oxford-cs-deepnlp-2017 / lectures Goto Github PK

PDF file is damaged. Please reissue !

I found this course material has been extremely helpful and think it would make the best use of it complementing with solutions for practicals for self-examination.

Meanwhile, I've organised the practicals I've done into a repo. Reviews and comments would be very much appreciated.

On Slide 57, in the table's third row (n=1), (y-y^)**2=36, not 64, so C(4,2)=108, not 136.
On Slides 63, in the table's sum row the total is 14, not 13, so C(3,0)=14, not 13.
On Slide 67, the table's sum is incorrectly labeled C(2,0) instead of C(4,0), the table's sum row the total is 48, not 54, and so C(w4,b4)=48, not 104.
On Slide 73, in the table's third row (n=1), y=15.05, not 15.01, so (y-y^)**2=0.9025, not 0.91, in the table's fourth row (n=2), y=18.06, not 18.01, so (y-y^)**2=3.7636, not 3.96, and the table's sum is 13.7262, not 12.82. Since C(3,0)=14 not 13 (see above), the slide's conclusion is still correct.

What is the License under which this repo's materials are being shared?

Thank you.

There exists a similar task that is named text classification.

But I want to find a kind of model that the inputs are keyword set. And the keyword set is not from a sentence.

For example:

input ["apple", "pear", "water melon"] --> target class "fruit"
input ["tomato", "potato"] --> target class "vegetable"

Another example:

input ["apple", "Peking", "in summer"]  -->  target class "Chinese fruit"
input ["tomato", "New York", "in winter"]  -->  target class "American vegetable"
input ["apple", "Peking", "in winter"]  -->  target class "Chinese fruit"
input ["tomato", "Peking", "in winter"]  -->  target class "Chinese vegetable"

Thank you.

Would it be possible to also get the video from the world's leading expert on acoustic modelling and speech transcription that Chris mentions at the beginning of Video 9 (Lecture 7)?

Thank you.

We can use CNN to classify more than 10 thousands of images of the ImageNet.
I find that CNN could only classify 10-20 text classes as this paper write.
So what is the high limit of short text classification's classes' number?

https://www.youtube.com/watch?v=RP3tZFcC2e8&list=PL613dYIGMXoZBtZhbyiBqb0QtgK6oJbpm

Thanks for the videos. But the videos don't have primary video stream, i.e., you cannot see the face of the instructor and can see lectures slides only. Without seeing the face expression of instructors, the learning experience is not complete. Please do something for this.

Would it be possible to post lecture 9 in mp4 form as the other lectures?

Thanks!

Thank you for a great course, I (and I think many others) will appreciate if you add missing slides for lectures 10-13.

UPD. I've missed somehow 10-12 lectures' slides, but the request is still actual for 13th lecture slides.

Best Regards,
Valentin

Just wondering if the materials for the practicals will be made available somewhere online for the public.

• I've read your slides in lecture 1b (Deep neural network are our friends). In slide: "Gradient are our friends" explaining arg min C(w, b): w0, b0 = 2, 2; C(w0, b0) = 68. This's correct. But after that, I don't understand why the results of expression sum(-2(y^ - y)*x) are: 8, -40, -72. I think that: -8, 40, 72 are correct.
• By the way, I implemented this simple network but when I trained it through 100 times, the value of cost function was not convergent. Here is my code:

import numpy as np 
x=np.array([1,5,6])
y=np.array([0,16,20])
w = 2
b = 2
epoches = 101
learning_rate = 0.05
for epoch in range(epoches):
    out = x*w + b
    cost = np.sum((y - out)**2) 
    if(epoch % 10 ==0):
        print('Epoch:', epoch, ', cost:', cost)
    dcdw = np.sum(-2*(out - y)*x)
    dcdb = np.sum(-2*(out - y))
    w = w - learning_rate*dcdw
    b = b - learning_rate*dcdb

, and here is result:
Epoch: 0 , cost: 68
Epoch: 10 , cost: 1.1268304493e+19
Epoch: 20 , cost: 3.00027905999e+36
Epoch: 30 , cost: 7.98849058743e+53
Epoch: 40 , cost: 2.12700154184e+71
Epoch: 50 , cost: 5.66331713039e+88
Epoch: 60 , cost: 1.50790492101e+106
Epoch: 70 , cost: 4.01492128811e+123
Epoch: 80 , cost: 1.06900592505e+141
Epoch: 90 , cost: 2.84631649237e+158
Epoch: 100 , cost: 7.57855254577e+175

Please explain for me. Thank you in advance!