All rights reserve. 5. Industry Challenge "Predicting ad click–through rates (CTR) is a massive-scale learning problem that is central to the multi-billion dollar online advertising industry" Google Context Ad platforms collect huge data to help them predict ad clicks A good predictive model is essential to serve ads efficiently to optimize over all economic value Sponsored search advertising, contextual advertising, display advertising, and real- time bidding auctions have all relied heavily on the ability of learned models to predict ad click–through rates accurately, quickly, and reliably 5 Private and confidential.

Role of RPA and AI in automation Why RPA and AI implementations don't achieve their goals? Planning and implementing automation initiatives that deliver high ROI How can you maximize the benefits from RPA and AI in your automation initiatives? 1 2 3 7.

Business Sensitive Bot Great market opportunity Guaranteed success Bot framework Inference Framework domain services Let's get started If you are interested, contact us 25. Pramati's M&A's of leading products Innovation enablement Serving over 200 product companies Agile methodology User-centric design Serving from 5 global locations Products built from conception-code-cash Unique products & services 1300 engineers About Imaginea Private and confidential.

Front- end Bot Patient Databases No Update Create authorization request Bot pre-fills authorization request, reviewer completes it and submits Check status Update status Yes No Yes 1 34 5 6 7 8 9 10 26.

All rights reserve. 5. DEEP LEARNING JOURNEY SO FAR 1980 Training them becomes difficult due to inadequate compute & dataThe power of neural networks is recognized GPUs make the required compute easily accessible for a broad range of problems Vast amounts of data flow through the internet Techniques for training "deep" networks open up new opportunities "Convolutional" networks – DCNNs, in particular – have changed the image analysis landscape 1990 2012 2015 2017 5Private and confidential. DEEP NEURAL NETWORK SHALLOW LAYERS Sensory features DEEP LAYERS Higher cognitive features Deep Neural Networks 10Private and confidential. FONTLI: 45,000 images used to train the network 14Private and confidential. Which whale is it, anyway? Face recognition to identify whales using deep learning A non-random sample from the dataset 4544 training images only & 447 categories 15Private and confidential.