circuits:
linear resistor + voltage source + current source
22:27 2013-7-25
Voc == open circuit voltage, // current is zero
22:28 2013-7-25
Isc == short circuit current, // voltage is zero
22:31 2013-7-25
Rth == thevenin equivalent circuit
22:32 2013-7-25
equivalent in the sense that they all
share the same i-v curve
22:34 2013-7-25
identical in the sense that they have
the same v-i curve
22:49 2013-7-25
split the circuit into 2 pieces
22:52 2013-7-25
pratical value & conceptual simplicity
22:53 2013-7-25
superposition
22:54 2013-7-25
drive(-> response)
23:02 2013-7-25
current divider
23:11 2013-7-25
finish introduction to EECS I circuits part,
///
15:33 2013-8-2 Friday
introduction to EECS I, lec 10
mapping + localization + planning
15:42 2013-8-2
probability theory
15:42 2013-8-2
uncertainty
15:43 2013-8-2
infer .. from unreliable observations

16:33 2013-8-4 Sunday
introduction to EECS I, lec 10
16:33 2013-8-4
mapping + localization + planning
16:48 2013-8-4
probability theory
17:08 2013-8-4
atomic event
17:08 2013-8-4
finest grain
17:09 2013-8-4
sample space
17:10 2013-8-4
collectively exhaustive
17:19 2013-8-4
enumerate the sample space
17:19 2013-8-4
atomic events
17:19 2013-8-4
compound event
17:20 2013-8-4
conditional probability
17:47 2013-8-4
distribution
17:47 2013-8-4
random variable
17:48 2013-8-4
event <-> probability
17:48 2013-8-4
atomic events
17:48 2013-8-4
mutually exclusive & collectively exhaustive
17:51 2013-8-4
joint probability distribution
17:55 2013-8-4
reducing dimensionality:
1. marginalizing
2. conditioning (apply Bayes' rule)
17:57 2013-8-4
joint probability distribution
18:01 2013-8-4
marginalization
18:11 2013-8-4
Random Variable:
event <-> probability
use dictionary to represent this....
18:12 2013-8-4
DDist == Discrete Distribution
18:14 2013-8-4
dictionary: key + value
18:17 2013-8-4
conditional probability
18:20 2013-8-4
joint probability distribution
18:24 2013-8-4
Markov models
18:25 2013-8-4
state transition
18:35 2013-8-4
maximum likelyhood
21:37 2013-8-4
modeling stage, prototyping stage
21:41 2013-8-4
uncertainty
21:47 2013-8-4
unreliable observation -> reliable inference
21:56 2013-8-4
event
21:56 2013-8-4
atomic event (finest grain)
21:57 2013-8-4
coin toss experiment
21:58 2013-8-4
mutually exclusive,
collectively exhaustively
21:59 2013-8-4
sample space
21:59 2013-8-4
how to chunk outcomes into a sample space
22:00 2013-8-4
probability
22:01 2013-8-4
intersection, union
22:02 2013-8-4
Pr == Probability
22:05 2013-8-4
enumerate the sample space, find the "event of interest"
22:06 2013-8-4
conditional probability
22:07 2013-8-4
Bayes' rule
22:07 2013-8-4
we normalize the intersection by the probability of B
22:09 2013-8-4
to shrink the universe
22:12 2013-8-4
occupies greater part of the universe
22:14 2013-8-4
atomic units
22:15 2013-8-4
event of interest
22:19 2013-8-4
conditioning
22:19 2013-8-4
conditioning can increase or decrease
the probability of an event
22:20 2013-8-4
Random Variable(distribution):
event <-> probability
22:22 2013-8-4
joint probability distribution
22:23 2013-8-4
multi-dimensional sample space
22:24 2013-8-4
P(V) // Random Variable
P(v) // events
22:24 2013-8-4
Reducing dimensionality:
1. marginalizing
2. conditioning
22:29 2013-8-4
marginalizing: sum up + rescaling
22:38 2013-8-4
conditioning
22:39 2013-8-4
conditioned on the fact that ....
22:39 2013-8-4
conditioning event
22:41 2013-8-4
marginal probability
22:43 2013-8-4
DDist == Discrete Distribution
22:43 2013-8-4
event(key) <-> probability(value)
22:43 2013-8-4
dictionary
22:44 2013-8-4
Random Variable: probability distribution
event <-> probability
22:45 2013-8-4
atomic event <-> probability
22:52 2013-8-4
joint probability distribution
22:52 2013-8-4
JDist == Joint Distribution
/
21:16 2013-8-5 Monday
discrete probability & state estimation
21:20 2013-8-5
modeling stage <-> prototyping stage
21:21 2013-8-5
mapping + localization + planning
21:24 2013-8-5
backtrack (try to get out of here)
21:27 2013-8-5
probability theory
21:28 2013-8-5
uncertainty
21:28 2013-8-5
draw reliable inference from unreliable observance
21:30 2013-8-5
lego brick
21:39 2013-8-5
Random Variable:
event <-> probability
21:39 2013-8-5
event
atomic event (finest grain)
21:43 2013-8-5
mutually exclusive
21:43 2013-8-5
sample space
21:45 2013-8-5
union & intersection
21:48 2013-8-5
coin toss
21:49 2013-8-5
event of interest
21:49 2013-8-5
compound event
21:51 2013-8-5
conditional probability
21:51 2013-8-5
Bayes' rule
21:52 2013-8-5
normalize
21:53 2013-8-5
condition on B == shrink the universe
21:58 2013-8-5
the universe shrank
21:58 2013-8-5
rescale my probability
22:03 2013-8-5
Random Variable:
event <-> probability
22:03 2013-8-5
Random Variable represents a "distribution"
22:04 2013-8-5
mutually exclusive & collectively exhaustive
22:05 2013-8-5
mathematical representation
22:05 2013-8-5
multidimensional random variable
22:07 2013-8-5
joint probability distribution
22:07 2013-8-5
Random Variable: W
Event: w
22:08 2013-8-5
it's just a convenient notation to write
more complicated things concisely.
22:09 2013-8-5
Reducing dimensionality:
1. marginalizing (summing out)
2. conditioning (rescale)
22:10 2013-8-5
joint probability space
22:19 2013-8-5
marginalization
22:24 2013-8-5
relative likelihood
22:27 2013-8-5
DDist == Discrete Distribution
key(event) <-> value(probability)
22:33 2013-8-5
conditional probability
22:36 2013-8-5
JDist == joint probability distribution
22:39 2013-8-5
Markov model
22:40 2013-8-5
state distribution
22:40 2013-8-5
state transition
22:45 2013-8-5
mathematical expectation
22:45 2013-8-5
maximum likelihood
22:53 2013-8-5
state probability
///
21:27 2013-8-6 Tuesday
start lec 12 of introduction to EECS I
state estimation
21:32 2013-8-6
planning(search)
21:34 2013-8-6
eight puzzles
21:34 2013-8-6
free spot
21:35 2013-8-6
start state -> goal state
21:41 2013-8-6
solve in a systematic way
21:44 2013-8-6
construct the tree on the fly
21:45 2013-8-6
incrementally build the tree
21:47 2013-8-6
goalTest
21:48 2013-8-6
successor,
state, action, newState
21:51 2013-8-6
SearchNode class
21:54 2013-8-6
search through the tree
21:57 2013-8-6
agenda:
1. initialize agenda
2. remove node from front
3. add children node at back
21:57 2013-8-6
successor subroutine
22:02 2013-8-6
parent <-> children
22:07 2013-8-6
pull that guy out, and replace with its children
22:10 2013-8-6
DFS == Depth First Search
22:14 2013-8-6
agenda, DFS, BFS
22:18 2013-8-6
order matters,
stack, DFS
queue, BFS
22:20 2013-8-6
LIFO == Last-In First-Out
22:24 2013-8-6
implement DFS with stack()
22:25 2013-8-6
implement BFS with queue()
22:29 2013-8-6
terminal node
22:35 2013-8-6
pruning rule
22:40 2013-8-6
pruning rule II
22:52 2013-8-6
dynamic programming
22:58 2013-8-6
BFS + dynamic programming
23:01 2013-8-6
optimal
/
23:29 2013-8-7 Wednesday
review introduction to EECS I, lec 12,
Search
23:30 2013-8-7
state estimation problems
23:33 2013-8-7
start state -> goal state
23:40 2013-8-7
board configuration
23:53 2013-8-7
sucessor: current state + action == next state
23:53 2013-8-7
construct the tree on the fly
23:54 2013-8-7
goalTest
0:00 2013-8-8
agenda
0:04 2013-8-8
terminating condition:
goal is met,
agenda is empty
0:05 2013-8-8
DFS -> stack
BFS -> queue
0:11 2013-8-8
order matters
0:18 2013-8-8
I'm going deeper & deeper into the tree
without fully exploring all the horizontal spaces.
0:19 2013-8-8
DFS == Depth First Search
0:21 2013-8-8
decision tree
0:23 2013-8-8
BFS == Breadth First Search
0:24 2013-8-8
stack, queue
0:29 2013-8-8
pseudo procedure -> real procedure
0:32 2013-8-8
tree of all possible paths
0:35 2013-8-8
walking around manhattens
0:40 2013-8-8
pruning rule I
0:47 2013-8-8
LIFO == stack
FIFO == queue
0:51 2013-8-8
BFS + memorization
0:58 2013-8-8
dynamic programming
0:58 2013-8-8
keep track of the node that have already visited
0:58 2013-8-8
BFS with dynamic programming
1:01 2013-8-8
agenda list, visit list
1:06 2013-8-8
pruning rule
1:07 2013-8-8
done!
/
20:43 2013-8-8 Thursday
start introduction to EECS I, lec 13,
optimizing search algorithm
20:46 2013-8-8
queue-based ordering
breadth-first search(BFS)
20:48 2013-8-8
uniform cost search
20:49 2013-8-8
action costs
20:51 2013-8-8
BFS + dynamic programming
20:52 2013-8-8
agenda list + visited list
20:59 2013-8-8
path cost
20:59 2013-8-8
priority queue
21:01 2013-8-8
pq == priority queue
21:07 2013-8-8
actionCost
21:12 2013-8-8
uniform cost search
21:15 2013-8-8
ucSearch == uniform cost search
21:16 2013-8-8
dynamic programming list
21:31 2013-8-8
starting point -> goal
21:32 2013-8-8
starting state centric
21:37 2013-8-8
heuristics
21:40 2013-8-8
Manhattan distance
21:41 2013-8-8
the pointer under consideration -> goal
21:50 2013-8-8
admissible heuristic
21:53 2013-8-8
Manhattan distance
21:56 2013-8-8
eight puzzle
///
21:46 2013-8-11 Sunday
uniform cost search
21:53 2013-8-11
BFS + dynamic programming
21:58 2013-8-11
visited list
22:03 2013-8-11
uniform cost search:
priority queue
22:04 2013-8-11
DFS: stack
BFS: queue
22:11 2013-8-11
hop <-> miles
22:19 2013-8-11
expanded list, agenda list
22:30 2013-8-11
starting point -> goal
22:30 2013-8-11
stumble into the goal
22:31 2013-8-11
starting state centric
22:34 2013-8-11
stumbling upon the goal
22:39 2013-8-11
the search is symmetric around E even the goal is not.
22:44 2013-8-11
How do I fix it?
Heuristics
22:46 2013-8-11
the pointer under consideration -> goal
22:48 2013-8-11
Heuristics:
approximate the amount of work we have to do
22:49 2013-8-11
Manhattan distance
22:54 2013-8-11
add Manhattan distance to complete the path to the goal
22:58 2013-8-11
admissible heuristic function
22:58 2013-8-11
do not overestimate!(do not pruning too much)
22:59 2013-8-11
always underestimates the actual distance
23:12 2013-8-11
admissible heuristic

21:43 2013-8-15 Thursday
introduction to EECS I, Recitation 1
OOP
21:45 2013-8-15
class, object
21:45 2013-8-15
method, attribute
21:57 2013-8-15
Recitation 2: inheritance
22:18 2013-8-15
functional programming
22:18 2013-8-15
functions as first-class object
22:20 2013-8-15
function can be
1. return value
2. pass as argument
3. assignment
22:25 2013-8-15
Lambda: anonymous function
22:30 2013-8-15
list manipulation
22:31 2013-8-15
Lambdas + list manipulation
22:32 2013-8-15
list comprehension
22:36 2013-8-15
mutability
22:36 2013-8-15
aliasing
22:37 2013-8-15
that object is changed in place
22:43 2013-8-15
SM == State Machine
22:43 2013-8-15
internal state
22:45 2013-8-15
FSM == Finite State Machine
22:47 2013-8-15
state transition diagram
22:50 2013-8-15
start state
0:25 2013-8-16
go to bed!
//
23:12 2013-8-18 Sunday
poles -> to predict the future
23:29 2013-8-18
feedback -> pole
-> predict the future
23:31 2013-8-18
largest magnitude -> dominant pole
23:34 2013-8-18
unit sample response
23:43 2013-8-18
dorminant pole -> long term behavior
/
21:03 2013-8-19 Monday
feedforward, feedback
21:04 2013-8-19
dorminant pole: pole with greatest magnitude
21:04 2013-8-19
dorminant pole -> long term behavior
21:05 2013-8-19
pole/zero cancellation
21:06 2013-8-19
implementation of real systems
21:07 2013-8-19
repeated roots
21:08 2013-8-19
unit sample response: h[n]
21:15 2013-8-19
converge monotonically
21:36 2013-8-19
circuit diagram
21:55 2013-8-19
current divider
22:02 2013-8-19
NVCC == Node Voltage Component Current
22:02 2013-8-19
NVCC method
22:03 2013-8-19
node voltage
component current
22:07 2013-8-19
super node
22:07 2013-8-19
node analysis <-> NVCC
22:24 2013-8-19
Op-Amp == Operational Amplifier
22:33 2013-8-19
buffer/voltage follower
22:33 2013-8-19
power rails limit range
22:39 2013-8-19
abstract away circuit details
22:39 2013-8-19
superposition
22:40 2013-8-19
Thevenin/Norton Equivalence
22:41 2013-8-19
open-circuit voltage,
short-circuit current
23:00 2013-8-19
probability: model uncertainty
23:00 2013-8-19
autonomy
23:04 2013-8-19
sample space
23:09 2013-8-19
Random variable
23:11 2013-8-19
Joint probability,
conditional probability
23:14 2013-8-19
Bayesian vs frequentist
23:14 2013-8-19
likelihood
23:21 2013-8-19
marginalization total probability
23:25 2013-8-19
state estimation
23:27 2013-8-19
stochastic state machine model
23:27 2013-8-19
starting distribution,
observation distribution,
transition distribution
23:30 2013-8-19
O == Observation,
I == action
23:32 2013-8-19
Bayesian Reasoning
23:58 2013-8-19
S == State,
O == Observation
///
21:24 2013-8-20 Tuesday
search state space
21:25 2013-8-20
autonomous system
21:25 2013-8-20
state transition diagram
21:25 2013-8-20
goal test
21:26 2013-8-20
legal action list
21:27 2013-8-20
search tree
21:29 2013-8-20
node
21:29 2013-8-20
parent node
backtrack to parent node
21:29 2013-8-20
agenda
21:29 2013-8-20
BFS == Breadth First Search // queue based
DFS == Depth First Search // stack based
21:31 2013-8-20
partial path
21:31 2013-8-20
state transition diagram
21:34 2013-8-20
to expand node
21:43 2013-8-20
search:
dynamic programming
21:45 2013-8-20
DFS + DP
21:46 2013-8-20
dynamic programming:
keep a list of visited
21:48 2013-8-20
cost
21:48 2013-8-20
priority queue
21:50 2013-8-20
search + DP + cost + Heuristic
21:57 2013-8-20
admissible heuristic:
nonstrict underestimate
THE END OF 6.01 INTRODUCTION TO EECS I!