NN : Introduction to Neural Network
Q1. Weights
impact
For
the neural network shown above, which of these statements is true?
Choose
the correct answer from below:
A. -5 weight is bad for the neural network.
B. The neuron with weight 10 will have the most impact on the output.
C. The neuron with weight -5 will have the most impact on the output.
D. The neuron with weight 2 will have the most impact on the output.
Ans:
B
Correct
option : The neuron with weight 10 will have the most
impact on the output.
Explanation :
There is no such thing that a neuron with a negative weight will be bad for the
output. The negative or positive weight of a neuron simply means whether it has
an increasing or decreasing effect on the output value. A neuron with the
largest magnitude will have the most significant effect on the output value.
Q2. Calculate
Forward Pass
The
neuron n has the weights 1,2,3,4, and 5. The
values of inputs are 4,10,5,20, and 0. We are using a linear
activation function with the constant of proportionality being equal to 2 here.
The
output will be:
Choose
the correct answer from below:
A. 193.5
B. 59.5
C. 238
D. 119
Ans:
C
Correct
option : 238
Explanation
:
Multiplying
weights with their corresponging inputs, and then adding everything together:
-> Output=(1∗4+2∗10+3∗5+4∗20+5∗0)=119
But
since the output is linear with a proportionality constant 2, hence :
-> Finaloutput=2∗119=238
Q3. Need
for NN
Are
classic ML Algorithms not powerful enough? Why exactly do we need to use Neural
Networks?
Check
all that apply.
Choose
the correct answer from below, please note that this question may have
multiple correct answers
A. Unlike Classic ML Algos, NN does not require us to do manual feature engineering.
B. NNs can work with both structured and unstructured data
C. For a large dataset, classic ML Algos can outperform NNs
D. NNs are able to work better with sparse data
Ans:
A, B, D
Correct
Options:-
- Unlike Classic ML Algos, NN does not
require us to do manual feature engineering.
- NNs can work with both structured and
unstructured data
- NNs are able to work better with
sparse data
Explanation:-
- In order to create a complex decision
boundary, classic ML Algorithms require us to do heavy feature
engineering, manually.
On the other hand, NN are able to find complex relations between features on their own. - NN are excellent in working with
unstructured data (image / text / audio data), whereas classic ML algos
are unable to handle this data.
- The performance of NN might be
comparable to that of classic ML Algos for small datasets.
However, if given a big enough dataset, NNs will always give better performance. - NNs work better with sparse data.
Q4. Scale
drives DL
Refer
to the given plot.
Which
of the following generally does not hurts an algorithm's performance, and may
help significantly?
Choose
the correct answer from below, please note that this question may have multiple
correct answers
A. Decreasing the size of a NN
B. Increasing the size of a NN
C. Decreasing the training set size
D. Increasing the training set size
Ans:
B, D
Correct
Options:-
- Increasing the size of a NN
- Increasing the training set size
Explanation:-
- According to the trends in the given
figure, big networks usually performs better than small networks.
- Also, bringing more data to a NN
model is almost always beneficial.
Q5. Factors
of DL performance
Which
of the following factors can help achieve high performance with Deep Learning
algorithms?
Choose
the correct answer from below, please note that this question may have
multiple correct answers
A. Large amount of data
B. Smaller models
C. Better designed features to use
D. Large models
Ans:
A, D
Correct
Options:-
- Large amount of data
- Large models
Explanation:-
- Over the last 20 years, we have
accumulated a lot of data. Traditional algorithms were not able to benefit
from this. Whereas, this large amount of data has been the fundamental
reason why DL took off in the past decade.
- In order to take advantage of this
large amount of data available to us, we need a big enough model also.
- Smaller models will not be able to
yeild very high performace, as they will not be able to take advantage of
the large amount of data.
- One main difference between classical
ML algos and DL algos is that DL models are able to “figure out” the
best features using hidden layers.
Q6. NN
true false
Mark
the following statement as true or false:-
"Neural
networks are good at figuring out functions, relating an input x to
an output y, given enough examples."
Choose
the correct answer from below:
A. True
B. False
Ans:
A
Correct
Option: True
Explanation:
- With NN, we don’t need to design
features by ourselves.
- The NN figures out the necessary
relations given enough data.
Q7. Why
NN
Why
might Neural Networks be preferred over Classic ML Algorithms?
B. NNs can work with both structured and unstructured data
C. For a large dataset, classic ML Algos can always outperform NNs
Choose
the correct answer from below:
A. Only Option A is correct
B. Only Options A and B are correct
C. Only Options B and C are correct
D. All the give options are correct
Ans:
A, B
Correct
answer: Only Options A and B are correct
Explanation:
- In order to create a complex decision
boundary, classic ML Algorithms require us to do heavy feature
engineering, manually. On the other hand, NN are able to find complex
relations between features on their own.
- NN are excellent in working with
unstructured data (image / text / audio data), whereas classic ML algos
are unable to handle this data
- The performance of NN might be
comparable to that of classic ML Algos for small datasets. However, if
given a big enough dataset, NNs will always give better performance.
Q8. Artificial neuron
Which
of the following statements is/are true about a single artificial neuron?
Statements:
A.
It is loosely inspired from biological neurons
B. It computes a weighted sum
C.
It applies an activation function
D. It is capable of performing non-linear classification using sigmoid
Choose
the correct answer from below:
A. Only statement A and B are correct
B. Only statement B,C and D are correct
C. Only statement A, B and C are correct
D. All the given statements are true
Ans: A, B, C
Correct
answer: Only statements A, B and C are correct
Explanation:
- The basic inspiration for artificial
neurons did come from biological neurons.
Biological neurons form a network a network within themselves.
Each connection, like the synapses in a biological brain, can transmit a signal to other neurons.
An artificial neuron receives signals then processes them and can signal neurons connected to it. - A neuron does it’s computation in 2
steps:
- First it computes the weighted sum
as: z=w1x1+w2x2+…+wdxd+b
- Then it applies an activation
function on top of this sum: a=f(z)
- A single neuron can perform binary
classification if it’s activation is the sigmoid function.
However, it cannot perform non linear classification on it’s own. We would need a network of multiple neurons to do that.
Q8. And
perceptron
We
want to design a perception that performs AND operation. Refer to the table
below:-
For
this, first, a weighted sum is calculated: z=w1x1+w2x2+b
Here, w1 and w2 are
weights, and b is the bias for the neuron.
The
activation function applied on this is as follows:-
Which of the following values of weights and bias will give the desired results?
A. w1=1, w2=1, b=-2
B. w1=1, w2=0, b=-2
C. w1=2, w2=1, b=4
Ans: A
Correct
answer: w1=1,w2=1,b=−2
Explanation:
- For the input (0,0) the perceptron
will perform calculation something like this:
z=w1x1+x2x2+b=(1)(0)+(1)(0)+(−2)=−2
Therefore f(z)=0 - Similarly for the inputs (0,1) and
(1,0)
For both cases, z=−1,
Therefore, f(z)=0 - But for the input (1,1) the value
of z=w1.x1+w2.x2+b=0
Therefore, f(z) = 1 - Therefore, among the options, 1,1,
and -2 give the required results.
- We can use any values for w1,w2,
and b which satisfy our conditions and output.
Q9.
Biological Neuron
Ans:
A neuron takes input, processes it, and sends output to other neurons
Q10:
Artificial Neuron
Ans:
A simplified model of a biological neuron used in neural networks
Q11: Dendrites
Ans:
Structures that receive input signals in a neuron, akin to features in ML
models
Q12:
Weights
Ans:
Parameters that determine the importance of features in a neural network
Q13:
Activation Function
Ans:
A function applied to the linear combination of inputs to introduce
non-linearity
Q14:
Sigmoid Function
Ans: An S-shaped activation
function used in logistic regression
Q15: Nonlinearity
Ans:
A
characteristic of models that can handle complex patterns beyond linear
relationships
Q16: Linear
Regression
Ans:
A
model representing a linear equation without an activation function
Q17: Logistic
Regression
Ans: A linear model with a
sigmoid activation function
Q18: Feature
Engineering
Ans: The process of creating
features to improve model performance
Q19: Encoder-Decoder
Ans: Neural network model for
compressing and then reconstructing data
Q20: Deep
Learning
Ans: A subset of ML involving
multi-layered neural networks to model complex patterns
