Dr. Ganapathi Pulipaka is a Chief Data Scientist for AI strategy, architecture, application development of Machine learning, Deep Learning algorithms with experience in deep learning reinforcement learning algorithms, IoT platforms, Python, R, and TensorFlow, Big Data, IaaS, IoT, Data Science, Blockchain, Apache Hadoop, Apache Kafka, Apache Spark, Apache Storm, Apache Flink, SQL, NoSQL, Mathematics, Data Mining, Statistical Framework, SIEM with 7+ Years of AI Research and Development Experience in SAP Cloud Platform Integration, AWS, Azure, and GCP. He is also an SAP Technical Lead and comes with 20+ years of experience as SAP Technical Development and Integration Lead. Dr. Ganapathi works for Accenture, a large technology corporation in the world with $41.60 billion in revenues for 2018, employing 477,000 people globally in 2019. Accenture provides geospatial platforms, developer platforms, cloud computing, distributed ledger technology, Artificial Intelligence (AI), Extended Reality (XR), SAP, and Quantum Computing technologies to the customers.
Dr. Ganapathi as Chief Data Scientist sets up the strategy for AI and provides CXO insights to build an intelligent enterprise by designing, developing, and deploying machine learning and deep learning applications to solve the real-world problems with reinforcement learning algorithms in natural language processing, speech recognition, text to speech, chatbots, and speech to text analytics in PyTorch, Python, TensorFlow, and R. He has a vast experience in data exploration, data preparation, data mining, statistics, and applying supervised and unsupervised machine learning algorithms, machine learning model training, machine learning model evaluation, predictive analytics, bio-inspired algorithms, genetic algorithms, and natural language processing. Dr. Ganapathi has experience in building recommendation systems and applying algorithms for anomaly detection in the financial industry. He also has experience in deep reinforcement learning algorithms for robotics and IoT. He works with the customers in determining the AI strategy to applying convolutional neural networks, recurrent neural networks, and long-term short memory with deep learning techniques to solve various business conundrums at scale. He has developed a number of machine learning and deep learning programs applying various algorithms and published articles with the architecture and practical project implementations on GitHub, medium.com, and data-driven investor. He also works as an advisory board for clients with experience in Python, TensorFlow, Caffe, Theano, Keras, Java, and R Programming languages implementing stacked auto encoders, backpropagation, perceptron, Restricted Boltzmann machines, and Deep Belief Networks.
Dr. Ganapathi is interested in helping the clients in setting up the strategy, coming up with statistics, applied mathematics, and developing algorithms for biology, drug discovery in healthcare, utilities, energy, financial applications, customer services, predictive maintenance applications and analytics in aerospace, and developing algorithms to run autonomous vehicles. In his role, he comes with experience in multiple IoT platforms. He wrote around 400 AI research papers and management strategies with a vast number of big data tool installations, SQL, NoSQL, practical machine learning project implementations, data analytics implementations, and statistics for publishing with the Universities as part of academic research programs. Dr. Ganapathi implemented 29 projects for Fortune 100 corporations Aerospace, semiconductor, cloud computing, hardware, software, AI, manufacturing, IS-AFS (Apparel footwear solutions), IS-MEDIA (Media and Entertainment), ISUCCS (Customer care services), IS-AUTOMOTIVE (Automotive), IS-Utilities, retail, high-tech, life sciences, healthcare, chemical industry, banking, and service management.
Enriched Industry Performance
Dr. Ganapathi is also a Bestselling Author of two books that stayed #1 on Amazon for the last few years with five-star ratings. The first book titled “Big Data Appliances for In-Memory Computing: A Real-World Research Guide for Corporations to Tame and Wrangle Their Data,” was published on December 8, 2015. The book also has an audible version available on Audible.com, iTunes, and Amazon.
The second book titled “The Future of Data Science and Parallel Computing: A Road to Technological Singularity” was published on June 29, 2018.
He is currently authoring a third book with Packt Publications PyTorch 1.0 – Deep Reinforcement Learning cookbook covering the most advanced reinforcement learning with deep learning techniques such as Trust Policy Region Optimization, reinforce algorithms, policy gradients, Montel Carlo methods, temporal difference learning, discretization, tile coding, Deep Q Networks, continuous exploration with adaptive noise scaling, and cross-entropy methods algorithms. Currently, there is no literature, book, or video on these topics developed in PyTorch.
Experience that Glorifies the Achievements
Dr. Ganapathi is a public figure on Twitter with a massive following of 57K+ followers, who manages a public data science, artificial intelligence, machine learning, and deep learning community. He secured the fifth spot on the Top 100 Influencers, Brands and Publications list by Onalytica in 2018 in Data Science category. He was also listed in Onalytica’s Top 25 Entrepreneurs in business intelligence in 2018. The other awards won by Dr. Ganapathi Pulipaka include Levi Strauss and Co. Jeff Gordon Award in 2011, for On-time and On-budget delivery of the projects, HP Award of Excellence for performance as a Senior SAP Technical Lead Consultant consecutively in 2002 and 2003, and successful implementation of North America’s first SAP CRM 7.0 project and SAP Pinnacle Award.
Challenging Stairway Towards Success
Building deep deterministic policy gradients in a parameterized-continuous action space requires single agent learning and multi-agent learning with deep reinforcement learning. Combining the deep neural networks and reinforcement learning can create intelligent agents (robotics) to work in continuous action spaces. Deep-Q-Learning is a model-free reinforcement learning technique/algorithm that can estimate the value function of the discrete action space and maximize the output for multiple states of each input. Generally, the robust neural networks work well in continuous state spaces, but making it work for continuous state-action spaces is a challenge because the neural network output nodes are trained to produce the outputs with Q-value estimates rather than generating the continuous actions. A novel architecture of actor and critic can decouple the action selection and value learning.
We can represent the original Q-Learning equation, where
Actor network = k, parameterization of actor network with θk, which will take the input as state s and generates a continuous action output as a. The critic network Q is parametrized with θQ, which will take the input as state s and transforms into a continuous action scalar Q value Q(s,a).
Q(s,a) =Q(s,a) +α(r+γmaxa′Q(s′,a′)−Q(s,a))
When we apply, this equation to deep neural networks, it will minimize the loss function as shown below.
The various approaches with applied mathematics, applied statistics, and code in Jupyter Notebooks, Google Colaboratory notebooks in Python and PyTorch can be found in the upcoming book PyTorch 1.0 – Deep Reinforcement Learning Cookbook sponsored by Packt Publications.
Leadership Beyond Convention
Dr. Ganapathi believes leadership is not a soft determinant of what statistics can do for you or it is not about building a blockchain with a mix of Ethereum and cryptocurrencies for financial gains. Leadership is going beyond formulas and shepherding the nation in a strategic advisory role and guiding the corporations in setting them up for success with innovation in building smart cities and boosting the real GDP with infrastructure that drives the economic output of the nation powered by data governance, data architecture, data science, machine learning, deep learning, IoT, and artificial intelligence with AI initiatives by improving the research in the field of machine learning and deep learning to advance medicine, neuroscience, bioinformatics, healthcare, hpc, hardware, FPGA, data centers, exascale compute power, and genomics. “Future communication networks will be powered by AI. AI will play a big role in defining the 6G, where innovation can fuel the economic wheels of the nation with massive MIMO techniques. When the government, academia, and business industries work together, one can see a future of innovation and a great nation,” he added.