Workshop Content

Workshop Agenda

Pre-requisite downloads Some of the workshops below require that you download certain software, such as node.js or cURL plus creating an account with the Weather Company. Please begin with this document that resides on Github, as do all other workshops. Once you download the PDF from Github, then the links within the document will become active.

Lecture, demo, QA (1.5 hours) Let us explore the cognitive era and the role that algorithms play in the API economy: Netflix recommends movies from the long tail, Amazon recommends the popular, cars that learn when to take over a vehicle that is in front, robots that distinguish trash from dishes to be washed. In this lecture, let's uncover the blackbox that lurks behind artificial neural networks powering IBM Watson cognitive services that enables the magic of machine learning. let's construct the stochastic gradient descent algorithm. How fitting that we will end the lecture with a glimpse into year 2025: the rise of the bots and the sunset of apps, natural language processing (NLP) has lead to natural language undertanding (NLU), and blockchain is now in the fabric of bots. The presentation PDF varies as do the universities, venues and the focus groups; therefore, it is only available upon request from the instructor.

Face Recognition (1.0 hour) In this workshop you will build an app from scratch using the graphical interface of Node-RED. Your app will use IBM Watson Visual Recognition service to identify faces of people; if those people happen to be in the millions of faces that the service is trained on, it will reveal their name; otherwise, distinguish gender and approximate age, with the confidence value visible. Try it with an image of multiple people.

Earthquake detection (45 minutes) More fun to be had with Node-RED with a bit of an IoT twist and you get to install an open-sourced node from the community called worldmap. Let's discover just where in the world...real-time...an Earthquake is taking place.

Train your custom images (2.0 hours) In this workshop, you can go a step further and upload your own set of images (positives and negatives) and train the system to distinguish various dog breeds or a broken windshield due to vandalism versus an accident.

Build a simple Conversation In this workshop, you build a simple Conversation chatbot or concierge and in the process gain an understanding of Intents, Entities and Dialog.

Advanced Conversation In this workshop, you build and train a cognitive chatbot, or better yet, a smart virtual help desk. You will first specify the intent; second, capture the entity that further narrows down your intent and finally, design a dialog. Additionally, you will connect this chat bot to an external system (the Weather Company) and obtain the actual weather forecast at your location. For a quick glance at the PDF, click here. For topics such as slots and context variables, watch this video.

Discovery With this new flagship cognitive service, you ingest, enrich, normalize and query you personal and public documents. Specifically, the Discovery APIs allow you to ingest and enrich documents that contain structured and unstructured data, and its query language allows you to carry out analytics on the resulting corpus. You can also query pre-enriched content sources, such as the News service. For a quick glance at the PDF, click here. ...and this one for Discovery video

Virtual Agent This workshop combines the cognitive capabilities of the Conversation and the Discovery services. You must complete the prior two workshops before beginning the exercises in this lab. To acceses the instructions, click here.

You can also click
here to access the car conversation JSON workspace and all 101 AirBnB JSON files.
Watch this demo to see the combined services in action video

Empathetic chatbot This is a simple conversation to demonstrate how a user's emotional tone can be used to provide more tailored and empathetic responses by integrating Watson Conversation and Watson Tone Analyzer. Use the README within the package for instructions until the detailed PDF is published here. Try the demo for yourself right from here.