August 23, 2016
After several tests, the fourth HonkinNews video is available on YouTube. You can view the six minute video at https://youtu.be/AIYdu54p2Mg. The HonkinNews highlights a half dozen stories from the previous week’s Beyond Search stream. The commentary adds a tiny twist to most of the stories. We know that search and content processing are not the core interests of the millennials. We don’t expect to attract much of a following from teens or from “real” search experts. Nevertheless, we will continue with the weekly news program because Google has an appetite for videos. We will continue with the backwoods theme and the 16 mm black and white film. We think it adds a high tech look to endless recycling of search and content jargon which fuels information access today.
Kenny Toth, August 23, 2016
August 9, 2016
You can view the August 8, 2016, HonkinNews program at this link. The video comes from Goodwill-grade 8 mm film equipment. The program highlights recent stories from the free (yep, no cost whatsoever) Beyond Search Web log. Learn about the how one Google executive “escaped” life in the fast lane. The Verizon acquisition of Yahoo reminds Stephen of Washington’s wooden false teeth. The deal allows Verizon to own two Internet artifacts. Hewlett Packard Enterprise, owner of Autonomy, faces an uncertain future as its sells units and thinks about selling itself. And there’s more in the six minute news program; for example, a restrained MBA cheer for Big Data. But that’s a sotte voce rah, rah. Like Beyond Search, the honking video version tries to separate the giblets from the goose feathers in the thrilling world of search, content processing, and related disciplines. That’s not easy in today’s search-centric world where relevance is mostly good enough and jargon is its own virtual reality.
Ken Toth, August 9, 2016
June 18, 2016
Years ago I read “The FBI’s Next Generation Identification Program: Helping Law Enforcement Track and Share Suspect Information across State Lines.” That write up identified, probably semi accurately, Lockheed Martin as the “lead contractor” for the NGI IPS (Next Generation Identification Interstate Photo System). I mention this because the write up “FBI Has 411 Million Photos in Its Facial Recognition System, and a Federal Watchdog Isn’t Happy” does not dig into the contractor or contractors involved in this nine zero project. (An older list of some FBI contractors appears at this link.)
The GAO report about the program also lacks some details. If you are interested in what a government report of the controversial system offers, you can download for now a copy of the document at this link.
I realize that the marketing of smart systems which can make sense of images suggests three functions:
- High speed matching
- High precision
- High recall.
The reality is a bit different. Please, keep in mind that the beliefs created by over inflated marketing claims and carefully staged demonstrations often are at odds with how the system actually performs in real life.
Government entities have to look to technology to help deal with the ever increasing and possibly unstoppable flood of digital information. The actual systems, whether the UK’s NHS systems or the US Army’s DCGS systems, are works in progress. In many cases, the progress is halting, and the work has unanticipated consequences.
I have pointed out that enterprise search, content management, and similar and smart software are not the slam dunks many managers think they are. Hope springs eternal, but that hope has to be gated with what happens in the real, disorganized, and time starved reality in which the magic is supposed to happen.
Stephen E Arnold, June 20, 2016
June 7, 2016
I read “Google Voice Search Records and Keeps Conversations People Haver Around their Phones but You Can Delete the Files.” I like the “you can delete the files. How does one know what has or has not been deleted in this era of real time cloud goodness?
I assume that the information in the write up is accurate.
The write up states:
The feature works as a way of letting people search with their voice, and storing those recordings presumably lets Google improve its language recognition tools as well as the results that it gives to people.
If you want to “delete” these recordings, the write up asserts:
It’s found by heading to Google’s history page and looking at the long list of recordings. The company has a specific audio page and another for activity on the web, which will show you everywhere Google has a record of you being on the internet.
Optimism is good. One presidential hopeful believed certain emails had been deleted. I am not sure that the FSB agrees. It seems that the Independent’s “real journalist” was not aware of “Your Data Is Forever.”
Stephen E Arnold, June 7, 2016
April 18, 2016
I read a story about matching up user queries with images. I don’t think Google’s image search is particularly good. Examples range from Google’s obsession with taking a query like “truth” and returning images of pictures with the word “truth” in them. And this image:
What about the query for “watson.” Google showed a picture of a computer, a person named “sherlock,” and images of this guy:
The write up “Do Google’s ‘Unprofessional Hair’ Results Show It Is Racist?” wants to point out that Google’s methods have a nasty side. I noted this passage:
We’ve always conceived of search engines as arcane but neutral creatures, obedient only to our will and to the precious logic of information. Older engines from the advent of the internet reflected this: Remember “Ask Jeeves,” the genteel butler? Dogpile, which would “fetch” things for you? Despite this fantasy, the things engines and their algorithms are able to know and to find are influenced by the content we give them to work with, which means they may reflect our own biases.
AskJeeves was a human powered system. The Google is algorithmic. Google does not “give” its image search system content. The image search system indexes what it finds, within the depth settings for the crawl. Sorry, gentle reader, Google does not index everything available via the Internet. Bummer, right?
I circled this statement:
is its [image search’s] purpose to reflect and reinforce what its users feel, do and believe? Or is it to show us a fuller picture of the world and all things contained in it as they really are? Google Images was conceived in response to what people most wanted to see. Maybe it hasn’t decided yet what we most need to see.
The Guardian itself is an interesting legal search. Run the query “guardian” on Google Images and what does one find? Here you go:
The logo of the “real” journalistic thing and the word “truth.” Now is that biased?
Stephen E Arnold, April 18, 2016
April 15, 2016
I marvel at the baloney I read about smart software. The most effective systems blend humans with sort of smart software. The interaction of the human with the artificial intelligence can speed some work processes. But right now, I am not sure that I want a smart software driven automobile to navigate near the bus on which I am riding. I don’t need smart automobile keys which don’t work when the temperature drops, do you? I am not keen on reading about the wonders of IBM Watson type systems when IBM struggles to generate revenue.
I read “Why Our Crazy-Smart AI Still Sucks at Transcribing Speech.” Frankly I was surprised with the candor about the difficulty software has in figuring out human speech. I highlighted this passage:
“If you have people transcribe conversational speech over the telephone, the error rate is around 4 percent,” says Xuedong Huang, a senior scientist at Microsoft, whose Project Oxford has provided a public API for budding voice recognition entrepreneurs to play with. “If you put all the systems together—IBM and Google and Microsoft and all the best combined—amazingly the error rate will be around 8 percent.” Huang also estimates commercially available systems are probably closer to 12 percent. “This is not as good as humans,” Huang admits, “but it’s the best the speech community can do. It’s about as twice as bad as humans.”
I suggest your read the article. My view is that speech recognition is just one area which requires more time, effort, research, and innovation.
The situation today is that as vendor struggle to prove their relevance and importance to investors, many companies are struggling to generate sustainable revenue. In case anyone has not noticed, Microsoft’s smart system Tay was a source of humor and outrage. IBM Watson spends more on marketing the wonders of its Lucene, acquired technology, and home brew confection than many companies earn in a year.
There are folks who insist that speech to text is not that hard. It may not be hard, but this one tiny niche in the search and content processing sector seems to be lagging. Hyperbole, assurance, and marketing depict one reality. The software often delivers a different one.
Who is the leader? The write up points out:
…most transcription start-ups seem to be mainly licensing Google’s API and going from there.
Yep, the Alphabet Google thing.
Stephen E Arnold, April 15, 2016
April 4, 2016
I love the capitalist tool. The founder rode a motorcycle. When I was in Manhattan, I had the pleasure of listening to the Malcolm-cycle burble and grunt when talking with a couple of pals. Wonderful that noise and odor.
I read “The Content Pyramid: And Why Video Must Be at the Top.” I am not sure the founder of the capitalist tool was into video. Well, the capitalist tool is an an article with a parental “must” makes this point:
Video is the Matryoshka doll of content.
I did not know that. I know that some folks who shoot videos write scripts, sell them and then other people (who know better than the author) rewrite them.
The write up points out that a video has a script. But the video has pictures and audio.
I need to take a couple of deep breaths. My heart is racing with the impact of these comments.
As more and more content consumption goes mobile, it’s usually a necessity to create multiple lengths and optimized formats of video content, so you should always have tiered, multi-channel thinking built in to your editorial process.
So how much video does the capitalist tool have on YouTube? 4,900 videos. But that’s not too many. I ran the query “Forbes” on Google Video and learned that there are 16,900 videos available. I checked Vimeo and learned there were 521 videos. I checked Blinkx and found quite a few false drops.
The problem is that I have never seen a reference to a Forbes video. I do receive mail addressed to my deceased father enjoining him to re-subscribe to the print edition of Forbes Magazine. But the video thing with the podcast, the clips, and the use of video in marketing. Not on my radar.
Remember the “must.” How about adding the concept of “effective”?
Video by itself is a bit of an ego play in my opinion. When no one watches the video or knows a video exists, what’s the point? Right, right. I forget. Some ad agencies love to do video shoots in Half Moon Bay. It is fun. How bright the video shines depends on more the height of the pyramid in my opinion.
Stephen E Arnold, April 4, 2016
March 28, 2016
I survived the Go games. In case you have been on an extended vacation, Google’s smart software beat a human at the game of Go. I assume that this smart software did not drive the car which ran into a bus, but that’s another issue.
I then noted “IBM Watson Could Soon Use Artificial Intelligence to Beat You at a Game of I Spy.” I love the use of the word “could.” I prefer supposition to reality. Contrast the satisfaction of “I could go to the gym” with “I am eating potato chips.” Which does IBM prefer? If you answered, “Generate substantial revenue”, you are incorrect.
The write up in question reports that IBM has “updated” Watson. I noted this statement about the updated Watson:
IBM has created a ‘Visual Recognition Demo’ to showcase Watson’s latest trick, which allows users to feed Watson an image before it tells you what it believes it sees. For example, supplying Watson with the image of a tiger throws up the result 77 per cent tiger, 26 per cent wild cat and 63 per cent cat.
In my experience, determining if an animal is a real live and possibly hungry tiger, that error could be darned interesting. On my last trip to Africa, I learned that a hapless trekker discovered that confusing “cat” with “tiger” can have interesting consequences.,
Sigh. IBM appears to be making news out of some image processing capabilities which I have seen in action before. How long “before”? Think more years than IBM has been reporting declining revenues. Watson, what can one do about that? Hello, Watson. Are you there?
Stephen E Arnold, March 28, 2016
March 23, 2016
I read “DeepGram Lets You Search through Lectures and Podcasts for Your Favorite Quotes.” I don’t think the system is available at this time. The article states:
Search engines make it easy to look through text files for specific words, but finding phrases and keywords in audio and video recordings could be a hassle. Fortunately, California-based startup DeepGram is working on a tool that will make this process simpler.
The hint is the “is working.” Not surprisingly, the system is infused with artificial intelligence. The process is to covert speech to text and then index the result.
Exalead had an interesting system seven or eight years ago. I am not sure what happened to that demonstration. My recollection is that the challenge is to have sufficient processing power to handle the volume of audio and video content available for indexing.
When an outfit like Google is not able to pull off a comprehensive search system for its audio and video content, my hunch is that the task for a robust volume of content might be a challenge.
But if there is sufficient money, engineering talent, and processing power, perhaps I will no longer have to watch serial videos and listen to lousy audio to figure out what some folks are trying to communicate in their presentations.
Stephen E Arnold, March 23, 2016
February 3, 2016
I read “Knowledge Discovery using Various Multimedia Data Mining Technique.” The write up is an Encyclopedia Britannica type summary of the components required to make sense of audio and video.
I noted this passage:
In this paper, we addressed data mining for multimedia data such as text, image, video and audio. In particular, we have reviewed and analyzed the multimedia data mining process with different tasks. This paper also described the clustering models using video for multimedia mining.
The methods used by the systems the author considered use the same numerical recipes which most search vendors know, love, rely upon, and ignore the known biases of the methods: Regression, time series, etc.
My take away is that talk about making sense of the flood of rich media is a heck of a lot easier than processing the video uploaded to Facebook and YouTube in a single hour.
The write up does not mention companies working in this farm yard. There are some nifty case studies to reference as well; for example, Exalead’s video search and my touchstone, Google YouTube and Google Video Search. Blinkx (spun out of Autonomy, a semi famous search outfit) is a juicy tale as well.
In short, if you want to locate videos, one has to use multiple tools, ask people where a video may be found, or code your own solution.
Stephen E Arnold, February 3, 2016