IBM Watson: AI For The Real World - InformationWeek

InformationWeek is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

IoT
IoT
Comments
IBM Watson: AI For The Real World
Threaded  |  Newest First  |  Oldest First
Ashu001
50%
50%
Ashu001,
User Rank: Ninja
12/26/2015 | 4:47:14 AM
Re: IBM Watson
Ilya,

Great points raised by you here.

As far my reading of all the "Marketing Literature" from IBM Watson goes it seems that the trick which they are currently pushing is for improved Machine learning/analysis of the entire Data an enterprise can house.

Do you feel this is'nt good enough on its own?

Great Blog otherwise.

Best of Luck with all your Court cases.I strongly believe if you pushing in the right direction;eventually you will prevail.

 
Li Tan
50%
50%
Li Tan,
User Rank: Ninja
1/2/2016 | 1:51:44 AM
Re: IBM Watson
I am eager to see the actual and wider utilization of IBM Watson in industry - can it make us work smarter and more efficient?
yalanand
50%
50%
yalanand,
User Rank: Ninja
12/26/2015 | 12:28:03 PM
Re: IBM Watson
@Ilya: That is why it's called Structured Query Language. I hear Hashing on indexed SQL tables are lightning fast.
Brian.Dean
50%
50%
Brian.Dean,
User Rank: Ninja
12/26/2015 | 8:26:15 PM
Re: IBM Watson
Good point about knowledge. I feel that machine learning's first task would be to help businesses allocate resources in an efficient fashion. For instance, if user sentiment is negative for a certain feature of a product that is driving sales figures to be low and similarly, user sentiment is positive for a feature, causing sales figures to increase then, identifying the variables at play would deliver value to the business. The variables could be locally correct (due to culture) but, universally incorrect. Nevertheless, value would still be generated for the business.
Ashu001
100%
0%
Ashu001,
User Rank: Ninja
12/27/2015 | 11:07:49 PM
Re: IBM Watson
Brian,

Fair points made regarding Localized Machine Learning experience and the ability of Managers to learn from that Experience.

However,the problem routinely arises when someone tries to apply those lessons learnt in other Jurisdictions where the same Local Features/Values don't arise.

Just for example's sake-Ford sees their Best-Selling Vehicles in the US are currently Gas-Guzzling Trucks and tries to superimpose the same product in Europe or India as anyone who has experience of all these markets will tell you it won't be successful ;primarily because Fuel Taxes are much-much higher in Europe and India than in America.

In both these Markets Taxes on Gasoline and Diesel come to over 60% of the Retail Cost of the Fuel sold to End-users which are much-much higher than in most parts of America today(not excluding High-Tax Jurisdictions like California,Hawaii and New York City).

When a Consumer has to pay a Dollar for a Litre of Gasoline their behavior is very different than when they have to pay 50-60 Cents /Litre.

This kind of customization and fine-tuning needs to fed into the Machine Learning first before it takes off in a big-big way.

Just my two cents on this fascinating topic.

 
PedroGonzales
50%
50%
PedroGonzales,
User Rank: Ninja
12/28/2015 | 5:18:56 PM
Re: IBM Watson
@Dean. All organization which collects huge amounts of data would really benefit from Watson. Imagine NASA being able to manage rocket launching by allowing watson to make sense of the millions of sensor it contains during rocket launch.  I hope that somehow Watson can be use in the federal government.  if watson can reduce the number of budget gridlocks the government, we should Watson manage congress.
SachinEE
50%
50%
SachinEE,
User Rank: Ninja
12/28/2015 | 10:22:34 PM
Re: IBM Watson
DeepMind seems to be developing a very powerful AI as well. 2016 is going to be the year of AI I think.
SachinEE
50%
50%
SachinEE,
User Rank: Ninja
12/28/2015 | 10:23:50 PM
Re: IBM Watson
I think someone should integrate nanotechnology with an AI and see where it takes them. The work on Nanotechnology has been stagnant for months.
Broadway0474
50%
50%
Broadway0474,
User Rank: Ninja
12/29/2015 | 11:38:28 PM
Re: IBM Watson
Unfortunately, Pedro, Watson cannot learn how to end political gridlock in Washington. I think if we begin to train it to do that, then we'll push him closer to actually being HAL. (Speaking of which, I don't completely buy the claim that Watson's nothing like HAL ... can't he just be taught to be crazy and murderous? "Watson, what's the best way to do away with Joe Bloe with extreme prejudice?")
Stratustician
50%
50%
Stratustician,
User Rank: Ninja
12/30/2015 | 8:21:02 PM
Re: IBM Watson
Great suggestions.  Given the right type of sensor data, Watson could be a great help in emergency situations as it would be able to better understand the data without human emotion getting in the way and delaying making a decision.  It could also look at larger data that might not normally be associated with a particular situation or project for correlations that might be overlooked by others.  Humans tend to have biases built on experience, if we can tap Watson to help us move past these biases, we could see improvement in many industries.
Ashu001
50%
50%
Ashu001,
User Rank: Ninja
12/31/2015 | 11:49:43 AM
Re: IBM Watson
Stratustican,

While for not even one instance do I dispute your assertion that Human Beings have biases ;I would tend to argue that those Biases tend to stop us form making the wrong decisions most of the time(collectively).

Think about it carefully how many times has the world come close to Nuclear Annihilation,etc during the Cold War and also recently-What is the Number one reason;we stepped back from the breach each and everytime??

Its these very built-in biases where we fear from our Kids Futures.

Unless we are able to structure AI with such built-in biases ;I wouldnt trust AI With nothing.

 

 
Stratustician
50%
50%
Stratustician,
User Rank: Ninja
12/31/2015 | 12:50:52 PM
Re: IBM Watson
I agree 100%, there are certain instances where AI should be kept separate, and yes, anything that has the potential to cause damage, such as in the case of weapons systems, should be kept separate.  As folks like Elon Musk point out, we cannot give AI systems access to anything where they can override human safeguards. 

AI is great for data, not for running critical infrastructure or weapons systems.
Ashu001
100%
0%
Ashu001,
User Rank: Ninja
12/27/2015 | 11:11:22 PM
Re: IBM Watson
Ilya,

Really Fascinating points shared here by you.

I must say I always learn a lot in this field from your discussions here.

The very fact that Yes/No can and should only be issues manually is super-critical atleast in the current state of AI/Machine-Learning.

Maybe in the future when this Technology develops and grows further we will see the Machines doing the Yes/No selections themselves successfully.

But that stage is still some way away today.

 
TerryB
100%
0%
TerryB,
User Rank: Ninja
12/29/2015 | 1:03:27 PM
Re: IBM Watson
@Ilya, you're such a brilliant guy, I must admit I've having a little trouble keeping up with you on your thoughts of Watson versus Google's search (for example).

Isn't a computer always going to have trouble with unstructured data? Leaving metadata tags out of it, how would a computer know that Jules Verne's 20,000 Leagues is fiction and War and Peace is not? At a very high level, I guess I get that non fictional history still struggles to differentiate pure fact from the opinion/supposition injected by the author. But wouldn't AI struggle to even know what is real and what is not when just parsing these books as unstructered data? I'm not sure most humans can tell difference sometimes.

Isn't that at the heart of why IBM designed Watson the way it did? While I agree Watson is struggling to find commercial success right now, I was extremely impressed at what it did on Jeopardy. It certainly defeated the best human Jeopardy players we have?
nasimson
100%
0%
nasimson,
User Rank: Ninja
12/30/2015 | 7:56:04 PM
Watson - open for public
Why doesn't IBM opens up the access to Watson to anybody over the Internet? It's really saddening to see that IBM has to justify that Watson will never take decisions on its own or will never answer questions unasked. Why AI has to be apologetic?
Broadway0474
50%
50%
Broadway0474,
User Rank: Ninja
12/30/2015 | 9:40:20 PM
Re: Watson - open for public
nasimson, please tell me you're kidding. IBM has to say that sort of disclaimer to fend off the anti-AI crowd, who has a very good reason to fear the singularity. People pushing for the singularity to see what will happen are similar to religious extremists pushing for war in the Middle East to speed Armageddon along.
Ashu001
50%
50%
Ashu001,
User Rank: Ninja
12/31/2015 | 11:53:24 AM
Re: Watson - open for public
Nasimson,

The reason for that is not very hard to take sight of.

Its primarily because IBM does'nt want anyone on the Internet "exposing " the various flaws in their Watson system in its entirety.

LOL!

 
Ashu001
50%
50%
Ashu001,
User Rank: Ninja
12/31/2015 | 11:42:26 AM
Re: Watson - open for public
Ilya,

I have to agree wholeheartedly with your point of view here.

We are far-far away from a Perfect AI machine today.

I dont think we will be even Half-way there by 2020.

Sure,we will proceed closer and closer by leaps and bounds with every passing year but to say that we will achieve full perfection in this space very soon is to say too much.

Not sure,how much help Stem Cells or even Basic Neurons can achieve here.

 
Ashu001
50%
50%
Ashu001,
User Rank: Ninja
12/31/2015 | 11:58:34 AM
Re: Watson - open for public
Ilya,

If that is the fact(IBM already has all the tools they need at their disposal);Why are'nt they doing it and releasing AI to the whole wide world Today?

Makes one wonder what else Big Blue is hiding.

Isnt it?


The State of IT & Cybersecurity Operations 2020
The State of IT & Cybersecurity Operations 2020
Download this report from InformationWeek, in partnership with Dark Reading, to learn more about how today's IT operations teams work with cybersecurity operations, what technologies they are using, and how they communicate and share responsibility--or create risk by failing to do so. Get it now!
Commentary
Gartner Forecast Sees 7.3% Shrinkage in IT Spending for 2020
Joao-Pierre S. Ruth, Senior Writer,  7/15/2020
Slideshows
10 Ways AI Is Transforming Enterprise Software
Cynthia Harvey, Freelance Journalist, InformationWeek,  7/13/2020
Commentary
IT Career Paths You May Not Have Considered
Lisa Morgan, Freelance Writer,  6/30/2020
Register for InformationWeek Newsletters
Video
Current Issue
Key to Cloud Success: The Right Management
This IT Trend highlights some of the steps IT teams can take to keep their cloud environments running in a safe, efficient manner.
White Papers
Slideshows
Twitter Feed
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.
Sponsored Video
Flash Poll