IBM Watson Deep Learning: A Great Leap Forward

August 16, 2017

I read in the IBM marketing publication Fortune Magazine. Oh, sorry, I meant the independent real business news outfit Fortune, the following article: “IBM Claims Big Breakthrough in Deep Learning.” (I know the write up is objective because the headline includes the word “claims.”)

The main point is that the IBM Watson super game winning thing can now do certain computational tasks more quickly is mildly interesting. I noticed that one of our local tire discounters has a sale on a brand called Primewell. That struck me as more interesting than this IBM claim.

First, what’s the great leap forward the article touts? I highlighted this passage:

IBM says it has come up with software that can divvy those tasks among 64 servers running up to 256 processors total, and still reap huge benefits in speed. The company is making that technology available to customers using IBM Power System servers and to other techies who want to test it.

How many IBM Power 8 servers does it take to speed up Watson’s indexing? I learned:

IBM used 64 of its own Power 8 servers—each of which links both general-purpose Intel microprocessors with Nvidia graphical processors with a fast NVLink interconnection to facilitate fast data flow between the two types of chips

A couple of questions:

  1. How much does it cost to outfit 64 IBM Power 8 servers to perform this magic?
  2. How many Nvidia GPUs are needed?
  3. How many Intel CPUs are needed?
  4. How much RAM is required in each server?
  5. How much time does it require to configure, tune, and deploy the set up referenced in the article?

My hunch is that this set up is slightly more costly than buying a Chrome book or signing on for some Amazon cloud computing cycles. These questions, not surprisingly, are not of interest to the “real” business magazine Fortune. That’s okay. I understand that one can get only so much information from a news release, a PowerPoint deck, or a lunch? No problem.

The other thought that crossed my mind as I read the story, “Does Fortune think that IBM is the only outfit using GPUs to speed up certain types of content processing?” Ah, well, IBM is probably so sophisticated that it is working on engineering problems that other companies cannot conceive let alone tackle.

Now the second point: Content processing to generate a Watson index is a bottleneck. However, the processing is what I call a downstream bottleneck. The really big hurdle for IBM Watson is the manual work required to set up the rules which the Watson system has to follow. Compared to the data crunching, training and rule making are the giant black holes of time and complexity. Fancy Dan servers don’t get to strut their stuff until the days, weeks, months, and years of setting up the rules is completed, tuned, and updated.

Fortune Magazine obviously considers this bottleneck of zero interest. My hunch is that IBM did not explain this characteristic of IBM Watson or the Achilles’ heel of figuring out the rules. Who wants to sit in a room with subject matter experts and three or four IBM engineers talking about what’s important, what questions are asked, and what data are required.

AskJeeves demonstrated decades ago that human crafted rules are Black Diamond ski runs. IBM Watson’s approach is interesting. But what’s fascinating is the uncritical acceptance of IBM’s assertions and the lack of interest in tackling substantive questions. Maybe lunch was cut short?

Stephen E Arnold, August 16, 2017

Decoding IBM Watson

August 14, 2017

IBM Watson is one of the leading programs in natural language processing. However, apart from understanding human interactions, Watson can do much more.

TechRepublic in an article titled IBM Watson: The Smart Person’s Guide says:

IBM Watson’s cognitive and analytical capabilities enable it to respond to human speech, process vast stores of data, and return answers to questions that companies could never solve before.

Named after founding father of IBM, Thomas Watson, the program is already part of several organizations. Multi-million dollar setup fee, however, is a stumbling block for most companies who want to utilize the potential of Watson.

Watson though operates in seven different verticals, it also been customized for specialties like cyber security. After impacting IT and related industries, Watson slowly is making inroads into industries like legal, customer service and human resources, which comfortably can be said are on the verge of disruption.

Vishal Ingole, August 14, 2017

IBM Watson: Horning in on the WKS Model

August 7, 2017

If you paid attention in sociology class, you might have heard about the Willem Kleine Schaars Model or WKS for short. This is a reasonably well known way to make sense of individuals with psychological disabilities. I was interested when I read an email from IBM to me stating:

Tired of wasting time creating and fine tuning your custom machine learning models? Rules-based approaches can often shorten development time and improve accuracy. Register for our webinar, Accelerate WKS model development with a rule-based approach, on Wednesday, August 9 at 10AM PT to learn how to build rule-based models and save time by using them to pre-annotate machine learning models.

There is nothing so disappointing as the idea of a better way to to perform WKS classifications.

Bummer.

IBM’s WKS means the Watson Knowledge Studio. There’s no getting around the fact that humans or wizards like my goslings have to write rules and plug them into Watson.

My suggestion is:

Either get a better acronym or automate Watson’s expensive, tedious, error prone, difficult, and time consuming preparatory steps for a Watson deployment.

Otherwise I might slip into a different Willem Kleine Schaars’s category. Yikes!

Stephen E Arnold, August 7, 2017

Watson Powers New Translation Earpiece, No Connection Required

August 4, 2017

A start-up out of Australia is leveraging the prowess of IBM’s Watson AI to bring us a wearable translator, dubbed the Translate One2One, that does not require connectivity to function, we learn from “Lingmo Language Translator Earpiece Powered by IBM Watson” at New Atlas. Writer Rich Haridy notes that last year, Waverly Labs found success with its Pilot earpiece. That device was impressive with its near real time translation, but it did depend on a Bluetooth connection. Haridy asserts that New Atlas’ device is the first of its connection-independent kind; he writes:

Lingmo is poised to jump to the head of the class with a system that incorporates proprietary translation algorithms and IBM’s Watson Natural Language Understanding and Language Translator APIs to deal with difficult aspects of language, such as local slang and dialects, without the need for Bluetooth or Wi-Fi connectivity. …

The system currently supports eight languages: Mandarin Chinese, Japanese, French, Italian, German, Brazilian Portuguese, English and Spanish. The in-built microphone picks up spoken phrases, which are translated to a second language within three to five seconds. An app version for iOS is also available that includes speech-to-text and text-to-speech capabilities for a greater number of languages.

The device is expected to be available in July and can be pre-ordered now. A single unit is $179, while a two-piece pack goes for $229. Lingmo launched its first translation device in 2013 and has been refining its tech ever since. Who will be next in the field to go connection-free?

Cynthia Murrell, August 4, 2017

Lost in Translation?

August 3, 2017

Real-time translation is a reality with a host of apps. However, all these apps rely on real-time Cloud Computing for proverbial accuracy. Lingmo One2One Universal Translator seems to be different.

According to a product review published by Forbes and titled Lingmo One2One Universal Translator Preview, the reviewer says:

What gives me pause about the Lingmo, like the other universal translator devices, is the company has no track record in making hardware. Getting the translation stuff right is, I’m sure, hard enough. Getting all that to work in a portable device adds a whole other level of complexity.

Attempts have been made earlier to perfect the translation system, but so far no one has succeeded even decently. The problem is the complexity of human interactions. Though the device is powered by IBM’s AI program Watson, how it manages to store and process the humongous amount of text or voice based communication within the small box is not understandable.

Scientists have been trying to crack the natural language processing problem for a couple of years. Even with the vast amount of resources, it still looks like a distant possibility.

Vishal Ingole, August 3, 2017

IBM Watson: Making AI Wizards Do Eye Ball Rolls

August 1, 2017

Three quotes from Wired’s analysis of IBM Watson in “Watson Won Jeopardy, but Is It Smart Enough to Spin Big Clue’s AI into Green?” My answer is, “No.” I circled these items.

  1. Reference to the IBM Watson ad in which Bob Dylan allegedly says, “I have never known love.”
  2. Quote from a wizard at the Allen Institute for AI, “The only intelligent thing about Watson is their [sic] PR department.” The Beyond Search grammatical goose notes that the phrase might be crafted by a person in the Harrod’s Creek general store as “The only thing about Watson is its PR department.” But the quip is an inspirational one.
  3. The MD Anderson Center pulled out of its $62 million dollar deal for the IBM Watson cancer treating marvel.

None of this info is new, but I found the assessment on point with some of my thoughts.

However, the major flaw in Watson is the time, cost, and Ask Jeeves-type approach to making Watson smart. There are better methods, but good old IBM is apparently not interested. There’s nothing like a big company which allows spokespeople to sign their blog posts with their first name, not a full identifier. What? You don’t know Vijay? Neither do we.

Stephen E Arnold, August 1, 2017

Doctors Fearful of Technology? Too Bad for Them and Maybe the Patients?

July 27, 2017

IBM, Google, and other outfits want doctors to get with the technology program. Sure, docs use mobile phones, but email and such wonderful innovations as selfies from the operating theatre have not yet caught on. Watching my doc fumble with the required online medical record system is interesting. Try it sometime. Puzzled expressions, eye squinting, and sloooow keyboarding are part of the show. One of my docs expressed interest in my Dark Web Notebook. I sent him a link so he could download a comp copy. Guess what? He couldn’t figure out how to download the book. Amazing expertise.

I read a Thomson Reuters’ article which seems to stray dangerously close to my view of technology in the medical profession. Mind you, here in Louisville sales people are in the operating room to provide information to a doc who may not be familiar with a new gadget. Get enough gadgets and peddlers in the facility and the patients may have to rest on gurneys in the hall.

But I digress. The write up i noticed was “Doctors View Technology as Largely Problematic.” I highlighted this “real” news statement:

69 percent of the 100 doctors in the audience said increased reliance on technology and electronic health records only served to separate them from their patients….But the biggest problem stemming from technology for the doctors, and the bane of many doctors’ existence, is the electronic health record, also known as an EHR.

Now think about the over the top marketing from IBM about Watson’s ability in a narrow field like bladder cancer. Put that Anderson affair out of your main. Google continues to push forward with an even more interesting approach. I recall the phrase was “solving death.” And there are other outfits which believe that their technologists can make life so much better for doctors.

Seems like the revolution may take a bit more time. The good news is that since Google has not solved death, the doubting docs will die. Their replacements may be more into the IBM, Google, et al approach to health care.

No worries in Harrod’s Creek. We just use a mixture of black powder and bourbon to cure all manner of ills.

Stephen E Arnold, July 27, 2017

Banks Learn Sentiment Analysis Equals Money

July 26, 2017

The International Business Times reported on the Unicorn conference “AI, Machine Learning and Sentiment Analysis Applied To Finance” that discussed how sentiment analysis and other data are changing the financing industry in the article: “AI And Machine Learning On Social Media Data Is Giving Hedge Funds A Competitive Edge.”  The article discusses the new approach to understanding social media and other Internet data.

The old and popular method of extracting data relies on a “bag of words” approach.  Basically, this means that an algorithm matches up a word with its intended meaning in a lexicon.  However, machine learning and artificial intelligence are adding more brains to the data extraction.  AI and machine learning algorithms are actually able to understand the context of the data.

An example of this in action could be the sentence: “IBM surpasses Microsoft”. A simple bag of words approach would give IBM and Microsoft the same sentiment score. DePalma’s news analytics engine recognises “IBM” is the subject, “Microsoft” is the object and “surpasses” as the verb and the positive/negative relationships between subject and the object, which the sentiment scores reflect: IBM positive, Microsoft, negative.

This technology is used for sentiment analytics to understand how consumers feel about brands.  In turn, that data can determine a brand’s worth and even volatility of stocks.  This translates to that sentiment analytics will shape financial leanings in the future and it is an industry to invest in

Whitney Grace, July 26, 2017

Watson Does Whiteboards

July 24, 2017

A write-up at Helge Scherlund’s eLearning News describes a very useful tool in, “World’s Smartest Active Virtual Meeting Assistant Ricoh.” The tool integrates the IBM Watson AI into an interactive whiteboard system. The press release positions the tool as the future of meetings, but we wonder whether small businesses and schools can afford these gizmos. The write-up includes a nine-minute promotional video that describes the system, so interested readers should check it out. We’re also given a list of key features.

*Easy-to-join meetings: With the swipe of a badge the Intelligent Workplace Solution can log attendance and track key agenda items to ensure all key topics are discussed.

 

*Simple, global voice control of meetings: Once a meeting begins, any employee, whether in-person or located remotely in another country, can easily control what’s on the screen, including advancing slides, all through simple voice commands.

 

*Ability to capture side discussions: During a meeting, team members can also hold side conversations that are displayed on the same Ricoh interactive whiteboard.

 

*Translation of the meeting into another language: The Cognitive Whiteboard can translate speakers’ words into several other languages and display them on screen or in transcript.

I suppose one feature here may also be a thorn in the side of some old-school business people—the system creates a transcript of everything said in each meeting, including side conversations, and sends it to each participant. Auto CYA. The process would take some getting used to, but we can see the advantages for many organizations. Headquartered in Tokyo, Ricoh’s history stretches back to 1936.

Cynthia Murrell, July 24, 2017

IBM Watson: Two Views of the Same Pile of Tinker Toys

July 19, 2017

I find IBM an interesting outfit to watch. But more entertaining is watching how the Watson product and service is perceived by smart people. On the side of the doubters is a Wharton grad, James Kisner, who analyzes for a living at Jeffries & Co. His report “Creating Shareholder Value with AI? Not So Elementary, My Dear Watson?” suggests that IBM is struggling to makes its big bet pay off. If not a Google moon shot, Mr. Kisner thinks the low orbit satellite launch is in an orbit which will result in Watson burring up upon re-entry to reality.

Image result for chihuahua costume

The Big Dog of artificial intelligence and smart software may be a Chihuahua dressed up like a turkey, not a very big dog, not much of a bark, and certainly not equipped to take a big bite out a Wharton trained analyst’s foot.

On the rah rah side is Vijay, a blogger who does not put his name on his blog or on his About page. (One of my intrepid researchers thinks this Vijay’s last name is “Vijayasankar?.” Maybe?) I assume he is famous, just not here in Harrod’s Creek. His most recent write up about Watson is “IBM Watson Is Just Fine, Thank You!” His motivation for the write up is that the attention given to the Jeffries’ report caught his attention. He is a straight shooter; for example:

I am a big fan of criticism of technology – and as folks who have known me over time can vouch, I seldom hold back what is in my mind on any topic. I strongly believe that criticism is healthy for all of us – including businesses, and without it we cannot grow. If you go through my previous blogs, you can see first hand how I throw cold water on hype.

I like the cold water on hype from a person who is an IBM executive, and one who has been involved in the IBM Watson health initiatives. (I think this includes the star crossed Anderson project in Houston. I hear, “Houston, we have a problem,” but you may not.) I highlighted these points in this blog post:

  1. Hey, world, IBM is an enterprise product, not a consumer product. This seems obvious, but apparently IBM’s ability to communicate what it is selling and to whom is not working at peak efficiency or maybe not working because everyone is confused about Watson?
  2. IBM does not do the data federation things with its customer data. That’s good. I know that IBM sells a mainframe that encrypts everything. Interesting but I am not sure how this addresses flat revenue growth, massive layoffs, and the baffling Watson marketing which recently had a white cube floating in a tax preparer’s office. A white cube?
  3. IBM Watson has lots of successes. That’s a great assertion. The problem is that Watson started out as the next big thing. There was a promise of billions in revenue. There was a big office commitment in Manhattan. Then there was the implosion at the Houston health center. “Watson, do you read me?” I once tracked some of the Watson craziness in a series called the “Weakly Watson.” I gave up. The actual examples struck me as a painful type of fake news. What’s interesting is that the “weakly” stories were “real.” Scary to me and to stakeholders.
  4. Watson is not a product. Watson is an API to the IBM ecosystem. Vendor lock in beckons. And, of course, lots of APIs. These digital tinker toys can be snapped together. The problems range from the cost and time required for system training, the consulting and engineering services price tag, and the massaging required to explain that Watson is something that requires a lot of work. For the Instagram crowd that’s a problem. “Houston. Houston. Do you copy? Tinker toys. Lego blocks. Do you copy?”
  5. Watson “some times needs consulting.” Talk about an understatement. Watson needs lots and lots of consulting, engineering services, training, configuring, tuning. and training. Because Watson is a confection of open source, acquired technologies, and home brew code—a lot of work is needed. That’s because Watson was designed to generate high margin services, not the trivial revenue from online ads or from people ordering laundry detergent by pressing a button on their washing machine.
  6. Watson has two things in its bag of tricks: “Great marketing” and “AI talent.” Okay, marketing and smart people. The basic problem IBM has to solve before investors get frisky is generating significant, sustainable revenues and healthy margins. Spending money buys marketing and people. Effective management orchestrates what can be bought into stuff that can be sold at a profit.

The Vijay write up ends with a question. Here you go: “So why is IBM not publishing Watson revenue specifically?” This Vijay fellow who assumes that I know his last name does not answer the question. In the deafening silence, we need an answer.

That brings me to the Jeffries & Co. report by James Kisner, who is certified to do financial analysis. The answer to Vijay’s question consumes 53 pages of verbiage, charts, and tables of numbers. The entire document was available on July 18, 2017, at this link, but it may disappear. Many analyst documents disappear for the average guy. (If the link is dead, head over to Investext or give Jeffries & Co. a quick call to see if that will get you the meaty document.

Image result for snarling guard dog

A Jeffries & Co. analyst with teeth bites into the IBM financial data and seems to be unsatisfied.

In a nutshell, the Jeffries’ report says that IBM Watson is a limp noodle. Among the Watson characteristics are unhappy customers, wild and crazy marketing, misfires on deep learning, and the incredibly difficult, time consuming, and expensive data preparation required to make the system say, “Woof, woof” or maybe “Wolf, wolf” when there is something important for a human to notice.

Net net: IBM’s explanations of Watson have not produced the revenues and profits stakeholders expect. Jeffries & Co. goes MBA crazy providing a wide range of data to support the argument that Watson is struggling.

That “woof, woof” is the sound of a Chihuahua barking with the help of IBM spokespeople and lots of PR and marketing minions. The Wharton guy is a larger dog, barks ferociously, and has a bite backed up by data. IBM has to prove that it can solve problems for clients, generate sustainable revenue, and keep the competition from chowing down on a Watson weighted down with digital tinker toys.

Stephen E Arnold, July 19, 2017

Next Page »

  • Archives

  • Recent Posts

  • Meta