ADAVANCED FACIAL RECOGNITION USING ARTIFICIAL INTELLIGENCE

 

STUDENT WEBINAR-30/11/2020

 

ADAVANCED FACIAL RECOGNITION USING 

ARTIFICIAL INTELLIGENCE

 

Nano Technology- Thin Films

SUPREME BEAUTY OF MATHEMATICS

 Sneha H
3rd Sem B.Sc(PME)
NCMS

SELF ORGANIZING MAPS

 Nandini D
5th BCA NCMS

 SOM is developed by Finnish professor and researcher Dr. TeuvoKohonen in 1982 was sometimes called a Kohonen map. It is a type of artificial neural network which follows unsupervised learning to generate a low dimensional (two dimensional), discretized representation of the input space of the training samples are referred as map, and is therefore a method to do dimensionality reduction. They apply competitive learning instead as opposed to error correction learning.

Working:

                     The data points in the data set recognize themselves by competing for themselves. Initializing the weight vectors is the first step of mapping in SOM. A weight that best represents that sample is searched from the randomly selected sample vector and the map of the weight vectors. The weight vectors have neighboring weights that are close to it. The chosen weight  and the neighboring weights are rewarded to become more like that randomly selected sample vector. This will allow the SOM to take different shapes. Commonly they form square/rectangle/hexagonal/L shape/in 2D feature space.

Algorithm:

                    For 0 to X number of training epochs

                           Select a sample from the input set

                           Find the “winning”  neuron for the sample input

                           Adjust the weights of nearby neurons

                   End for loop

Pros:

●       Data is interpreted and understood easily.

●       Dimensional reduction helps in observing similarities in data.

Cons:

●       Extraneous data in weight vectors will add randomness to the groups.

For more information about the SOM do watch Click the link below

video:https://www.youtube.com/watch?v=0qtvb_Nx2tA

BLOCK CHAIN TECHNOLOGY AND AI IN AGRICULTURE

Webinar by
Shagufta
5th Sem BCA

 

INTERESTING FACTS ABOUT MATHEMATICS

Jaidivya H N
1st Sem B.Sc (PME)

 

ISRO-2020: PSLV-C49

 Chandan L M
3rd Sem B.Sc (PME)

FEED FORWARD NEURAL NETWORK

 Shagufta
5th Sem BCA

 

The Feed Forward neural network was the first and simplest type of artificial neural network devised. The network is called feed forward because it is forward in nature. The information travels forward in the network without any loops, first through the input nodes, then through the hidden nodes (if present), and finally through the output nodes. The series of hidden layers defines the accuracy of the generated output. Here the connections between units do not form a circle.

Feed forward neural network are primarily used for supervised learning in cases where the data to be learned is neither sequential nor time-dependent. That is, the network compute a function f on fixed size input x such that

f(x)=y f(x) \ approx. y f(x)=y for training pairs (x,y) (y,x) (x,y).

The major limitation is that it is not capable of processing sequential data. For example, we cannot do speech recognition using a feedforward neural net – we need a recurrent neural network for this.

When we create a deeper feed forward neural network, we are giving the model the ability to capture more complex representations. This network is best used in image recognition tasks, Natural language processing, voice recognition tasks and many more.

The feedforward neural networks are applicable to many spaces where the classic machine learning techniques are applied; the major success of it has been in computer vision and speech recognition where the classification spaces are quite complicated. Also, these kind of neural networks are susceptible to noisy data and easy to maintain.

LAUNCHING PRAGYAN-SCIENCE FORUM PODCAST

PRAGYAN PODCAST PROMO

Click the above link to listen

CONVOLUTIONAL NEURAL NETWORKS

Lakshmi
5th Sem BCA, NCMS

One of the most popular neural networks is the convolutional neural networks. In recent years it has given breaking results in different fields related to pattern recognition, image processing, voice recognition etc. The most beneficial aspect of CNNs is reducing the parameters in artificial neural networks. This achievement prompted researcher's and developers to solve the complex problems. In deep learning a convolutional neural network is a class of Deep neural networks it is most commonly applied to analyzing visual imagery. The term "convolution" is a mathematical operation on two functions that produces third function. CNNs are regularized versions of multilayer perceptron , multilayer perceptron  usually mean fully connected networks that is each neuron  in one layer  is connected to all neurons in the next layer. CNNs are inspired by biological processes in that the connectivity pattern between neurons. The architecture of the CNNs is different from the traditional multilayer perceptron  . This is to guarantee a certain degree of shift and distortion invariance . To do so, three design ideas are merged, which are, local receptive fields, common weights, and spatial and temporal subsampling. The CNNs consists of an input and output layers as well as multiple hidden layers. The hidden layers are convolutional layer, pooling layer, fully connected layer, receptive field layer, weights.  A typical architecture consists of several convolution layers and pooling layers followed by one or more fully connected layers.

The convolution function

The applications of convolutional neural networks are

Computer vision

Scene labelling     

Natural language processing                                                                                                        

Face Recognition      

Action Recognition                                                                                           

Text and image classification                                                                                                                       

Speech recognition

Convolutional neural networks have accomplished astonishing achievements across various domains including medical research, and an increasing interest has emerged in radiology. Therefore CNN offers better accuracy when compared with other standard approaches .

RADIAL BASIS FUNCTION NEURAL NETWORK

 Rukba Firdose
5th Sem BCA, NCMS

Radial basis function neural network (RBFNN) is very prominent in data processing. However, improving this technique is vital for the NN training process. Neural -symbolic integration combines both the connectionist system and symbolic artificial intelligence (AI) systems. Radial basis function (RBF) networks typically have three layers: an input layer .a hidden layer with a non -linear RBF activation function and a linear output layer. It is used for the approximation and recognition pattern. And also multi quality function. The input layer is made up of some node (sensory units) that connect its network to its environment. The secondary layer consisting of hidden layer which has hidden units. Apply a non linear transformation from the input space to its hidden spaces. The output layer is linear it is basically designed to supply the input from the source node to the activation function and produce the output and it is work as supervised learning layer. From the output layer we can calculate the errors as well.

In single perceptron / multi -layer perceptron (MLP), we only have linear separability because they are composed of input layers (some hidden layers in MLP).

For example:

The local modelling capability of RBF is demonstrated on the logical operators AND and OR. The examples used are demonstrative for the structure and physical meaning of the weights of the RBF network. Two datasets of four training objects operators AND and OR. The datasets and the network topology are graphically depicted . the AND operator yields a positive output on inputs of identical sign, whereas the OR-operator responds positivity on any positive input.

Different learning algorithm may be used for learning the RBF network parameters. We describe three possible methods for learning , spreads and weights.

The second or hidden layer performs a non-linear mapping from the input space into a (usually) higher dimensional space in which the patterns become linearly separable.

ADVANTAGES OF RBFNN:

Easy design

Good generation

Strong tolerance to input noise

Online learning ability

RBF networks make it very suitable to design flexible control systems.

 

DISADVANTAGES OF RBFNN

Although the training is faster in RBF network but classification is slow in comparison to multi layer perceptron due to fact that every node in hidden layer have to compute the RBF function for the input sample vector during classification.

A function radial basis if its output depends on (is a non-increasing function of) the distance of the input from a given stored vector.

Click below to learn more 

https://www.youtube.com/watch?v=j6lhPMOPgsQ

INDIAN SPACECRAFT

 Chandan L M
2nd B.Sc, NCMS

Click the below link to know more
https://www.isro.gov.in/spacecraft

LiDAR [Light Detection and Ranging] used in Autonomous Vehicles

Navaneetha
3rd Sem BCA, NCMS 

LiDAR is a remote sensing device or a sensor, which uses laser light to detect and measure the distance. The name itself represents that it works on the light detection.  LiDAR is used in an Autonomous Vehicles  also known as self-driving cars like robot. And that sensor, lidar is installed on the top of the cars. 

LiDAR is capable of detecting ultraviolet rays and also infrared rays to measure. Each light rays means to a different distance. LiDAR senses thousands of laser lights per second. So that, the sensor placed on the top of the cars, emits light which is measured with pulses and waits for the return signal also it transmittes rapid update point cloud into animated 3D Representation, where a person can see the map. And that 3D  Representation is been created by measuring the speed of light and distance covered by it,  which helps to locate the position of vehicles around it. On that basis, sensor commands to put break or to speed down/speed up or to stop the vehicle. The main advantage of this LiDAR sensor in an Autonomous Vehicles is- we need not to confront the issues such as sharp curves, rash driving or traffic jams and some other issues confronted by humans. And it provides a smoother and the safe driving.

Click the below link to know more about LiDAR
https://www.youtube.com/watch?v=uOhD8PYAVF4

SPINNAKER COMPUTER

 Sanjay.R
3^rd Sem BCA, NCMS

Spinnaker (Spiking Neural Network Architecture) is a massive parallel, supercomputer architecture designed by the Advanced Processor Technologies research group (APT) at the department of computer science, University of Manchester.  Spinnaker is composed of 57,600 processing nodes , each with 18 ARM9 processors(specially ARM968) and 128MB of mobile DDR SDRAM, totaling 1,036,800 cores and over 7TB of RAM. The computing platform is based on spiking neural networks, useful in simulating the human brain.

  The completed design is housed in 10 19-inch racks, with each rack holding over 100,000 crores. The card’s holding the chip’s are held in 5 blade enclosures, and each core emulates 1000 neurons. In total, the goal is to simulate the behavior of aggregates of up to a billion neurons in real time. This machine requires about 100kw from a 24 V supply and an air conditioned environment.  Spinnaker is being used as one component of the neuromorphic computing platform for the human brain project.

On 14 October 2018 the HBP announced that the million core milestone had been achieved.

On 24 September 2019 HBP announced that a 8 million euro grant ,that will fund construction of the second generation machine, (called spincloud) has been given to TU Dresden.

Developer                Steve Furber

Type                         Neuromorphic

Release date            2019

CPU                          ARM968E-S@ 200MHz

Memory                    7 TB

Searching for Life on Mars: Part-1

 Dr. Santhosh TCM
Asst. Professor,
Department of Physics, NCMS

CONGRATULATIONS!!!

Congratulations to Shagufta and Chandana S of 5th Sem BCA

Got Selected for Bolt IoT Student Partner Internship Program

 

RFID-RADIO FREQUENCY IDENTIFICATION

Ravindra Reddy P M
5th Sem BCA, NCMS

First convinced in 1948, Radio Frequency Identification(RFID) has taken many years for the technology to mature to the point where it is sufficiently affordable and reliable for widespread use. From Electronic Article Surveillance(EAS) for article security to more sophisticated users in mainly clothing. RFID is seen by some as the inevitable replacement for bar codes. With increasing use comes increasing concern on privacy and security. Clearly there is considerable work to be undertaken before RFID becomes as pervasive as a bar codes although the tempo of change is increasing rapidly. Active RFID systems typically operate in the ultra-high frequency (UHF) ban and offer a range of up to 100 m. The ultra-high frequency ranges include frequencies from 300 to 1000 MHz, but only two frequency ranges, 433 MHz and 860-960 MHz, are used for RFID applications.

Security

Like other technologies,
RFID is exposed to security threats and, specifically, to attacks on the confidentiality, integrity, and availability of the data stored on the tags or on the information exchanged between a reader and a tag.
RANGE:
Ø  UP TO 100 m.
FREQUENCY:
Ø  300 to 1000 MHz 

Extract text from Images/PDF with Tesseract OCR

 Puneetha D S
5th Sem BCA, NCMS

Do you want to convert image/pdf into text? Then you will need an application that can recognize text using OCR(optical character recognition). OCR uses artificial intelligence for text search and its recognition on image.

Tesseract is an open source text recognition engine which is available under Apache license2.0. It is free software, originally developed by Hewlett-Packard  and has been funded by GOOGLE. Tesseract can extract text from image directly or using API. Even it can recognize wide variety of languages. It is easy to install and use.

Tesseract is written in C and C++ and it is available for Linux, Windows and macOS operating systems. It is compatible with many programming languages and frame works. It recognizes the pixels, letters, words, and sentences on the image. When we input image it will analyze the adaptive thresholding and converts to binary image and it analyze the components and find lines and recognize the image and at last it will provide the text which is recognized from the input image.

The main pros of using Tesseract OCR are building training is easy, it can recognize many languages. Better the image quality (like size, contrast, lightning) better recognition. No OCR software is 100 percent accurate. The number of errors mainly depends on type of document and quality of the image. The cons of using Tesseract are recognition rate, some text blocks were recognized more than 1 time.

Indian languages accuracy
Bengali -98%
Gujarati-97.21%
Hindi-90%
Kannada-33.3%
Oriya-96.3%
Gurmukhi-97%
Tamil-99%
Telegu-98.5%
Malayalam-90.22%

Click the link to install Tesseract OCR

https://digi.bib.uni-mannheim.de/tesseract/tesseract-ocr-w64-setup-v5.0.0-alpha.20200328.exe

https://digi.bib.uni-mannheim.de/tesseract/tesseract-ocr-w32-setup-v5.0.0-alpha.20200328.exe

NAIVE BAYES CLASSIFIER ALGORITHM

Chandana. S

5^th  SEM BCA

Naïve bayes classifier is the collection of classification algorithms based on bayes theorem. It is the family of algorithm where a value of a particular feature in a class is unrelated to the value of any other feature. The statistician and philosopher, Thomas bayes and the theorem named after him, bayes ’theorem, which is the base for naïve bayes algorithm. This algorithm or model is easy to build and used for very large data sets. Naïve bayes is also known to highly sophisticated classification methods. We also use machine learning, python, R etc. languages in this model. P(c|x) from p(c), p(x) and p(x|c) formulae is used to calculate naïve bayes algorithm.

P(c|x)=P(x|c)p(c)/p(x)

Naïve bayes algorithm is used to solve the probabilistic queries in different class based on various attributes. Naïve bayes algorithm is a real time predictor which is a fast learning algorithm used to make predictions in real time. It is also used for binary classification and multiclass classification. It has three types of algorithm called GaussianNB, multinomialNB, BernoulliNB. The main applications of bayes algorithm is real time prediction, multiclass classification, text classification. This algorithm is also used for spam messages or spam mails.

Examples: To mark an email as spam, or not spam.

                 Classify a news article about technology, politics, or sports.

                 This is also used for face recognition algorithm software.

The main limitation of naïve bayes algorithm is the considering of independent predictor features.

This algorithm is used in classifying problems which helps in identifying the category of a new observation from a set of categories. The category is determined on the basis of a training set of data which contains observations whose category membership is already known.

Click the link to learn more about NAIVE BAYES CLASSIFIER ALGORITHM

https://www.youtube.com/watch?v=sjUDlJfdnKM

Enki:Learn data science, coding and tech skills

Nandini D

5th sem BCA, NCMS

You can get a quick boost of skills anywhere if you have a good app that allows you to practice on a bus or in your leisure time. A good app can fuel your professional growth and development. One such app is Enki which helps you to learn coding in the easiest way.

The topics included are:

1. SQL
2. Data Science
3. JavaScript
4. Python
5. Docker
6. MongoDB
7. Linux
8. Java

This app supports everyone from beginners to the experienced developers. It provides you with the expert written lessons, daily code workouts, Interactive quizzes etc. You can keep track of your progress which helps you to improve yourself. One can create community /join a community to ask questions or to practice the code. The modules in this app are rich with contents. This app will test your understanding right after a concept by conducting a small quiz to make sure that  you are learning the concepts thoroughly and not just browsing. This is simple and the best app for a beginner to learn coding. This app is not going to charge a penny for you to learn the concepts. The only con is that you have to pay for additional workouts(exercises) and you are not provided video lectures. Other than that it’s a fantastic app.

Link to download the app: https://play.google.com/store/apps/details?id=com.enki.insights