Event Schedule

Assisted or Supervised models of machine learning are a great boon to automate many mundane tasks in ERP world. Supervised model combined with hypothesis models helps in creating "situations" and in predicting possible outcomes to a given situation. Simple cases of Critical Value Methods and p-value methods can be used to provide insights before user takes an action in ERP. In this presentation we will use the above methods to solve a problem of how invoices are matched with vendor payments.

Use Case:
Account Receivable accountants have to match incoming payments with open customer invoices that the payments are intended to clear. This is a laborious task riddled with many incorrect matches and subsequent reversals and painstakingly slow process. Using Machine Learning capabilities, the matching engine can learn from accountants’ past manual actions, capturing much richer detail of customer- and country-specific behavior and use hypothesis models to propose possible outcomes.

Demo:
A demo of this usecase will be shown

Srivatsan Santhanam, Sr Director of S4HANA Product Management APJ unit has been with SAP for past 15 years. Srivatsan drives the strategy, product features and adoption of SAP S/4HANA
Cloud product across APJ market unit.
Srivatsan has 4 US patent awarded to his credit and many more in filed state.
He is an alumni of IIM-Ahmedabad
Srivatsan is a regular speaker in many technology & innovation events across APJ.

Types of chatbot & applications around it.
- Developing a chatbot which can reason, infer & logic

CoFounder of zekeLabs & edyoda. Consulted & Trained for more than 25 companies in Machine Learning & related technologies. More than 10+ industry experience with companies like CISCO, Juniper. NIT Allahabad 2007 CSE.

Media and Entertainment industry is one of the key consumers of AI and Machine Learning solutions to help discover and search content based on various facets that machines can recognize. There are off the shelf solutions recognizing faces, some objects, keywords, sounds, etc, but the need of the M&E industry is significantly larger. Machine wisdom, from Prime Focus Technologies, synthesizes layers of AI to bring transformations to M&E workflows, making possible much simpler and easier operations.

Muralidhar Sridhar (Murali), is the Vice President for AI and ML, Centre of Excellence at Prime Focus Technologies and is passionate about how AI can positively impact the workflows of the Media and Entertainment industry. Murali has also been an Entrepreneur and Founder of Apptarix, been the Head of Product Management at Nokia Smart devices, Essential Software and served and consulted at many Silicon Valley Startups before. Over the last 21 years, he has lead product teams across various geos including India, China, Finland and US. He has filed 7 patents in India, US and EU with 5 granted ones.

Key Analytics Fundamentals - AI Vs Machine Learning Vs Deep Learning
Important Stages/Steps in E2E Analytics life-cycle
Case Study - Telecom Network Analytics
Machine Learning Selection Criteria, Methods, Algorithms, Model – Mathematical Space Representation
Machine Learning Categories – Supervised/Un-Supervised with applicable algorithms

There are various levels in the use of AI/ML in the field of QA. The industry is currently at various levels in the use of AI and ML. My talk will attempt to present a snapshot of where the industry currently is with the adoption of AI/ML and various paths of adoption can follow from here.

The talk will also focus on the use of AI/ML in the form of case studies after the above.

With over 15 years of experience in multiple industries and technologies, Bharath is a technologist who is passionate about solving business problems for clients. Currently he is a Principal Consultant at Thoughtworks. Apart from trying to marry technology and business, he loves to write, emote, listen to music and play any sport with a stick and ball.

- Internet = at least 20 Bn pages and exponentially increasing, hence there is a huge need of NLP expertise

- Natural Language Processing broadly refers to the study and development of computer systems that can interpret speech and text as humans naturally speak and type it.

- Accurate application of NLP can make life easier and increase revenue for different business lines

Siddheshwar Kumar Jain (SK) is a Data Scientist, Analytics professional and Transformation leader with over 22 years of experience. He drives data sciences & advanced analytics projects across sales & marketing, biz operations and e-commerce in his functional role. He is passionate about learning new technologies. He has experience of driving Digital Transformation in various organizations like Oracle India, Societie Generale, Fidelity Investments and Samsung. He has great experience at driving smart automation in industries like finance, pharma and investment banking. He has an engineering degree from IIT Roorkee, Lean Six Sigma Master Black belt from Indian Statistical Institute and conflict management leadership from IIM. Additionally, he has presented about Machine Learning and Analytics in various forums.

Bayesian networks are probabilistic graphical models that are capable of modeling domains comprising uncertainty.
Recent advances in quantitative finance have shown that the adoption of probabilistic graphical models is critical in order to understand
causality and uncertainty.Given the vast amount of macroeconomic factors affecting market movement such as supply of oil from OPEC countries,
geopolitical and geoeconomic changes among many other variables,Probabilistic graphical models (PGMs) allow us to understand
the market movement by learning the graphical structure from the historical data.

I am a polymath and unicorn data scientist with strong foundations in Economics, Finance, Business Foundations, Business Analytics and Psychology. I specialize in Probabilistic Graphical Models, Machine Learning and Deep Learning. I have completed Financial Engineering and Risk Management program from Columbia University with top honors, micromasters in Marketing Analytics from UC Berkeley and statistical analysis in Life Sciences specialization from Harvard. I have around 5 years of technical experience working in various companies like Infosys, Temenos and NeoEYED . I am part of dedicated group of experts and enthusiasts who explore Coursera courses before they open to the public , an ambassador at AIMed (an initiative which brings together physicians and AI experts) , participant @ Stanford Scholar initiative ,community member in NumFOCUS and volunteer at Statistics without Borders.

With the rapid pace of connectivity and product innovations that we've seen in the past decade across the online marketplace, simple A/B testing for deciding the efficacy of an action is not enough. In the era of intense competition, it is essential that we take that leap to learn not just population summary from A/B tests, but to make inferences on personalised levels - and spend on marketing only where it's necessary.

The applications are open to all industries - especially the ones where actions need to be determined and learned from big data anlaytics - for instance, who are the people whom this product feature should be targeted to - and that “who” can be answered on an individual level by building models on test-control testing and can be applied to the overall population.

Machine Learning Takeaways: Uplift Forest Models to determine the population fit for targeting. The idea of uplift, and how it relates and compares to traditional statistical models.

Analogous to the natural events, there is a phenomenon called Social Event, which captures the key issues that is relevant to the society. An early detection of a Social Event can be of immense value to marketing professionals and political leaders. With the advent of micro-blogging and online news, there is data set ready for detecting Social Events. In this talk we will identify key approaches for Social Event Detection, that work real time on internet scale.

Practical Deep Convolutional Neural Network architecture for Image segmentation. I will discuss on general semantic segmentation architecture based on encoder network followed by a decoder network.

I will cover popular medical imaging datasets, exploratory analysis on this dataset and the approach of using fully convolutional network (FCN) end to end and pixel to pixel for the task of image segmentation on medical images.

I will also show how it can be implemented in PyTorch a popular deep learning framework.

AI Senior Principal Engineer at Dell IT working on transforming Finance Cash Application Operations through Deep learning based solutions using Computer Vision. Working on a book project with Apress (Springer) publication on Deep Learning with PyTorch. Working on cutting edge deep learning ideas in Computer Vision - Object Detection/Semantic Segmentation, autoencoders. He has vast experience working with terabyte scale data and has been developing solutions in AI, Machine Learning space. Have worked for consulting companies in India delivering end to end solutions. Has strong experience in building data science practice from scratch and has diverse skills ranging from data engineering, bridge between business and technology, Information Modeling, data science. Has experience working with Banking and Insurance, FMCG & Retail. His research interests include Deep Learning, Computer Vision, Information Retrieval.

Contrary to popular beliefs, deep learning is not a magic AI wand but merely an effective mechanism to learn deep, robust, and abstract patterns. While it is quite easy these days to start with buzz-worthy and pre-trained models open sourced by hard-working research-scientists, it is equally hard to make the same models work in practice to solve custom problems at scale. Such problems fall into the broad categories of classification, detection, segmentation and embedding in case of computer vision. In this talk, I will consider object detection as an example in computer vision and show the importance of choosing the right set of data augmentation techniques to improve data distribution, the right set of loss functions to tackle data imbalance, an effective hard negative mining approach towards faster convergence and active learning approach towards better convergence. I will also show visualizations and interpretations of models trained with and without such strategies to argue that todays deep neural networks are hardly intelligent. The main take away would be a thorough understanding modern single-shot object detectors. I will show a few applications of our approach in physical retail domain. These applications parse photos taken in physical retail domain at scale, extract information via object detection, improve decision making of various stakeholders in FMCG brands.

Vijay Gabale is Co-founder and CTO of https://infilect.com. Prior to starting Infilect in late 2015, Vijay worked with IBM as Research Scientist. Vijay has a PhD in Computer Science from IIT Bombay, has several research papers and patents in computer technologies, and has been working in the field of machine and deep learning since last 8 years.

Neural Networks have become the state of the art and lie at the core of modern Artificial Intelligence algorithms that are used for varied applications across a wide landscape, from specific industrial scale usage for Image Recognition, Sentiment Analysis, Machine Translation down to the ever present "Alexa" that responds to your commands in the sitting room.

This talk begins by looking at the background and need for the development of Neural Networks, presents two typical use cases that would not be feasible without the use of Neural Nets. We then examine the internals of a Neural Network to understand clearly the mechanics behind their working. Without getting into a detailed mathematical explanation, we focus on creating an "intuitive" feel among participants of how Neural Nets work and the reason for their success.

The talk then demonstrates the working of a Neural Network, using a real data set and shows the working of "Tensor Board" for visualizing a Neural Network. We conclude the talk with references to 3 to 4 very well known network architectures and aim at generating curiousity among participants so that they can move ahead in this vast domain based on their individual interests.

Arnab Ganguly, presently working with Capgemini as a Director for Predictive Analytics based out of Bengaluru. With a background and education in Engineering from the IIT in Mumbai and having experience of using data for solving practical client problems, moved into the domain of Neural Networks out of personal interest, certified in Machine Learning and Deep Learning as a way to enhance knowledge in the domain. Applied this knowledge to help create valuable solutions to business problems using specific Neural Networks and advanced algorithms.

This presentation will talk of basics of deep learning and how it differs from traditional machine learning. It will define popular architectures of CNN and RNN.

It will also go through popular use cases of deep learning in Finance, Retail, Automotive and Other sectors.

Talk will cover the following subjects:

1: Introduction to machine learning recommendation system
2: Use case of Netflix movie recommendation system

Rohit Garg is a Techo-Functional-Banker and started in career with Citibank Corporate and Investment banking. In his current role as Vice President, Operations at RRD, he manages a number of financial services teams. As a speaker he regularly shares his views on the impact of automation on the financial services industry. Rohit did his graduation from MNIT, Allahabad and master from NITIE, Mumbai.

The current AI buzz is the outcome of 50 years of continuous research. The buzz around us is high not because current AI is perfect because it is still imperfect. We believe AI will take another 10 years where things will be as popular as internet. I will start with basics of AI/ML/DL and extend the talk with it’s working process. The talk will include different aspects (Use Cases, Challenges, frameworks, APIs, demo etc.) of Deep Learning for Computer Vision and where we can make improvements. I will take Face recognition as an example and will compare different models and available APIs. I will add, use cases of Facial recognition in different industries such as retail, education, smart city etc. and end the talk with a note for audience, why it is necessary for any country to develop core AI technology in house.

Suman is founder and CEO of Giscle Systems (https://giscle.ml), prior to Giscle he was Co-founder of Maptags and Navi Samagri. He has published research papers in different International Conferences and own few patents.

A session which will holistically cover what, why and how of Reinforcement learning, a branch of AI along with a live demo.

45 minutes of Hands-on knowledge - Participants need not replicate but note down the techniques.

Shubhradeep heads the AI and Machine learning division(Digital technologies) at MSys. He comes with a decade plus of experience in setting up departments delving into BI, Data Science,AI& ML verticals at many reputed companies like Market Share, IBM, Adobe, Oracle and recently MSys. He founded MAVIP(Massive vision and Intelligence project) the largest opensourced hyperlocal machine vision project; which has now grown up to a team of 11 people. He is also an Adjunct faculty in Deep Learning at Symbiosis Centre of Information Technology. He is the co-organizer of India's largest Datascience Meetup group that consists of 5500+ members. He has founded a Social Fintech startup in 2011 and has successfully exited in 2014. He is also an alumni of IIM Bangalore. He has spoken at multiple conferences like MODS, IndiaHacks, TIE Summit, Indian Technology Congress, OpenSource India etc. to name a few. This year(i.e. 2018) he will mostly be speaking at RedhatDevconf(Bangalore), STEPIN(Bangalore), DataHack 2018(Bangalore), 6th Global Summit on Artificial Intelligence and Neural Networks(Helsinki), AI Summit 2018, DLS(Toronto), AI World(Boston) etc.

This session will discuss some recent innovations in Machine Learning and Deep Learning related to Computer Vision, Image Processing etc. The presenter will cover emerging applications of vision algorithms such as generative modeling, style transfer and pose estimation.

Vivek Singhal, Co-Founder and Chief Data Scientist, CellStrat
Vivek is a leading Data Scientist and thought leader with expertise in Artificial Intelligence and Deep Learning.

He is a Serial Entrepreneur with strong experience in India and USA in AI, telecom and emerging technology in leading MNCs such as IBM, AT&T, Schlumberger as well as several startups like Healthiply, SalesGlobe and LocVille.com.

Currently, Vivek is a Co-Founder and Chief Data Scientist at CellStrat, India's leading AI startup.

Vivek holds a BE degree from IIT Roorkee and an MBA from Georgia State Univ, Atlanta, USA.

1. Will cover Computer Vision from basic image processing to pixel wise semantic segmentation. Will cover latest state of art deep learning techniques - FCN, U-Net on medical imaging dataset.

2. Will discuss PyTorch deep learning framework in detail and show how to prepare image dataset to carry out semantic segmentation.

Learning Objectives / Outline:

- Computer vision – Image Processing, Image Classification and Semantic Segmentation
- PyTorch deep learning framework

Saurabh Jha : AI Senior Principal Engineer at Dell IT working on transforming Finance Cash Application Operations through Deep learning based solutions using Computer Vision. Working on a book project with Apress (Springer) publication on Deep Learning with PyTorch. Working on cutting edge deep learning ideas in Computer Vision - Object Detection/Semantic Segmentation, autoencoders. He has vast experience working with terabyte scale data and has been developing solutions in AI, Machine Learning space. Have worked for consulting companies in India delivering end to end solutions. Has strong experience in building data science practice from scratch and has diverse skills ranging from data engineering, bridge between business and technology, Information Modeling, data science. Has experience working with Banking and Insurance, FMCG & Retail. His research interests include Deep Learning, Computer Vision, Information Retrieval.

How to move from a few bots to scaled up model
From centralized to distributed model
Standardized bot framework
Create utility model to provide RPA as a service

The Use Cases for automation using RPA tools are certainly compelling. However, the heavy lifting begins after the business buys into the promise of RPA. In this presentation, Mohan will share practical insights on establishing and scaling up an Automation Center of Excellence.

Mohan Krishnamoorthy is a Director of Enterprise Architecture at a Fortune 500 company. He is result oriented Enterprise Architect and Business Partner with a diverse background that includes translating strategy to actionable programs and engaging in transformation programs. His most recent experience has been in establishing and scaling up an Automation Center of Excellence. More about him on his linkedin

Insurance, as an industry has lagged in terms of transformation and has been traditional and product focused. The world is changing and so are the customer demographics. To be able to tap into new markets, new preferences, it is important for Insurance to be completely customer centric. Today, there is an opportunity to lower costs and better care for patients. Today's technologies can help. That is what digital transformation in the health insurance industry is about. This presentation will be about the breadth and depth of Digital Transformation in Health Insurance, and the potential roadblocks.

As we talk about creating a hyper-personalised and seamlessly connected world, it’s important to understand both ‘powers' and ‘limitations' of AI technologies. A business can exponentially grow by harnessing the powers of AI. In parallel, an understanding of it’s limitations is required to define right strategy and road-map for the growth.

Further, it is clear that the growth is an step-by-step journey — where — today, different business are in different stages of readiness for being benefited by power of data and AI. Hence, it is important to locate ‘your' business in this journey to blend the best of available Human Intelligence with the best possibilities of AI for defining right short-term goals for a long-term exponential growth for your business.

Bottom line: Before you can fix the problem, you have to understand it.

Anvita Bajpai – Founder of "SunvAI", Speaker, and Bestselling Author of 3 Books – is an IIT Madras and IIM Bangalore alumni, with about 14+ years of experience of building smart products using AI with MNCs, like, Oracle, Satyam, Wipro, and startups, like, Decidyn (as AVP), Vimagino and her own ventures. Anvita has published a number of Research Papers, Patents, and business and technology Articles; and has been invited as speaker at leading conferences like UXINDIA (2013 & 2014 both), WIAD 2015, GIDS 2015, RSD 2016, Antaragni 2016, HoG@IIMB 2017, RISE 2018, AI Connect 2018 etc.

AI implementations today, are largely hand-crafted & built for experimentation. Thinking about AI and its scenarios is happening in pockets. The larger picture of how all of this will work together to help business is yet to emerge. How should organizations think of AI as an integral piece of their overall strategy? What are the broad building blocks to be considered in this journey?

Padmashree is a global leader, bringing business transformations through Analytics & Data Sciences, across domains.
She heads Analytics, Data Sciences and Business intelligence @ Capgemini, India.
She has led large & complex transformational projects using Machine Learning, Artificial Intelligence, NLP and Cognitive computing across leading Banks, Insurance, Re-insurance & Multinationals;
She was instrumental in structuring, incubating and scaling Data Sciences & Analytics practice in Cognizant and Max Life, arguably the thought leaders in establishing such a practice in the respective industries.
She has proven track record in crafting strategic & business centric offerings & solutions involving Advanced analytics, Data & Technology, that have generated extensive; quantifiable, business value.

RPA has tremendous opportunity to automate repetitive processes in Service Management which results in considerable cost saving by reducing tickets effort. In this session we will walk through a typical development life cycle of creating a RPA BOT and its practical applicability in Service Management. We will also walk-through a practical scenario on how RPA BOT can be used in Life Sciences Regulatory environment.And we will end the session by discussing some challenges faced during RPA automation.

Subeed Ahmed is a computer science graduate with Executive General Management from IIM Bangalore. He is a technologist with over 19 years of Architecture & Design, Enterprise Agile transformations, Account Management and Service Delivery in a large complex, cross-functional global programs. His current focus is on Digital Technologies, Intelligent Automation., AI/ML strategy, digital workforce technology, and operational excellence through the use of cognitive technology.

Manish Rathi is a computer science engineer with over 12 years of work experience in Architecture & Design, Agile Distributed Delivery of complex global programs at enterprise level.

His current focus area is RPA tied up AI/Machine learning to produce a human like intelligent automation.