IBM Cognitive Computing versus Papyrus UTA

I am pleased to see that I am not the only one to believe in using human intelligence as a model for building better software and applications as we did with the User-Trained Agent of the Papyrus Platform.

In its 2006 conference the IBM Almaden Institute focused on the theme of “Cognitive Computing” and examined scientific and technological issues around how the human brain works.  Approaches to understanding cognition that would unify neurological, biological, psychological, mathematical, computational, and information-theoretic insights were discussed.

As a major next step, IBM disclosed its Cognitive Computing project a few days ago. “Today’s computers are unable to deal with ambiguities or consolidating information from multiple sources into a holistic picture of an event. Without the ability to monitor, analyze, and react to the continuous growth of information in real time, the majority of its value may be lost,” IBM said in the press release.

IBM hopes that such a cognitive machine could integrate data from consumers’ credit reports, tax returns, pay stubs, and mortgage statements in order to allow lending companies to make instant decisions on loan eligibility. IBM and its research partners will look closely at the brain’s internal structure and functions and attempt to emulate those patterns on silicon and transistors. IBM’s partners include Stanford University, University of Wisconsin-Madison, Columbia University Medical Center, Cornell, and University of California at Merced.

My take: THAT IS A FAIRLY LONG SHOT. Neural net computing has been around for some time and little practical value has been achieved except in very special applications. IBM and its partners are fairly naive in the goals for this project. I don’t believe that a hardware approach will be the right one because it is too expensive and too slow to iterate in development. How such a system will model and understand data will be interesting to see. Further, it will be a hard sell once the solution should actually work.

Our Papyrus User-Trained Agent is a simple cognitive REAL-TIME functionality that models basic human decision-making on trained patterns. Our model can still be expanded substantially in depth and capability. I have been telling our customers about the UTA for two years now and they simply are not that interested. If they are we have one major problem: Cognitive computing that learns from large scale data patterns does not follow easy to understand logic. All the people that are interested in the UTA ask: “What has the UTA learned?” and “Why does it take one decision rather than another?” Also our study work with Vienna University brought up the same questions.

Therefore our main work item for the UTA at this time is to visualize the knowledge that it acquired. That is only possible because it works as a piece of software and has access to the data models and real-time data while it processes. It can use that information to write historic process information that can be used to create process decision graphs that a human can understand.

So IBM’s cognitive computing projects will face the same hurdles. Should such a cognitive HW device be commercially feasable at some stage (ten years down the road?), people will not want to use it because they would not understand its decisions and reasoning. People don’t trust business intelligence data right now, so why should they trust business intelligence that was analyzed by a computer whose thinking they don’t understand.

What is the solution? As always, it is: ‘Keep it simple.’ Cognitive computing has to happen on a small scale and on limited data that are plausible like we do with the UTA. Rather than huge computing devices that understand huge amounts of data, the right approach is to narrow the data volume down by ensuring that the context of information is the right one. Massive context-less data do not solve anything.

I have learned from human intelligence, that we are only so fast in our thinking because we always have limited processing to perform. We simply map information and knowledge we EXPECT because of the context, and compare it against the real-time input. Therefore the context of information reduces the data amount to be processed and simple pattern matching is enough to recognize information and to make decision based on past patterns trained. In the UTA the context is currently defined by the SCOPE of the state space that is monitored. Our current research tries to find better ways to identify context more automatically and efficiently. So in effect we are quite a few years ahead of IBM and its research partners.

7 Comments on “IBM Cognitive Computing versus Papyrus UTA

  1. I am not so certain that not understanding how something works will make a difference to anyone if it works.

    So setting up a UTA, training it, and then letting it perform its job will be enough to convince anyone. I guess the trick is to convince someone to give it a serious try.

    Barely anyone who uses it really understands (remembers) how electricity works, and everyone who can does use it in many different ways.

    Ten years from now the perhaps realized ubiquity of holistic computing will preclude anyone from trying to understand how it works.

    Like

  2. “Therefore the context of information reduces the data amount to be processed”.

    We human can quickly decide which “context” to use, but a machine can’t.

    UTA is just one small piece of the whole picture. The whole picture (where a lot of “contexts” exist) needs to be drawn from a large-scale data, then we really need large-scale machine learning or large-scale cognitive systems.

    So don’t criticize what you don’t understand, Max.

    Like

    • Kent, thanks for the reply. We can always agree to disagree, but I think I just failed to make my point in that short paragraph. I understand the (complete nonsense) subject of large-scale data analysis better than most experts, because I look at it from a human perspective. I have not been maths-PhD-brainwashed to believe that nature follows mathematical models. These models are an illusionary assumption that only work in small functional areas.

      I would honestly like to hear from you HOW you can draw a ‘whole picture’ from large-scale data. There aren’t any ways that I know of that can be taken seriously. Question: HOW does a human choose (a.k.a. ‘decide’) which context to use and why would large-scale data analysis do that any better than what a human does without large-scale data? IBM’s goal is to create human-like cognitive computing, right? So they are way off the right track. The main problem of large-scale data is a total lack of understanding correlation and causality and therefore a total lack of RELEVANCE.

      Follow that line of thought: Does a data-set have an implicit meaning? The answer is a simple no. The meaning is bestowed by metadata, context and interpretation. The meaning is bestowed by the receiver and not by the sender! You don’t know if the data has any meaning in the context that you choose. That is a natural physical concept that is still widely misunderstood by most quantum physicist who therefore have to deal with all these paradoxons. Therefore gathering huge amounts of data does not do anything beneficial as it just creates noise. If you read Gigerenzer then it will become quickly clear to you that decisions are always better made with limited data. The human brain works with local pattern recognition clusters that all work on local limited data and never try to figure the meaning of all the sensory inputs together.

      What we can do is pick up those patterns that were used by humans to make those choices and learn from them. This does not require ONE UTA but many and each one can learn a few of those patterns that were used in making those choices. Each one uses a particular scope/context but it will still be an assumption what the context was. Only when those contextual decision patterns are then reviewed and corrected by humans the decision patterns used in the UTA make any sense at all. A human can not do that for large-scale data and therefore the grand notion of business intelligence is utterly useless and an illusion.

      And then we have the final misunderstanding to deal with: Human decisions are always emotional (Minsky, Damasio) and therefore do not follow a logical choice. A machine will therefore not be able to make human-like cognitive choices of context. Large-scale data analysis will not improve that in any way.

      If you have a different opinion I would like to hear IN DETAIL AND WHY, please.

      Like

  3. Hello,
    Sorry could not reply to you earlier. My point was using machine as a low-cost solution to parse large-scale data. They are not yet “human-like”, since we have not yet understood how our brain functions.
    Some successful examples of how low-cost and efficient large-scale machine learning techniques are, are automatic speech recognition and statistical machine translation. (Of course those automatic quality are far worse than a good human hearing or translator).

    I will return to this soon, thanks.

    Like

  4. Hello Kent, no problem at all. We seem to understand pretty well how the brain works biologically today but as it is an emerging phenomenon that can’t be decomposed it will not help us to build like machinery in any case. And the machinery won’t have emotions …
    We use machine learning very successfully at ISIS Papyrus since many years and the results are very good. What I am trying to point out is that the machine learning does not improve with the amount of data and that is true also for humans. Therefore to create human-like cognition it makes no sense to look at large scale data analysis. Humans do very well with just a few data samples in the right context. Cognition on large-scale data sets won’t be human after all …

    Like

  5. That’s funny: quite finished Damasio I found your post. And I can folllow your statement. This is the reason why we need experts in case management situation for hopefully long time. Our brain can handle no more the a maximum of twelve points to make a decision. This is the reason why we say in Germany “Schlaf mal drüber” (sleep a night), if a decision has to be done on lots of point, simple buying a car.
    I am 51 years old and looking to my years as a teenager I see more and more what it means to be wise. Because a I made so many experiences I will decide probably without less errors than a youngster. And the decision is mostly done by emotion.
    Thanks for your post.

    Like

    • Martin, thanks for your comment. Yes, Damasio explained to me the power and importance of emotion. I always thought that I had to be rational but was still unhappy. The other really good reading on this subject is Gerd Gigerenzer ‘Gut Feeling’ and a few others, where he writes about the various decision mechanisms of our brain — the adaptive toolbox. Regards, Max

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 5,331 other followers

%d bloggers like this: