One Week Faculty Development Programme for CSE department on “Machine Learning Techniques & it’s Applications
January 04, 2016
University Sponsored one week FDP on “Machine Learning Techniques & it’s Applications” sponsored by Dr. A.P.J. Abdul Kalam Technical University, Lucknow, Uttar Pradesh.
The opening of program was made by Saraswati Vandana and lamp lightening by distinguished guest, Dr. Anurag Singh Baghel, Dr. Vineet Kansal, Director, ITS Engineering College, followed by welcome speech by Dr. Sanjay Yadav, HOD-ME.
1st session of 04/01/2016 on Computing with Artificial Neural Networks was initiated by Dr. Anurag Singh Baghel, Assistant Professor, GBU, Greater Noida comparing Artificial Neural Networks with human brain. He elaborated on working and training of single layer and multilayer Neural Network. He further explained in detail about decision boundary and learning strategy. He also took example to show how to use various operators in MATLAB. 2nd session of 04/01/2016 on Nature Inspired Techniques and Applications was initiated by Dr. T. V. Vijay Kumar, Professor, JNU, New Delhi briefly describing about behaviour exhibited and patterns formed in nature by Bees, Ants, Swarm and Fishes which help in designing algorithms to find solution to complex problems. Then he differentiated P and NP problems. This was followed by explanation of Classification of Nature-Inspired Algorithms and discussion of meta-heuristics, annealing, exploration and exploitation and also Crossover and Mutation.
1st session of 05/01/2016 on “Embedded Intelligence and IoT” was initiated by Dr. Jayadeva, Professor, IIT Delhi briefly explaining about working of Neural Network. He elaborated on Standard SVM formulation and various applications briefly describing about Application of SVM to Pattern Classification, & Regression Estimation Problems. This was followed by discussion about SVM with Soft Margin. 2nd (Post lunch) session of 05/01/2016 on “Support Vector Machines” for Pattern Classification Problems was initiated by Dr. S. Balasundaram, Professor, School of Computer & Systems Sciences, JNU, New Delhi briefly describing about Application of SVM to Pattern Classification, & Regression Estimation Problems. Then he briefed about 2 class linearly separable problem and Primal form & Dual form of Standard SVM. This was followed by explaining Kernel Transformation taking examples of Kernel functions and discussion of Mangasarian Approach and its matrix representation.
1st session of 06/01/2016 on “Neural Network with MATLAB” was initiated by Dr. Manu Pratap Singh, Professor, B. R. Amedkar University, Agra briefly explaining about Pattern, Pattern Class and Classifier. He elaborated on different Neural Network Architectures. He further briefed about decision boundary and PERCEPTRON learning Algorithm. He also explained in detail about various MATLAB functions and took examples to show PERCEPTRON LEARNING / TRAINING using various MATLAB functions. 2nd session of 06/01/2016 on “Signal Processing & Machine Learning” was initiated by Dr. Angshul Majumdar, Professor, IIT Delhi with a video showing how machine can be trained through signals from brain. This was followed by elaboration of concept of Dictionary Learning, Applications of Image Processing. This was followed by discussion about Greedy layer wise Learning Approaches-Restricted Boltzmann Machine, Auto-encoder and proposed approach (Dictionary Learning). Then he explained GROUSE (GROUP Sparse Encoding).
1st session of 07/01/2016 of FDP on Machine Learning applied to Text Processing was initiated by Dr. Vivek Kumar Singh, Professor, Department of Computer Science, BHU, Varanasi briefly explaining about Machine Learning through example. He elaborated on Information Retrieval and Text Processing and explained how Indexing and Inverted Index Construction is done and used for Information Retrieval. Then he shared Internet Access and Social Media Statistics taking examples of Wikipedia, Facebook and YouTube for developed and developing nations and how they can be used to draw inferences regarding a particular product. 2nd session of 07/01/2016 on Learning Semantics from Text using Word Space Model and Random Indexing was initiated by Dr. Niladri Chatterjee, Professor, Department of Mathematics, IIT Delhi explaining the difficulties faced while analysing the unstructured data volume & variety and handling the same words having different meaning in different context for machine learning taking various examples of synonyms. This was followed by discussion of statistical problems and modern view of Text Processing. Then he briefed about Document based and Word based Co-occurrence and discussion of usage of Co-occurrence Matrix and Context Matrix to find solution to problems. This was followed by discussion of Random Indexing to solve problems.
1st session of 08/01/2016 on Introduction to Map-Reduce and Join Processing was initiated by Mr. Himanshu Gupta from IBM Research, India with a interaction discussing about various issues in Parallel Processing Framework. Then he explained in detail taking an example how HADOOP provides a framework for creating distributed applications that process huge amount of data. This was followed by elaborated discussion on how data is stored and organised in HDFS and how data is processed using Map-Reduce Framework. Then process of mapping tasks using HDFS and HADOOP sorting and shuffling was explained through a figurative example. Finally a demonstration of how HADOOP engine helps to reduce processing by reducing tasks was given and join processing to get final processed output was explained. 2nd session of 08/01/2016 on Computational Intelligence (CI) was initiated by Dr. Jagdish Chand Bansal, Assistant Professor, South Asian University, New Delhi briefly explaining about Intelligent Agents. Then he talked about Classification of various CI Techniques including Particle Swarm optimization, Biogeography based optimization and Spider Monkey optimization. He cited various difficulties faced in the study of Computational Intelligence and various research issues. He also explained concepts like Stagnation, Exploration, Exploitation, HIS, SIV, BBO Operators during discussion of various CI Optimization Techniques.
Finally, Dr. Vineet Kansal, Director and Dr. Ashish Gupta, HOD CSE/IT congratulated and gave certificate of participation to all FDP participants.