Department of Electronics and Telecommunication Engineering
Report of Artificial Intelligence & Machine Vision Workshop- IETE FORTNIGHT 2019
Title: Artificial Intelligence & Machine Vision Workshop
Date: 26th, 27th January, 2019, 3rd & 24th February, 2019
Number of students participated: 40
Year: F.E., S.E. & T.E.
Resource Person: Mr. Sameer Kadam and Miss. Aanchal Sharma
The seminar was conducted over a course of three days, 26th, 27th of January and 3rd of February. It was conducted by Ms.Anchal and Sameer, who themselves are the esteemed alumni of the college.
The first day started with a round of introductions. The students were asked about the reason they wished to pursue artificial intelligence and what according to them was set in the future of Artificial Intelligence. As amateurs’ students don’t have much clue about what ai actually is and the intricacies of its functioning. As we move into the future the demand for automated systems has never been more. And this is why they spent the better part of an hour trying to make the students aware of the applications and benefits of AI, various companies working on their own AIs and the special functions of each. What followed was an interesting session on the comparison of the various AIs that have been developed by companies like Google, Amazon, Apple, etc. The video shown explained how the accuracies of each of these systems differed. And also showcased the reference to context feature, that is the ability of the system to retain the information of the initial question and answer the questions that follow in accordance with the previous ones, of the various systems. The google assistant was by far the most superior as compared to the others having been developed for a much longer time.
They then explained how AI is used for automated driving. The automated driving can reduce the number of accidents and can simplify and safty proof the process of driving all together. The automated driving can be used to help navigate easier. There are levels for the kind of automation systwm used. The more the number of levels the better the automated system is. The complexity of the system is an insight to the kind of operations performed by the systems.
The number and location of the cameras determinmes the kind of safety provided by the AI system. The more the number of cameras the better the estimation and accuracy. They even talked about the kind of sensors required and the position and functions of these sensors. The new kind of precomputed maps that are now available with the various Artificial Intelligence systems have improvised the navigation systems by leaps and bounds.
V2x, the technology used in the in the car automated systems, has been developed recently. It meams vehicle to everything. It analyzes the surroundings of the vehicle. Every motion of the vehicle and the vehicles surrounding it has been monitered. The tesla car has 8 cameras. Each camera has adifferent function. And the position of the camera defines the importance of the camera. The greater number of the cameras are used for more accurate estimation. The basic functions controlled by these cameras include the parking and auto driving.
The other place that this camera and AI is used is for object recognition and classification. But since this is a vast topic with immense amount of details it was covered on the third day.
The last theoretical topic covered on the first day was the speech recognition and speech output of different kinds of AI Systems. And then the students were made to practice speech recognition and speech to text conversion codes.
On the next day the students were divided by the year. The first year students were taught the basics of python and the second years were made to revise the language and do a mini project.
Next image processing and programs on image processing were taught.
On the third day Sameer started the day with neural networks, basic terminologies and concepts.
Perceptrons, epoch, learning rate and target value were some of the important terminologies discussed at first. Next an introduction to machine learning was given. Machine learning is essential to understanding the concept of AI System. Convolution neural network was taught next. This is a multi layer system and each layer of the system has specific task. The task of the layers is more or less to separate and identify the features from the object that needs to be identified. It helps in the classification of objects and also with YOLO, You Only Look Once which is Real time object recognition software.
Next Deep Learning and applications of Deep Learning were taught followed by software like Softmax that are used to for Convolution Neural network and the students were then given programs to realize the concepts taught through the morning.
Photographs of events: