These Emojis Are Logical
August 9, 2016
Emojis are a secondary language for many people, especially the younger sect, and whole messages can be conveyed within a few images. Someone needs to write an algorithm to translate emoji only messages, but machine learning has not yet reached the point where it can understand all the intricacies associated with emojis. Or has it? TechCrunch shares that “Dango Mind-Melds With Emoji Using Deeping Learning And Suggests Them While They Type.”
Dango is an emoji suggestion chatbot. Unlike the Microsoft chatbot that became anti-Semitic and misogynist in a matter of hour, Dango just wants to give you emoji suggestions to pep up your messages:
“Okay, so Dango is one of those virtual assistants that lives in your chat apps, and this one is based on a neural network that has been trained with millions of examples to understand what emoji mean. So not only can it suggest an appropriate one, but it can translate entire sentences. Its icon is a weird piece of cute cake, which sits above your keyboard watching you type. It’s free for Android right now, with an iOS version coming out eventually.”
Aww, it’s a little cake icon that sits above your keyboard. Is it not tempting already to download it make Dango your friend? The cute factor comes after the deep machine learning took place.
The Dango programmers used a recurrent neural network to teach Dango how to decipher the meaning of emoji. It would guess, then check against real world examples, then adjust its parameters when it was wrong. The guesses were assembled in a “semantic space” that relates the emojis to concepts (check the article for the visualization).
Dango is constantly updating itself to be on top of the latest slang and memes, including the negative aspects of the language. Dango is still learning, especially when it comes to translating entire sentences to pictures. Before you say that the written language cannot be replicated in little images, it was done eons ago by Egyptians, Sumerians, Phoenicians, and still by the Chinese, Japanese, and other Asian cultures.
Whitney Grace, August 9, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden /Dark Web meet up on August 23, 2016.
Information is at this link: https://www.meetup.com/Louisville-Hidden-Dark-Web-Meetup/events/233019199/
Summize, an App with the Technology to Make Our Children Learn. But Is They?
August 2, 2016
The article on TheNextWeb titled Teenagers Have Built a Summary App that Could Help Students Ace Exams might be difficult to read over the sound of a million teachers weeping into their syllabi. It’s no shock that students hate to read, and there is even some cause for alarm over the sheer amount of reading that some graduate students are expected to complete. But for middle schoolers, high schoolers, and even undergrads in college, there is a growing concern about the average reading comprehension level. This new app can only make matters worse by removing a student’s incentive to absorb the material and decide for themselves what is important. The article describes the app,
“Available for iOS, Summize is an intelligent summary generator that will automatically recap the contents of any textbook page (or news article) you take a photo of with your smartphone. The app also supports concept, keyword and bias analysis, which breaks down the summaries to make them more accessible. With this feature, users can easily isolate concepts and keywords from the rest of the text to focus precisely on the material that matters the most to them.”
There is nothing wrong with any of this if it is really about time management instead of supporting illiteracy and lazy study habits. This app is the result of the efforts of an 18-year-old Rami Ghanem using optical character recognition software. A product of the era of No Child Left Behind, not coincidentally, exposed to years of teaching to the test and forgetting the lesson, of rote memorization in favor of analysis and understanding. Yes, with Summize, little Jimmy might ace the test. But shouldn’t an education be more than talking point mcnuggets?
Chelsea Kerwin, August 2, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Microsoft Cortana Update Draws Users to Bing
January 22, 2016
The article titled Microsoft Updates Windows 10 Cortana With New Search Tools for Better Results on IB Times heralds the first good news for Bing in ages. The updates Microsoft implemented provide tremendous search power to users and focused search through a selection of filters. Previously, Cortana would search in every direction, but the filters enable a more targeted search for, say, applications instead of web results. The article explains,
“It’s a small change, but one that shows Microsoft’s dedication to making the assistant as useful as possible. Cortana is powered by Bing, so any improvements to the Windows 10 assistant will encourage more consumers to use Microsoft’s search engine. Microsoft made a big bet when it chose to deeply integrate Bing into Windows 10, and there is signs that it’s paying off. After the June 2015 Windows 10 launch, Bing attained profitability for the first time in October 2015.”
That positive note for Bing is deeply hedged on the company’s ability to improve mobile search, which has continued to grow as a major search platform while desktop search actually peaked, according to research. Microsoft launched Cortana on Android and iOS, but it is yet to be seen whether this was sufficient action to keep up the Bing momentum.
Chelsea Kerwin, January 22, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Google Continues to Improve Voice Search
November 5, 2015
Google’s research arm continues to make progress on voice search. The Google Research Blog updates us in, “Google Voice Search: Faster and More Accurate.” The Google Speech Team begins by referring back to 2012, when they announced their Deep Neural Network approach. They have since built on that concept; the team now employs a couple of models built upon recurrent neural networks, which they note are fast and accurate: connectionist temporal classification and sequence discriminative (machine) training techniques. The write-up goes into detail about how speech recognizers work and what makes their latest iteration the best yet. I found the technical explanation fascinating, but it is too lengthy to describe here; please see the post for those details.
I am still struck when I see any article mention that an algorithm has taken the initiative. This time, researchers had to rein in their model’s insightful decision:
“We now had a faster and more accurate acoustic model and were excited to launch it on real voice traffic. However, we had to solve another problem – the model was delaying its phoneme predictions by about 300 milliseconds: it had just learned it could make better predictions by listening further ahead in the speech signal! This was smart, but it would mean extra latency for our users, which was not acceptable. We solved this problem by training the model to output phoneme predictions much closer to the ground-truth timing of the speech.”
At least the AI will take direction. The post concludes:
“We are happy to announce that our new acoustic models are now used for voice searches and commands in the Google app (on Android and iOS), and for dictation on Android devices. In addition to requiring much lower computational resources, the new models are more accurate, robust to noise, and faster to respond to voice search queries – so give it a try, and happy (voice) searching!”
We always knew natural-language communication with machines would present huge challenges, ones many said could never be overcome. It seems such naysayers were mistaken.
Cynthia Murrell, November 5, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Facebook Program May Disintermediate Google
June 5, 2015
Soon, Facebook users may not have to navigate to Google for relevant links then copy-and-paste them into posts and comments. TechCrunch reports, “Skip Googling with Facebook’s New ‘Add a Link’ Mobile Status Search Engine.” If this program currently being tested on a sample group makes it to all users, you can impress your “friends” a few seconds faster, and with fewer clicks. Actually reading what you find before you share the link is up to you. The article describes:
“Alongside buttons to add photos or locations, some iOS users are seeing a new ‘Add A Link’ option. Just punch in a query, and Facebook will show a list of matching links you might want to share, allow you to preview what’s on those sites, and let you tap one to add it to your status with a caption or share statement. Results seem to be sorted by what users are most likely to share, highlighting recently published sites that have been posted by lots of people. …
“If rolled out to all users, it would let them avoid Googling or digging through Facebook’s News Feed to find a link to share. The ‘Add A Link’ button could get users sharing more news and other publisher-made content. Not only does that fill the News Feed with posts that Facebook can put ads next to. It also gives it structured data about what kind of news and publishers you care about, as well as the interests of your friends depending on if they click or Like your story.”
Writers Josh Constine and Kyle Russell observe that, as of last year, Facebook drives nearly 25 percent of “social” clicks, and publishers are becoming dependent on those clicks. Facebook stands to benefit if their Add A Link button enhances that dependency. Then there is the boost to ad revenue the site is likely to realize by keeping users inside their Facebook sessions, instead of wandering into the rest of the Web. A move that will both please users and the bottom line– well played, Facebook.
Cynthia Murrell, June 5, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
The Unacceptable Planned Obsolescence of the Android
June 5, 2015
The op ed on Tom’s Hardware titled Google Can’t Ignore The Android Update Problem Any Longer inspects the release process for Androids, particularly the Android 5.0 Lollipop and the 5.1 iteration. The problem Google faces with its major upgrade per year schedule is that while the Lollipop garners 9.7 percent of the market, it might be several years before the majority of android users catch up to this version, by which time Google might be releasing Android 8.0 (Snickers? M&M?) The article explains the issues with transitioning,
“Because Android is open source and because so many (essentially) OEM-tweaked “forks” of it exist, a “clean” upgrade path is almost impossible. To have a clean standardized update system would mean all the OEMs would have to agree to abide strictly by Google’s guidelines for what they can and cannot modify on the platform.
However, as soon as Google tries to do something like that, the OEMs usually cry foul that Google is making Android more proprietary.”
Obviously Google does not want to lose the business of those OEMs, either. But the article argues that this is unlikely due to Android and iOS cornering the market. The final point is the weakness in the update system due to users desiring more secure platforms, meaning Android adoption will only lessen.
Chelsea Kerwin, June 5, 2014
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
-
- Subscribe to Beyond Search
Feature archive
News archive
-
