Do Not Forget to Show Your Work
November 24, 2016
Showing work is messy, necessary step to prove how one arrived at a solution. Most of the time it is never reviewed, but with big data people wonder how computer algorithms arrive at their conclusions. Engadget explains that computers are being forced to prove their results in, “MIT Makes Neural Networks Show Their Work.”
Understanding neural networks is extremely difficult, but MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has developed a way to map the complex systems. CSAIL figured the task out by splitting networks in two smaller modules. One for extracting text segments and scoring according to their length and accordance and the second module predicts the segment’s subject and attempts to classify them. The mapping modules sounds almost as complex as the actual neural networks. To alleviate the stress and add a giggle to their research, CSAIL had the modules analyze beer reviews:
For their test, the team used online reviews from a beer rating website and had their network attempt to rank beers on a 5-star scale based on the brew’s aroma, palate, and appearance, using the site’s written reviews. After training the system, the CSAIL team found that their neural network rated beers based on aroma and appearance the same way that humans did 95 and 96 percent of the time, respectively. On the more subjective field of “palate,” the network agreed with people 80 percent of the time.
One set of data is as good as another to test CSAIL’s network mapping tool. CSAIL hopes to fine tune the machine learning project and use it in breast cancer research to analyze pathologist data.
Whitney Grace, November 24, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
MIT Embraces Google DeepMinds Intuitive Technology Focus
October 6, 2016
The article on MIT Technology Review titled How Google Plans to Solve Artificial Intelligence conveys the exciting world of Google DeepMind’s Labyrinth. Labyrinth is a 3D environment forged on an open-source platform where DeepMind is challeneged by tasks such as, say, finishing a maze. As DeepMind progresses, the challenges become increasingly complicated. The article says,
What passes for smart software today is specialized to a particular task—say, recognizing faces. Hassabis wants to create what he calls general artificial intelligence—something that, like a human, can learn to take on just about any task. He envisions it doing things as diverse as advancing medicine by formulating and testing scientific theories, and bounding around in agile robot bodies…The success of DeepMind’s reinforcement learning has surprised many machine-learning researchers.
Of the endless applications possible for intuitive technology, the article focuses on the medical, understanding text, and robotics. When questioned about the ethical implications of the latter, Demis Hassabis, the head of Google’s DeepMind team, gave the equivalent of a shrug, and said that those sorts of questions were premature. In spite of this, MIT’s Technology Review seems pretty pumped about Google, which makes us wonder whether IBM Watson has been abandoned. Our question for Watson is, what is the deal with MIT?
Chelsea Kerwin, October 6, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Hewlett Packard Makes Haven Commercially Available
July 19, 2016
The article InformationWeek titled HPE’s Machine Learning APIs, MIT’s Sports Analytics Trends: Big Data Roundup analyzes Haven OnDemand, a large part of Hewlett Packard Enterprise’s big data strategy. For a look at the smart software coming out of HP Enterprise, check out this video. The article states,
“HPE’s announcement this week brings HPE Haven OnDemand as a service on the Microsoft Azure platform and provides more than 60 APIs and services that deliver deep learning analytics on a wide range of data, including text, audio, image, social, Web, and video. Customers can start with a freemium service that enables development and testing for free, and grow into a usage and SLA-based commercial model for enterprises.”
You may notice from the video that the visualizations look a great deal like Autonomy IDOL’s visualizations from the early 2000s. That is, dated, especially when compared to visualizations from other firms. But Idol may have a new name: Haven. According to the article, that name is actually a relaxed acronym for Hadoop, Autonomy IDOL, HP Vertica, Enterprise Security Products, and “n” or infinite applications. HPE promises that this cloud platform with machine learning APIs will assist companies in growing mobile and enterprise applications. The question is, “Can 1990s technology provide what 2016 managers expects?”
Chelsea Kerwin, July 19, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark
Web meet up on July 26, 2016.
Information is at this link: http://bit.ly/29tVKpx.
The Machine Learning Textbook
July 19, 2016
Deep learning is another bit of technical jargon floating around and it is tied to artificial intelligence. We know that artificial intelligence is the process of replicating human thought patterns and actions through computer software. Deep learning is…well, what specifically? To get a primer on what deep learning is as well as it’s many applications check out “Deep Learning: An MIT Press Book” by Ian Goodfellow, Yoshua Bengio, and Aaron Courville.
Here is how the Deeping Learning book is described:
“The Deep Learning textbook is a resource intended to help students and practitioners enter the field of machine learning in general and deep learning in particular. The online version of the book is now complete and will remain available online for free. The print version will be available for sale soon.”
This is a fantastic resource to take advantage of. MIT is one of the leading technical schools in the nation, if not the world, and the information that is sponsored by them is more than guaranteed to round out your deep learning foundation. Also it is free, which cannot be beaten. Here is how the book explains the goal of machine learning:
“This book is about a solution to these more intuitive problems. This solution is to allow computers to learn from experience and understand the world in terms of a hierarchy of concepts, with each concept de?ned in terms of its relation to simpler concepts. By gathering knowledge from experience, this approach avoids the need for human operators to formally specify all of the knowledge that the computer needs.”
If you have time take a detour and read the book, or if you want to save time there is always Wikipedia.
Whitney Grace, July 19, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark
Web meet up on July 26, 2016.
Information is at this link: http://bit.ly/29tVKpx.
The Computer Chip Inspired by a Brain
July 6, 2016
Artificial intelligence is humanity’s attempt to replicate the complicated thought processes in their own brains through technology. IBM is trying to duplicate the human brain and they have been successful in many ways with supercomputer Watson. The Tech Republic reports that IBM has another success under their belt, except to what end? Check out the article, “IBM’s Brain-Inspired Chip TrueNorth Changes How Computers ‘Think,’ But Experts Question Its Purpose.”
IBM’s TrueNorth is the first computer chip with an one million neuron architecture. The chip is a collaboration between Cornell University and IBM with the BARPA SyNAPSE Program, using $100 million in public funding. Most computer chips use the Von Neumann architecture, but the TrueNorth chip better replicates the human brain. TrueNorth is also more energy efficient.
What is the purpose of the TrueNorth chip, however? IBM created an elaborate ecosystem that uses many state of the art processes, but people are still wondering what the real world applications are:
“ ‘…it provides ‘energy-efficient, always-on content generation for wearables, IoT devices, smartphones.’ It can also give ‘real-time contextual understanding in automobiles, robotics, medical imagers, and cameras.’ And, most importantly, he said, it can ‘provide volume-efficient, unprecedented neural network acceleration capability per unit volume for cloud-based streaming processing and provide volume, energy, and speed efficient multi-modal sensor fusion at an unprecedented neural network scale.’”
Other applications include cyber security, other defense goals, and large scale computing and hardware running on the cloud. While there might be practical applications, people still want to know why IBM made the chip?
” ‘It would be as if Henry Ford decided in 1920 that since he had managed to efficiently build a car, we would try to design a car that would take us to the moon,’ [said Nir Shavit, a professor at MIT’s Computer Science and Artificial Intelligence Laboratory]. ‘We know how to fabricate really efficient computer chips. But is this going to move us towards Human quality neural computation?’ Shavit fears that its simply too early to try to build neuromorphic chips. We should instead try much harder to understand how real neural networks compute.’”
Why would a car need to go to the moon? It would be fun to go to the moon, but it doesn’t solve a practical purpose (unless we build a civilization on the moon, although we are a long way from that). It continues:
” ‘The problem is,’ Shavit said, ‘that we don’t even know what the problem is. We don’t know what has to happen to a car to make the car go to the moon. It’s perhaps different technology that you need. But this is where neuromorphic computing is.’”
In other words, it is the theoretical physics of computer science.
Whitney Grace, July 6, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Drone and Balloon WiFi Coming to the Sky near You
November 10, 2015
Google and Facebook have put their differences aside to expand Internet access to four billion people. Technology Review explains in “Facebook;s Internet Drone Team Is Collaborating With Google’s Stratospheric Balloons Project” how both companies have filed documented with the US Federal Communications Commission to push international law to make it easier to have aircraft fly 12.5 miles or 20 kilometers above the Earth, placing it in the stratosphere.
Google has been working on balloons that float in the stratosphere that function as aerial cell towers and Facebook is designing drones the size of aircraft that are tethered to the ground that serve the same purpose. While the companies are working together, they will not state how. Both Google and Facebook are working on similar projects, but the aerial cell towers marks a joint effort where they putting aside their difference (for the most part) to improve information access.
“However, even if Google and Facebook work together, corporations alone cannot truly spread Internet access as widely as is needed to promote equitable access to education and other necessities, says Nicholas Negroponte, a professor at MIT’s Media Lab and founder of the One Laptop Per Child Project. ‘I think that connectivity will become a human right,’ said Negroponte, opening the session at which Facebook and Google’s Maguire and DeVaul spoke. Ensuring that everyone gets that right requires the Internet to be operated similar to public roads, and provided by governments, he said.”
Quality Internet access not only could curb poor education, but it could also improve daily living. People in developing countries would be able to browse information to remedy solutions and even combat traditional practices that do more harm than good.
Some of the biggest obstacles will be who will maintain the aerial cell towers and also if they will pose any sort of environmental danger.
Whitney Grace, November 10, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
The PurePower Geared Turbofan Little Engine That Could
October 29, 2015
The article on Bloomberg Business titled The Little Gear That Could Reshape the Jet Engine conveys the 30 year history of Pratt & Whitney’s new PurePower Geared Turbofan aircraft engines. These are impressive machines, they burn less fuel, pollute less, and produce 75% less noise. But thirty years in the making? The article explains,
“In Pratt’s case, it required the cooperation of hundreds of engineers across the company, a $10 billion investment commitment from management, and, above all, the buy-in of aircraft makers and airlines, which had to be convinced that the engine would be both safe and durable. “It’s the antithesis of a Silicon Valley innovation,” says Alan Epstein, a retired MIT professor who is the company’s vice president for technology and the environment. “The Silicon Valley guys seem to have the attention span of 3-year-olds.”
It is difficult to imagine what, if anything, “Silicon Valley guys” might develop if they spent three decades researching, collaborating, and testing a single project. Even more so because of the planned obsalesence of their typical products seeming to speed up every year. In the case of this engine, the article suggests that the time spent has positives and negatives for the company- certain opportunities with big clients were lost along the way, but the dedicated effort also attracted new clients.
Chelsea Kerwin, October 29, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Compare Cell Phone Usage in Various Cities
October 8, 2015
Ever wonder how cell phone usage varies around the globe? Gizmodo reports on a tool that can tell us, called ManyCities, in their article, “This Website Lets You Study Cell Phone Use in Cities Around the World.” The project is a team effort from MIT’s SENSEable City Laboratory and networking firm Ericsson. Writer Jamie Condliffe tells us that ManyCities:
“…compiles mobile phone data — such as text message traffic, number of phone calls, and the amount of data downloaded —from base stations in Los Angeles, New York, London, and Hong Kong between April 2013 and January 2014. It’s all anonymised, so there’s no sensitive information on display, but there is enough data to understand usage patterns, even down the scale of small neighbourhoods. What’s nice about the site is that there are plenty of intuitive interpretations of the data available from the get-go. So, you can see how phone use varies geographically, say, or by time, spotting the general upward trend in data use or how holidays affect the number of phone calls. And then you can dig deeper, to compare data use over time between different neighbourhoods or cities: like, how does the number of texts sent in Hong Kong compare to New York? (It peaks in Hong Kong in the morning, but in the evening in New York, by the way.)”
The software includes some tools that go a little further, as well; users can cluster areas by usage patterns or incorporate demographic data. Condliffe notes that this information could help with a lot of tasks; forecasting activity and demand, for example. If only it were available in real time, he laments, though he predicts that will happen soon. Stay tuned.
Cynthia Murrell, October 8, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Googles Chauvinistic Job Advertising Delivery
July 28, 2015
I thought we were working to get more women into the tech industry, not fewer. That’s why it was so disappointing to read, “Google Found to Specifically Target Men Over Women When It Comes to High-Paid Job Adverts” at IBTimes. It was a tool dubbed AdFisher, developed by some curious folks at Carnegie Mellon and the International Computer Science Institute, that confirmed the disparity. Knowing that internet-usage tracking determines what ads each of us sees, the researchers wondered whether such “tailored ad experiences” were limiting employment opportunities for half the population. Reporter Alistair Charlton writes:
“AdFisher works by acting as thousands of web users, each taking a carefully chosen route across the internet in such a way that an ad-targeting network like Google Ads will infer certain interests and characteristics from them. The programme then records which adverts are displayed when it later visits a news website that uses Google’s ad network. It can be set to act as a man or woman, then flag any differences in the adverts it is shown.
“Anupam Datta, an associate professor at Carnegie Mellon University, said in the MIT Technology Review: ‘I think our findings suggest that there are parts of the ad ecosystem where kinds of discrimination are beginning to emerge and there is a lack of transparency. This is concerning from a societal standpoint.’”
Indeed it is, good sir. The team has now turned AdFisher’s attention to Microsoft’s Bing; will that search platform prove to be just as chauvinistic? For Google’s part, they say they’re looking into the study’s methodology to “understand its findings.” It remains to be seen what sort of parent the search giant will be; will it simply defend its algorithmic offspring, or demand it mend its ways?
Cynthia Murrell, July 28, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Algorithmic Art Historians
July 14, 2015
Apparently, creativity itself is no longer subjective. MIT Technology Review announces, “Machine Vision Algorithm Chooses the Most Creative Paintings in History.” Traditionally, art historians judge how creative a work is based on its novelty and its influence on subsequent artists. The article notes that this is a challenging task, requiring an encyclopedic knowledge of art history and the judgement to decide what is novel and what has been influential. Now, a team at Rutgers University has developed an algorithm they say is qualified for the job.
Researchers Ahmed Elgammal and Babak Saleh credit several developments with bringing AI to this point. First, we’ve recently seen several breakthroughs in machine understanding of visual concepts, called classemes. that include recognition of factors from colors to specific objects. Another important factor: there now exist well-populated online artwork databases that the algorithms can, um, study. The article continues:
“The problem is to work out which paintings are the most novel compared to others that have gone before and then determine how many paintings in the future have uses similar features to work out their influence. Elgammal and Saleh approach this as a problem of network science. Their idea is to treat the history of art as a network in which each painting links to similar paintings in the future and is linked to by similar paintings from the past. The problem of determining the most creative is then one of working out when certain patterns of classemes first appear and how these patterns are adopted in the future. …
“The problem of finding the most creative paintings is similar to the problem of finding the most influential person on a social network, or the most important station in a city’s metro system or super spreaders of disease. These have become standard problems in network theory in recent years, and now Elgammal and Saleh apply it to creativity networks for the first time.”
Just what we needed. I have to admit the technology is quite intriguing, but I wonder: Will all creative human endeavors eventually have their algorithmic counterparts and, if so, how will that effect human expression?
Cynthia Murrell, July 14, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
-
- Subscribe to Beyond Search
Feature archive
News archive
-
