Welcome to our weekly selection of digital innovation news. Created using our opinionated automated selection algorithm with a twisted text rank summary creator. We present some top innovation news items to get you thinking, debating and take action in order to make our world even better.
1 DeepNude app harms the most vulnerable
An app called DeepNude that “undressed” photos of women, using generative adversarial networks to swap their clothes for highly realistic nude bodies, has been shut down after an outcry. However, the story feeds into growing fears over the unique way deepfakes can be weaponized against women and other vulnerable populations. The app only generated images of the female body, even when given a picture of a man. Deepfakes are not a new threat; mass manipulated media has been around since long personal computers exist. But the technology has accelerated and broadened existing trends.
2 Predicting Bus Delays with Machine Learning
Google Maps introduced live traffic delays for buses, forecasting bus delays in hundreds of cities world- wide, ranging from Atlanta to Zagreb to Istanbul to Manila and more. If we run the model on that day’s car traffic data, it gives us the green predictions below. Even within a small neighborhood, the model needs to translate car speed predictions into bus speeds differently on different streets. In the left panel below, we color-code our model’s predicted ratio between car speeds and bus speeds for a bus trip.
3 Drag-and-drop data analytics
In the same vein, researchers from MIT and Brown University have now developed a system for interactive data analytics that runs on touchscreens and lets everyone — not just billionaire tech geniuses — tackle real-world issues. Next, the researchers are looking to add a feature that alerts users to potential data bias or errors. Over four years, the researchers have published numerous papers detailing components of Northstar, including the interactive interface, operations on multiple platforms, accelerating results, and studies on user behavior. Users feed the system datasets, and manipulate, combine, and extract features on a user- friendly interface, using their fingers or a digital pen, to uncover trends and patterns. All data are stored and analyzed in the cloud. The researchers like to demonstrate the system on a public dataset that contains information on intensive care unit patients.
4 Ctizen science: Introducing the Cooperative Open Online Landslide Repository (COOLR)
The COOLR project (https://landslides.nasa.gov) contains Landslide Reporter, the first global citizen science project for landslides, and Landslide Viewer, a portal to visualize data from COOLR and other satellite and model products. The new Cooperative Open Online Landslide Repository (COOLR) supplements data in a NASA-developed Global Landslide Catalog (GLC) with citizen science reports to build a more robust, publicly available global inventory. COOLR continues to expand as an open platform of landslide inventories with new data from citizen scientists, NASA scientists, and other landslide groups. However, collecting landslide events in inventories is difficult at the global scale due to inconsistencies in or the absence of landslide reporting. Future work on the COOLR project will seek to increase participation and functionality of the platform as well as move towards collective post-disaster mapping.
5 New AI programming language goes beyond deep learning
In probabilistic AI, inference algorithms perform operations on data and continuously readjust probabilities based on new data to make predictions. Some other probabilistic programming systems are flexible enough to cover several kinds of AI techniques, but they run inefficiently. But there are plenty of other AI techniques available today, such as statistical and probabilistic models, and simulation engines. Gen does the same for probabilistic programming.” Gen’s source code is publicly available and is being presented at upcoming open-source developer conferences, including Strange Loop and JuliaCon. Gen also provides high-level infrastructure for inference tasks, using diverse approaches such as optimization, variational inference, certain probabilistic methods, and deep learning.
6 Testing Concurrent Systems
One common observation is that test planning and implementation does not explicitly address the development of test cases designed to uncover concurrency defects. This blog post ends with a set of recommendations for improving the effectiveness and efficiency of concurrency testing. In this blog post, I explore the different types of concurrency, the resulting types of concurrency defects, and the most important testing techniques for uncovering these defects. At the SEI, we are often called upon to review development planning documents including Test and Evaluation Master Plans (TEMPs) and Software Test Plans (STPs). Concurrency leads to nondeterministic behavior and numerous types of concurrency defects that require specialized approaches to uncover.
TensorWatch (from Microsoft, MIT licensed) is a debugging and visualization tool designed for data science, deep learning and reinforcement learning from Microsoft Research. It works in Jupyter Notebook to show real-time visualizations of your machine learning training and perform several other key analysis tasks for your models and data. TensorWatch is designed to be flexible and extensible so you can also build your own custom visualizations, UIs, and dashboards. Besides traditional “what-you-see-is-what-you-log” approach, it also has a unique capability to execute arbitrary queries against your live ML training process.
The Radical Open Innovation weekly overview is a brief overview of innovation news on Digital Innovation and Management Innovation from all over the world. Your input for our next edition is welcome! Send it to [info] at [bm-support]dot[org]