Daily Tech Digest - December 17, 2016

Bitcoin Is Being Monitored by An Increasingly Wary US Government

This ability to financially disrupt, disable and dismantle nefarious networks, is crucial to U.S. national security, Treasury officials say. It has proven effective for more than a decade and is often strongly preferable to deploying troops. “We have made it very difficult for members of the Islamic State to raise or move money around the world these days,” Zarate says. “Even Iran had a hard time finding safe havens.” In fact, years of financial pressure from the U.S. and its allies helped force Iran to negotiate with the White House and sign a landmark nuclear deal last year. The biggest concern the U.S. has about virtual currencies, Zarate says, is that terrorists and other enemies might create one so powerful and so untrackable, that they’ll no longer need the global banking system, which the U.S. uses to financially starve them.


10 Steps to Train a Chatbot and its Machine Learning Models to Maximize Performance

The Watson services rely on a variety of machine learning algorithms, most of which fall in the supervised machine learning category, which learn the specifics of the problem from sample labeled data and help make predictions on unlabeled data. Training a supervised machine learning system involves providing it with representative inputs and corresponding outputs and the system will learn by example. These pairs of representative inputs/outputs constitute the “groundtruth” from which the system learns. ... Training NLC would require providing a groundtruth which includes representative utterances (input) and the corresponding intents (output). NLC would then learn which utterances map to which intents. Note that it not only will be able to extract intent from utterances it has seen but it can also extract intent from any utterance based on similarity of such an utterance to what is available in the training data.


Talent Development for the Digital World

From a skills perspective, innovation and learning ability are becoming key requirements. Innovation is not limited to products anymore: It cuts across processes, organization design, reviews, performance management, and rewards. Hence this has become a requirement across all functions within the organization. While the debate on whether innovation is a trait or a skill that can be developed is still raging, the need for it is only burgeoning. In a tongue-in-cheek manner, we can certainly say that learning new skills is necessary to a company’s success. How organizations can create a platform that propels employees to learn and adapt is becoming a key success factor in reskilling the existing workforce and preparing them for the digital future.


How to Use Thought Experiments to De-Risk Your Startup

Sometimes you miss signs that things aren't working. Maybe engineers keep going to Google or Twitter instead of accepting your job offers, or maybe very few customers agree to see your demo after an initial discovery call. When you're scrambling day-to-day, you might think, "if 5% of people want to see a demo, then I should call at least 40 people daily." Months later, you realize that the low demo rate was a sign that your product didn't fit the market's needs. It's helpful to step back and ask yourself if the things that you're struggling with today are a sign that you need to optimize or double down on your processes, or if they're a sign of something more significant, like working on the wrong product or targeting the wrong job candidates.


Artificial intelligence creeps into daily life

A self-driving car, for example, can easily navigate around Google's home base in Mountain View, California, but may have more problems around the Arc de Triomphe in Paris, where driving behaviors are less predictable. Alahi said robotics needs to understand the unwritten social behaviors used in daily life, which can vary from one culture to another. A robot, for example, might cut through a group of people in a train station to find the most efficient path, unknowingly violating social rules on personal space. "There are situations where technology is not yet capable of understanding human behavior," said Alahi, who is part of a research project using a robot, with the aim of understanding pedestrian behavior. These kinds of robots may be technological marvels, but they also raise fears that they could get out of control, concerns heightened by movies like "Terminator."


Worm on the sensor: What happens when IoT data is bad?

The harsher the surrounding conditions and the more isolated the device, the worse the bad-data problem is likely to be. In addition to agriculture, industries like oil and gas and energy distribution face this. But it’s not just far-flung sensors that have problems. Even in a hospital, a blood oxygen sensor clamped on a patient’s finger can start giving bad data if it gets bumped into the wrong position. On top of that, some IoT devices malfunction on their own and start spewing out bad data, or stop reporting at all. In many other cases, human error is the culprit: The wrong settings mess up what the device generates. ... John Deere equips its giant farm tools with sensors that detect whether the machines are working right. The company’s ExactEmerge planter, which rolls behind a tractor planting seeds across a field, has three sensors per row of crops to detect how many seeds are being planted and at what rate.


Africa 2017: Smartphone penetration, Open Data and less online freedom

The demand for cheap smartphones is boosting penetration rates and is affecting the data bundle business for the majority of telecom companies. Safaricom, Kenya’s largest telecom company has seen its profit shoot up through mobile internet services. The company said during its half year result ending September 2016 that: “Mobile data revenue, which accounts for 13.7% of the firm’s service revenue, grew at 46.3% to Sh13.4 billion (US$134 million), driven by growth in active mobile data customers to 14.9 million, increased bundle users and smartphone penetration.” According to research firm Ovum, the smartphone penetration rate will grow at 52.9% year- on-year. Currently there are 293.8 million smartphone users across the continent. Ovum predicts that there will be 929.9 million smartphones by the year 2021.


Self-Driven Car Simulator Using a Neural Network and Genetic Algorithm Training

Artificial Intelligence impacts human life in many ways nowadays. An example is in the auto industry; many companies are trying to make their cars smarter. The cars can self-drive, avoid obstacles, find destinations … without controls from human. This paper is about a car simulation program in which the car itself will move without any controls from outside. The approach uses a Neural Network and a Genetic Algorithm to train the car by making the car learn after each time it fails to finish the track. ... Every computer has different speeds so we need a mechanism to normalize that to make the game run at the same speed in any computer. We have 2 main methods: update and render. Usually, the game runs at 60 fps so the update method will be called 60 times per second and the render method will be called as fast as the computer’s speed.


Getting Started With JMeter: A Basic Tutorial

Performance testing and load testing are the practices of ensuring that websites and apps perform under heavy loads, from different geolocations, and for different user scenarios. If you followed the Pokemon Go craze or heard about Macy’s crashing during the last Black Friday, you know the importance of performance testing and how crucial it is for businesses. Poor performance, whether website crashes or slow page loading, equals an immediate and long-term loss of revenue, as it creates a bad reputation and immediate churn. ... Open-source and JAVA-based, JMeter simulates browser behavior (though it’s not a browser!) by sending requests to web or application servers for different loads. JMeter can also parse the responses. On your local machine, you can scale up to approximately 100 virtual users, but you can go up to more than 1,000,000 VUs with CA BlazeMeter, which is kind of a JMeter in the cloud.


Why Artificial Intelligence Will Be Analog

In essence, analog is similar to the human brain. Consider the human body in comparison to a fitness monitor, like a FitBit. These devices are both analog and digital. The analog sensors are what collects the data about the number of steps your take, your heart rate, etc. That data is then converted using an analog-to-digital converter, so that the readings can be more easily processed using the algorithms in the microprocessor of the device, putting the raw data into a form that we can use. Over time, the device “learns” your patterns, so it can make recommendations as to the number of steps you should take, how much sleep you need, etc. This is a form of artificial intelligence, and the same concepts apply to other AI devices as well. Essentially, as science historian George Dyson points out, the brain itself is an analog computer. Our senses take in information in analog format, whether a sight, sound, smell, etc., which is then processed by the neural pathways of the brain.



Quote for the day:


"Judge your success by what you had to give up in order to get it." -- Unknown