In a multi class classification problem, the output (also called the label or class) takes a finite set of discrete values . In this post, system model for a multi class classification with a linear layer followed by softmax layer is defined. The softmax function transforms the output of a linear layer into values lying…
Question 34 on signals from GATE (Graduate Aptitude Test in Engineering) 2012 Electronics and Communication Engineering paper. Q34. Consider the differential equation with and The numerical value of is (A) -2 (B) -1 (C) 0 (D) 1
I happened to stumble on Prof. Andrew Ng’s Machine Learning classes which are available online as part of Stanford Center for Professional Development. The first lecture in the series discuss the topic of fitting parameters for a given data set using linear regression. For understanding this concept, I chose to take data from the top…
Question 13 on analog electronics from GATE (Graduate Aptitude Test in Engineering) 2012 Electronics and Communication Engineering paper. Q13. The diodes and the capacitors in the circuit shown are ideal. The voltage across the diode is (A) (B) (C) (D) Solution The first half of the circuit is a negative clamper circuit and the second half…
Understanding gradients is essential in machine learning, as they indicate the direction and rate of change in the loss function with respect to model parameters. This post covers the gradients for the vanilla Linear Regression case taking two loss functions Mean Square Error (MSE) and Mean Absolute Error (MAE) as examples. The gradients computed analytically…
In May 2008, we derived the theoretical symbol error rate for a general M-QAM modulation (in Embedded.com, DSPDesignLine.com and dsplog.com) under Additive White Gaussian Noise. While re-reading that post, felt that the article is nice and warrants a re-run, using OFDM as the underlying physical layer. This post discuss the derivation of symbol error rate for a general…
Thanks to the keen observation by Mr. Phan Minh Hoang, I was notified that the Matlab/Octave scripts provided along with the topic raised cosine filtering was not behaving properly. Reason: I was not taking care of the division by zero when creating the raised cosine filter taps. 🙁
For curve fitting using linear regression, there exists a minor variant of Batch Gradient Descent algorithm, called Stochastic Gradient Descent. In the Batch Gradient Descent, the parameter vector is updated as, . (loop over all elements of training set in one iteration) For Stochastic Gradient Descent, the vector gets updated as, at each iteration the…
In Problem 4.36 of DSP-Proakis [1], the task is to provide insights into harmonic distortion which may be present in practical sinusoidal generators. Consider the signal , where . My take: The discrete time signal of fundamental period can consist of frequency components separated by radians or cycles (Refer Section4.2 in [1]). The Fourier series…
In typical digital signal processing applications, there arises need to increase the sampling frequency of a signal sequence, where the higher sampling frequency is an integer multiple of the original sampling frequency i.e for a signal sequence with a sampling frequency , change the sampling frequency to , where is an integer.
Question 7 on digital from GATE (Graduate Aptitude Test in Engineering) 2012 Electronics and Communication Engineering paper. Q7. The output Y of a 2-bit comparator is logic 1 whenever the 2 bit input A is greater than 2 bit input B. The number of combinations for which output is logic 1 is (A) 4 (B)…
We have installed Google FriendConnect on dspLog.com. With Google Friend Connect, you can: (a) You can interact with other members who have similiar interests. You will come to know the list of other sites (apart from dspLog.com) where the members have joined. You can add a member as a friend and so on. (b) You…
Question 24 on math from GATE (Graduate Aptitude Test in Engineering) 2012 Electronics and Communication Engineering paper. Q24. Two independent random variables X and Y are uniformly distributed in the interval [-1, 1]. The probability that max[X,Y] is less than 1/2 is (A) 3/4 (B) 9/16 (C) 1/4 (D) 2/3