In this post on my series on "Optimization Methods for Large-Scale Machine Learning" by Bottou, Curtis, and Nocedal, I want to focus on model building in machine learning. Section 2 of the paper describes several case studies, with the purpose of showing how "the process of machine learning leads to the selection of a prediction function through solving an optimization problem." A prediction function is a mathematical function that links the model inputs to the quantity we wish to predict. From the practitioner's point of view, a prediction function is implicitly specified by the technique the data scientist has chosen (for example, regression or neural networks) and trained model parameters (what is actually learned when the technique is applied to data). For example, the structure of a neural network amounts to a description of a family of related functions.

The world's bestselling drone-maker has unveiled models that warn their operators when there is a risk posed by nearby aeroplanes or helicopters. The M200 series use ADS-B (automatic dependent surveillance broadcast) receivers to detect broadcasts from nearby manned flights. The transmissions allow users to see the position, altitude and velocity of surrounding aircraft so they can take evasive action if required. "It's an extra safety measure and will help drone operators work in restricted airspace," explained Dave Black from the commercial drone services firm Blackwing Aerial. "The way we tend to do this is to contact air traffic controllers before we fly, tell them where we are flying and then they contact us by phone if there's going to be any conflict. "But in time, the authorities may well also want this kind of feature fitted as standard before they approve operations with larger drones." Mr Black added that some enthusiasts had created their own ADS-B kit by connecting a receiver to a Raspberry Pi computer with customised software and then tying this to their drone. But he added that an integrated unit would be appreciated. The announcement was made at the Mobile World Congress trade show in Barcelona. DJI announced last July that it intended to develop an ADS-B collision avoidance system in conjunction with uAvionix, a specialist in the technology. The M200 series drones, which debut the feature, are designed for industrial applications such as inspecting power lines and mapping construction sites. They have a range of 7km (4.3 miles) and can stay airborne for up to 38 minutes, but some country's regulators currently require them to stay within line-of-sight. The aircraft are bigger and heavier than many consumer drones - they weigh 3.8kg (8.4lb) and can carry a further 2kg of equipment - so potentially pose a greater risk if involved in a crash. However, one analyst warned that the new safety system could give drone pilots a false sense of confidence. "The problem is that ADS-B isn't yet required for all aircraft, so even if you had a receiver you're not necessarily seeing all the traffic in the area," explained Colin Snow, chief executive of the Skylogic Research consultancy. "Its use is also not required at the low altitudes at which commercial drones tend to fly.

Tip One: Do Both Empire State and Top of Rock But do one in the morning and one as the sun is setting to see the city in a different light. I was so grateful for this tip.

A small insurance startup, Root, has launched a car insurance specifically designed for Tesla vehicle owners that reduces the price of the policy the longer the vehicle runs in autonomous mode, on the basis that this mode is much safer than driving manually. Thus, drivers who spend a lot of time on the highway or in conditions where they can activate the autonomous mode will pay less insurance. The idea is based on the fact that a vehicle is increasingly a connected platform, a smartphone on wheels from which we can obtain a constant flow of information. To sign up for a Root policy, which typically offers much lower prices than its competitors, you must download an app that allows the company to access GPS, accelerometers and gyroscopes data on your smartphone, making it possible for the company to evaluate your driving. After about two to three weeks driving with the app, enough for the average driver to forget about the app and go back to his or her typical driving habits, the algorithm has created a user profile that includes how much time the vehicle is in use, frequent destinations, whether drivers change lane excessively, their driving speeds, to what extent they respect traffic rules, or if they use their smartphone while driving, among many other things.

Machine learning is a trending topic today and for good reason. It has enormous potential to transform entire markets and industries. Successful machine-learning startups will be the ones targeting vertical applications with a clear need for the technology. The consumer packaged goods industry is a good example. Machine learning can more accurately predict inventory levels to better manage the supply chain, reduce inventory costs, minimize excess capacity requirements, and eliminate stockouts.

Until a few years ago, computers couldn't tell the difference between images of dogs and ones of cats. Today, computer programs use machine-learning algorithms to study piles of data and learn about the world and its people. Algorithms can tell a banker whether someone will pay back a loan. They can pick the right applicant out of thousands of resumes. They assist judges to determine a prisoner's risk of returning to a life of crime.

At this year's SXSW the theme for our conference tech coverage is "AI for good." Technologists will be discussing how they are using new developments in artificial intelligence to solve complex, long-standing social problems, including crimes against children. Federico Gomez Suarez (pictured), senior technical program manager at Microsoft, spoke to John Furrier (@furrier), host of theCUBE, SiliconANGLE Media's mobile live streaming studio, during SXSW about his involvement in Thorn, a special outfit that works with law enforcement to rescue traffic or exploited children. "The fact that we're making a difference in those lives is extremely encouraging," Suarez said. He credits Microsoft's Hack for Good program for paving his avenue to Thorn.

As business opportunities for artificial intelligence multiply, how can industry leaders design the principles and technical standards into their products that benefit society as a whole?

For Silicon Valley, the headline was sweet nectar: Google DeepMind, the world's hottest artificial intelligence lab, embraces the blockchain, the endlessly fascinating idea at the heart of the bitcoin digital currency. The lab's re-imagining of the blockchain has very little to do with AI--or the blockchain, for that matter. If you want AI crossed with the blockchain, try wrapping your head around Numerai, the world's strangest hedge fund. To DeepMind's credit, its new project depends less on trendy ideas than an apparent desire to solve a real problem in the real world--one that involves the most private and personal information. DeepMind is building an auditing system for healthcare data. That may not sound sexy, but it matters.

The Indus script has long challenged epigraphists because of the difficulty in reading and classifying text and symbols on the artefacts. Now, a Chennai-based team of scientists has built a programme which eases the process. Ronojoy Adhikari of The Institute of Mathematical Sciences and Satish Palaniappan, who is at Sri Sivasubramaniya Nadar College of Engineering, have developed a "deep-learning" algorithm that can read the Indus script from images of artefacts such as a seal or pottery that contain Indus writing. Scanning the image, the algorithm smartly "recognises" the region of the image that contains the script, breaks it up into individual graphemes (the term in linguistics for the smallest unit of the script) and finally identifies these using data from a standard corpus. In linguistics the term corpus is used to describe a large collection of texts which, among other things, are used to carry out statistical analyses of languages.