Google What-IF Tool for Code Free ML Visualization

Like the idea of visual tools that map with specific process, resource needs and output results.  Leads to better understandable and resilient results.

Google's What-If Tool And The Future Of Explainable AI
Kalev Leetaru Contributor in Forbes
AI & Big Data

(Excerpt)

" ..... As deep learning has matured sufficiently to find widespread adoption in industry and as developers require increasingly greater understanding of their creations in order to pioneer new advances, the AI community has begun investing heavily in explainable AI as a way to render their black boxes transparent.

Google has been an early leader in emphasizing interpretability and how practitioners can build more understandable, representative and resilient AI solutions. Last year the company unveiled its What-If Tool, which offers a range of interactive visualizations and guided explorations of a TensorFlow model, allowing developers to explore how their model interpreted its training data and how subtle changes to a given input would change its classification, yielding insights into the model’s robustness. .... " 

Google's Description:

The What-If Tool: Code-Free Probing of Machine Learning Models
Tuesday, September 11, 2018
Posted by James Wexler, Software Engineer, Google AI

Building effective machine learning (ML) systems means asking a lot of questions. It's not enough to train a model and walk away. Instead, good practitioners act as detectives, probing to understand their model better: How would changes to a datapoint affect my model’s prediction? Does it perform differently for various groups–for example, historically marginalized people? How diverse is the dataset I am testing my model on?

Answering these kinds of questions isn’t easy. Probing “what if” scenarios often means writing custom, one-off code to analyze a specific model. Not only is this process inefficient, it makes it hard for non-programmers to participate in the process of shaping and improving ML models. One focus of the Google AI PAIR initiative is making it easier for a broad set of people to examine, evaluate, and debug ML systems.


Today, we are launching the What-If Tool, a new feature of the open-source TensorBoard web application, which let users analyze an ML model without writing code. Given pointers to a TensorFlow model and a dataset, the What-If Tool offers an interactive visual interface for exploring model results.  .... " 

Byte Sized Deep Learning Tutorials

There seems to be quite a bit to learn in The AI Workbox.  We often learn best by example rather than rules.

Bite-size video tutorials for Deep Learning Developers

Learn the latest cutting-edge tools and frameworks  Level-up, accomplish more, and do great work ... 

And one particular example (there are many more):

Visualize Training Results With TensorFlow summary and TensorBoard ....

Moving Groceries

Interesting example of the delivery problem.

Space, Time and Groceries
Grocery delivery visualized in python with datashader.

At Instacart, we deliver a lot of groceries. By the end of next year, 80% of American households will be able to use Instacart. Our challenge: complete every delivery on-time, with the right groceries as fast as possible.

Over the course of a week, we traverse cities all over the United States many times over while delivering groceries:

In the remainder of this post, we’ll first introduce the logistics problem Instacart is solving, outline the architecture of our systems and describe the GPS data we collect. Then we will conclude by touring a series of datashader visualizations ...." 

How Viruses Spread

An impressive visual and description of of how viruses spread.  In Quartz   " ... A new genetic tool maps how deadly viruses spread around the world in real time ... ".   Recall a recent post on Kellogg work in using cell data for prediction.  Tag below on Epidemics.   And also links to our work on retail data for the same purpose.

The Future of Programming

Have become involved in looking how algorithmic knowledge is stored lately, and this has required looking at alternative coding methods.  Visualization of process is a key element to describe the related knowledge to decision makers.  The future of coding is important.  In Infoworld.

An Information Cartography

In the CACM: 'Metro maps' of information and its use in narratives.  Video.   A means to understand use architectures?  Fascinating, but I need to see it with real examples.

" .... Several methods have sought to summarize and visualize narratives.2,28,29 However, most work only for simple stories that are linear in nature. In contrast, complex stories exhibit a nonlinear structure; stories spaghetti into branches, side stories, dead ends, and intertwining narratives. To explore them, users need a map to guide them through unfamiliar territory.

We previously introduced a methodology for creating structured summaries of information we call "metro maps." The name is metaphoric; just as cartographic maps have been relied on for centuries to help us understand our surroundings, metro maps help us understand the information landscape. In this article, we explore methods we have developed for automatically creating metro maps of information. ,,, " 

Vinimaya Visual Marketplace Discovery

Mentioned last week in conjunction with delivering visual Marketplaces.  Can you understand your market visually through search?  Impressive approach.

 Vinimaya’s Visual Discovery capabilities provides users of MarketPlace an experience beyond  “Amazon-like” shopping. 
 vMarketPlace Procurement Shopping Experience Enhanced Through Visual Discovery To achieve success with any eProcurement system, it is essential that organizations overcome the challenges of user adoption and content enablement – getting quality catalog content from suppliers quickly into a user-friendly digital format. Employee buy-in and adoption are key to driving spend under management using eProcurement systems.  .... "

Visual Search for eProcurement

Have brought up this method a number of times for minimizing costs. Now new methods include a novel visualization supporting search to find better, cheaper, procurement solutions.  They describe it:

 " ... Vinimaya, Inc., a leader in e-Procurement marketplace and supplier catalog management solutions, announced today that it has added a dynamic Visual Search technology into the latest release of its cloud-based vMarketPlace solution. The new search capability will drive even greater efficiency for users within Vinimaya’s e-Procurement marketplace.

The proprietary Visual Search technology dynamically organizes cross-catalog supplier search results into a real-time, interactive visual heat map and helps corporate shoppers quickly and efficiently find items they are seeking. The heat map analyzes and displays multiple areas of user interest to enhance search relevance. With just a mouse click or two on the color-coded category segments within the heat map, shoppers can drill down to quickly zero in on the exact product they are seeking. ... " 

(Update) Saw this application recently again, and how it related to generalized knowledge and research management.   Interestingly updated.   Worth re-examining.