Online Hands-on Machine Learning and Artificial Intelligence Program

We propose a no-coding, non-programming but at the same time totally hands-on Program on Machine Learning & AI Techniques (ML & AI). It is an online, live, and totally interactive lab oriented program with the primary objective of disseminating techniques of Analytics using Machine Learning to enable one to apply them on data in one's research work or for teaching or in any industrial context.

KNIME:

We use KNIME to implement ML & AI Techniques. KNIME is to Data Science what SPSS is to Statistics. Just as in SPSS, it is easy to import data, analyse it and generate reports for statistical analysis, so also it is equally easy (if not easier) to implement ML&AI techniques and publish results using KNIME--no matter how small or how large your dataset is. Also, just as results from SPSS are recognised widely in research community, so also KNIME's credibility to Data Science is recognised the World over and for the sixth year in a row, Gartner has placed KNIME as a leader for Data Science and Machine Learning (ML) Platforms in its Magic Quadrant based on ability to execute and completeness of vision. See links here and here. KNIME also offers an additional bonus for those who are familiar with R. They can extend its utility very easily. This facility is in parallel with SPSS which also offers facility for programming. KNIME is being used extensively in industries in production oriented work. Please see this example link for KNIME related jobs.

H2O:

We also use another top-of-industry tool--H2O.ai. Both KNIME and H2O are open-source platforms, free to download and use (GPL ver 3 licence). It is very easy to install on your laptop. They are installable on Windows, Mac or Linux platform (for example Ubuntu).

Primary Objectives:

• i) Develop insights into data through visual analytics.
• ii) To discover if data has any structure.
• iii) To learn techniques to group/cluster data .
• iv) To develop models for predictive analytics.
• v) To optimize model performance, and
• vi) To discover attributes that contribute most towards higher performance-->Explainable AI.

Program Contents:

Briefly, program contents are as below. We practice techniques which consistently garner high performance and are well known in ML community:

• 1. Introduction to Machine Learning Technology
• 2. Data visualization (including t-sne, parallel coordinates, mosaic plots) and Feature importance
• 3. Unsupervised learning techniques
  • a) Kmeans clustering
  • b) Hierarchical clustering
  • c) DBScan algorithm
  • d) Expectation-Maximization algorithm
  • e) T-SNE manifold learning technique
• 4. Dimensionality reduction
  • a) Principal Component Analysis (PCA)
  • b) Random Projections
• 5. Supervised learning techniques for Classification and Regression
  • a) Decision trees
  • b) Ensemble modelling using Random Forest
  • c) Gradient Boosting Techniques
    • i) Adaboost
    • ii) Gradient Boosting Learner
    • iii) XGBoost
  • d) Handling imbalanced data—SMOTE and ADASYN
  • e) Performance measures: Accuracy, Precision and Recall, F-measure; Area Under the Curve, Cohen’s Kappa, Sensitivity, Specificity
• 6. Deep Learning Techniques
  • a) Neural Networks
  • b) Deep Learning models using H2O
• 7. Anomaly detection
  • a) Anomaly detection using isolation forests
  • b) PCA with Mahalanobis distance
  • c) Autoencoders
• 8. Automated Machine Learning (AML) using H2O

Pedagogy:

We strongly believe that a course in data analytics can only be practice-based rather than theory based. We also believe that a practice based course requires constant interaction with the teacher during lecture hours in real time. As it is an online course, the teaching pedagogy is like this: First the algorithm (or theory part) is conceptually explained without getting into mathematics and then a project is undertaken to implement the techniques. Datasets for implementation are made available in advance. During the lecture, we create workflow on KNIME and explain the steps. At his end, the student goes through the same steps on his laptop. Consequently, results are available at our end as also with the Students immediately. In short, both the teacher and students are working on their respective laptops simultaneously; students solve their problems and ask any questions to clarify. The whole experience is just as if everyone is sitting in a laboratory and working together. Students are required to have a laptop with minimum of 4GB of RAM.

Students will be given practical exercises (along with data) on every topic and they are expected to submit them for evaluation within the time prescribed---Learning-by-exercises to model with real data constitutes an important component of our program.

Who Should Attend?:

Today no field or discipline remains untouched by data. Most organizations today are driven by data. This program would, therefore, be useful to anyone who would like to advance his career growth opportunities or add value to his organization or do something useful for society or aspires to be a data scientist.In short, whether you are a teacher, Ph.d scholar, student (of any discipline), or a working professional--this program is for you. Machine learning is key to innovation.

Program Timings and Duration:

Total program duration is 30-hours spread over 5-weeks. Program will be delivered on Saturdays and Sundays from 2PM to 5PM (IST), If there is a sufficiently large group, We can find a mutually convenient time. Thus there are 6 hours of teaching per week. During the week students are expected to perform exercises. This methodology of "learning concept->performing class projects-->Do self-exercises" leads to better and stress-free absorption.

Program Fees:

INR 12000/- per participant (plus GST @18%) or equivalent in US dollars. Early bird discount of Rs. 2,400 per participant for nominations received on or before January 8, 2020. Fee is payable in advance by way of Local Cheque / DD in favor of “FORE School of Management” payable at New Delhi. Or deposit online through this link.

Webinar:

We propose to have a live webinar of 1 hour duration on 3rd January, 2020 (Friday), 5 PM to 6PM with the theme: "Wherever there is data, there is Machine Learning". To register for the webinar, please just send an acknowledgement email to address ashok@fsm.ac.in .

Program Faculty:

Prof. Ashok K Harnal: Graduated from IIT Delhi in Electronics and Communication; M. Phil with Distinction from Punjab University, Chandigarh, and MA (Economics) from Punjabi University. Expert in Big Data, Data Analytics and Deep Learning, both on the technology side as also on Analytics side. Extensively taught faculty and students on the subject of big data technology and analytics. Has been associated with University of California, Riverside, US, in one of the Executive Education programs on Big Data for the last three years. Participated in various machine learning projects with real world data in areas of business, environment, marketing and advertisement. Conceived, planned & implemented in Defence Estates three country-wide information systems: a) Raksha Bhoomi to computerize land records; b) Knowledge Management of land-title related files/maps in all Defence Estates offices; and c) Setting up of a Disaster Management organization, Archival Unit and Resource Center, at Delhi and at Pune for safe storage of land-title related records in paper, digital & microfilm forms. Authored two books: one on Programing Games on Computers and the other on Linux Applications and Administration; both books have been published by Tata McGraw-Hill. He can be reached at: ashok@fsm.ac.in.

Prof. Lalit K Jiwani: (PhD, IIT Delhi) is an Experienced academician and researcher with interest in Analytics and Decision Science. He has 14+ years of experience both in industry and academics. His primary interest is in the creation and application of Information Technology for Business and Management. His teaching and research interests are in the area of Machine Learning, Deep Learning, Statistics and Random Processes, Analytics and Information Technology for Business. He conducted research and published in various national and international journals namely IEEE, EURASIP to name a few. He can be reached at: lalit.jiwani@fsm.ac.in.

Prof. Sunita Daniel, PhD (IIT Kanpur); M.Phil., M.Sc. (University of Madras) is an Associate Professor at FORE. She has over 21 years experience in research and teaching. Extremely versatile in her interests, her research has covered a wide range of topics, from abstract algebra to creating algorithms for designing curves during her PhD, even venturing into disease modelling and epidemiology. She has presented various research papers at international conferences held at Dubai, Malaysia, Turkey, China and Zurich, and has also published her research work in reputed international journals. Her current area of research includes Big Data Analysis and Financial modelling, and their applications in real world scenarios like finance and prediction of patterns in various fields. She can be reached at: sunita@fsm.ac.in.