Building AI Is Hard—So Facebook Is Building AI That Builds AI

“…companies like Google and Facebook pay top dollar for some really smart people. Only a few hundred souls on Earth have the talent and the training needed to really push the state-of-the-art [AI] forward, and paying for these top minds is a lot like paying for an NFL quarterback. That’s a bottleneck in the continued progress of artificial intelligence. And it’s not the only one. Even the top researchers can’t build these services without trial and error on an enormous scale. To build a deep neural network that cracks the next big AI problem, researchers must first try countless options that don’t work, running each one across dozens and potentially hundreds of machines.”


This article represents a true picture of where we are today for the average consumer and producer of information, and the companies that repurpose information, e.g. in the form of advertisements.  
The advancement and current progress of Artificial Intelligence, Machine Learning, analogously paints a picture akin to the 1970s with computers that fill rooms, and accept punch cards as input.
Today’s consumers have mobile computing power that is on par to the whole rooms of the 1970s; however, “more compute power” in a tinier package may not be the path to AI sentience.  How AI algorithm models are computed might need to take an alternate approach.  
In a classical computation system, a bit would have to be in one state or the other. However quantum mechanics allows the qubit to be in a superposition of both states at the same time, a property which is fundamental to quantum computing.
The construction, and validation of Artificial Intelligence, Machine Learning, algorithm models should be engineered on a Quantum Computing framework.

Time Lock Encryption: Seal Files in Cloud Storage

Is there value in providing users the ability to apply “Time Lock Encryption” to files in cloud storage?  Files are securely uploaded by their Owner.  After upload no one, including the Owner, may decrypt and access / open the file(s).   Only after the date and time provided for the time lock passes, files will be decrypted, and optionally an action may be taken, e.g. Email a link to the decrypted files to a DL, or a specific person.

Additionally, files might only be decrypted ‘just in time’ and only for the specific recipients who had received the link.  More complex actions may be attached to the time lock release such as script execution using a simple set of rules as defined by the file Owner.

The encryption should be the highest available as defined by the regional law in which the files reside.  Note: issue with cloud storage and applicable regional laws, I.e. In the cloud.

Already exists as a 3rd party plugin to an existing cloud solution?Please send me a link to the cloud integration product / plug in.

AI Personal Assistants are “Life Partners”

Artificial Intelligent (AI)  “Assistants”, or “Bots” are taken to the ‘next level’ when the assistant becomes a proactive entity based on the input from human intelligent experts that grows with machine learning.

Even the implication of an ‘Assistant’ v.  ‘Life Partner’ implies a greater degree of dynamic, and proactive interaction.   The cross over to becoming ‘Life Partner’ is when we go ‘above and beyond’ to help our partners succeed, or even survive the day to day.

Once we experience our current [digital, mobile] ‘assistants’ positively influencing our lives in a more intelligent, proactive manner, an emotional bond ‘grows’, and the investment in this technology will also expand.

Practical Applications Range:

  • Alcoholics Anonymous Coach , Mentor – enabling the human partner to overcome temporary weakness. Knowledge,  and “triggers” need to be incorporated into the AI ‘Partner’;  “Location / Proximity” reminder if person enters a shopping area that has a liquor store.  [AI] “Partner” help “talk down”
  • Understanding ‘data points’ from multiple sources, such as alarms,  and calendar events,  to derive ‘knowledge’, and create an actionable trigger.
    • e.g. “Did you remember to take your medicine?” unprompted; “There is a new article in N periodical, that pertains to your medicine.  Would you like to read it?”
    • e.g. 2 unprompted, “Weather calls for N inches of Snow.  Did you remember to service your Snow Blower this season?”
  • FinTech – while in department store XYZ looking to purchase Y over a certain amount, unprompted “Your credit score indicates you are ‘most likely’ eligible to ‘sign up’ for a store credit card, and get N percentage off your first purchase”  Multiple input sources used to achieve a potential sales opportunity.

IBM has a cognitive cloud of AI solutions leveraging IBM’s Watson.  Most/All of the 18 web applications they have hosted (with source) are driven by human interactive triggers, as with the “Natural Language Classifier”, which helps build a question-and-answer repository.

There are four bits that need to occur to accelerate adoption of the ‘AI Life Partner’:

  1. Knowledge Experts, or Subject Matter Experts (SME) need to be able to “pass on” their knowledge to build repositories.   IBM Watson Natural Language Classifier may be used.
  2. The integration of this knowledge into an AI medium, such as a ‘Digital Assistant’ needs to occur with corresponding ‘triggers’ 
  3. Our current AI ‘Assistants’ need to become [more] proactive as they integrate into our ‘digital’ lives, such as going beyond the setting of an alarm clock, hands free calling, or checking the sports score.   Our [AI] “Life Partner” needs to ‘act’ like buddy and fan of ‘our’ sports team.  Without prompting, proactively serve up knowledge [based on correlated, multiple sources], and/or take [acceptable] actions.
    1. E.g. FinTech – “Our schedule is open tonight, and there are great seats available, Section N, Seat A for ABC dollars on Stubhub.  Shall I make the purchase?”
      1. Partner with vendors to drive FinTech business rules.
  4. Take ‘advantage’ of more knowledge sources, such as the applications we use that collect our data.  Use multiple knowledge sources in concert, enabling the AI to correlate data and propose ‘complex’ rules of interaction.

Our AI ‘Life Partners’ may grow in knowledge, and mature the relationship between man and machine.   Incorporating derived rules leveraging machine learning, without input of a human expert, will come with risk and reward.

Financial Technology – Categories of FinTech Solutions

FinTech refers to new solutions which demonstrate an incremental or radical / disruptive innovation development of applications, processes, products or business models in the financial services industry. These solutions can be differentiated in at least five areas.

  1. First, the banking or insurance sector are distinguished as potential business sectors. Solutions for the insurance industry are often more specifically named “InsurTech”.
  2. Second, the solutions differ with regard to their supported business processes such as financial information, payments, investments, financing, advisory and cross-process support.[4] An example is mobile payment solutions.
  3. Third, the targeted customer segment distinguishes between retail, private and corporate banking as well as life and non-life insurance. An example are telematics-based insurances that calculate the fees based on customer behaviour in the area of non-life insurances.
  4. Fourth, the interaction form can either be business-to-business (B2B), business-to-consumer (B2C) or consumer-to-consumer (C2C). An example are social trading solutions for C2C.
  5. Fifth, the solutions vary with regard to their market position. Some for example provide complementary services such as personal finance management systems, others focus on competitive solutions such as e.g. peer-to-peer lending.

Global investment in financial technology increased more than twelvefold from $930 million in 2008 to more than $12 billion in 2014

Source: Financial technology – Wikipedia, the free encyclopedia

FinTech: End to End Framework for Client, Intermediary, and Institutional Services

Is it all about being the most convenient,  payment processing partner, with an affinity to the payment processing brand?  It’s a good place to start; the Amazon Payments partner program.

FinTech noun : an economic industry composed of companies that use technology to make financial systems more efficient

Throughout my career, I’ve worked with several financial services  teams to engineer, test, and deploy solutions.  Here is a brief list of the FinTech solutions I helped construct, test,  and deploy:

  1. 3K Global Investment Bankers – proprietary CRM platform, including Business Analytics, Business Objects Universe.
  2. Equity Research platform, crafted based on business expertise.
    • Custom UI for research analysts, enabled the analysts to create their research, and push into the workflow.
    • Based on a set of rules,  ‘locked down’ part of the report would  “Build Discloses” , e.g. analyst holds 10% of co.
    • Custom Documentum workflow would route research to the distribution channels; or direct research to legal review.
  3. (Multiple Financial Org.) Data Warehouse middleware solutions to assist organizations in managing,  and monitoring usage of their DW.
  4. Global Derivatives firm, migration of mainframe system to C# client / Server platform
  5. Investment Bankers and Equity Capital Markets (ECMG)  build trading platform so teams may collaborate on Deals/Trades.
  6. Global Asset Management Firm: On boarding and Fund management solutions, custom UI and workflows in SharePoint

*****

A “Transaction Management Solution” targets a mixture of FinTech services, primarily “Payments” Processing.

Target State Capabilities of a Transaction Management Solution:

  1. Fraud Detection:  The ability to identify and prevent fraud exists within many levels of the transaction from facilitators of EFT to credit monitoring and scoring agencies.  Every touch point of a transaction has its own perspective of possible fraud, and must be evaluated to the extent it can be.
    • Business experts (SMEs)  and technologists continue to expand the practical applications of Artificial Intelligence (AI) every day.  Although extensive AI fraud detection applications  exists today incorporating human populated Rules Engines,  and AI Machine learning (independent rule creation).
  2. Consumer “Financial Insurance” Products
    • Observing a business, end to end transaction may provide visibility into areas of transaction risk.   Process  and/or technology may be adopted / augmented to minimize the risk.
      • E.g. eBay auction process has a risk regarding the changing hands of currency and merchandise.  A “delayed payment”, holding funds until the merchandise has been exchanged minimized the risk, implemented using PayPal.
    • In product lifecycle of Discovery, Development, and Delivery phases, converting concept to product.
  3. Transaction Data Usage for Analytics
    • Client initiating transaction,  intermediary parties, and destination of funds may all tell ‘a story’ about the transaction.
    • Every party within a transaction, beginning to end, may benefit from the use of the transaction data using analytics.
      • e.g. Quicken – personal finance management tool; collects, parses, and augments transaction data to provide client  analytics in the form of charts / graphs, and reports.
    • Clear, consistent, and comprehensive data set available at every point in the transaction lifecycle regardless of platform .
      • e.g. funds transferred between financial institutions may  have a descriptions that are not user friendly, or may not be actionable, e.g. cryptic name, and no contact details.
      • Normalizing data may occur at an abstracted layer
    • Abstracted, and aggregated data used for analytics
      • e.g. average car price given specs XYZ;
      • e.g. 2. avg. credit score in a particular zip code.
    • Continued growth opportunities, and challenges
      • e.g. data privacy v. allowable aggregated data
  4. Affinity Brand Opportunities Transaction Management Solution
    • eWallet affinity brand promotions,
      • e.g. based on transaction items’ rules; no shipping
      • e.g.2. “Cash Back” Rewards, and/or Market Points
      • e.g.3. Optional, “Fundraiser” options at time of purchase.
  5. Credit Umbrella: Monitoring Use Case
    • Transparency into newly, activated accounts enables the Transaction Management Solution (TMS) to trigger a rule to email the card holder, if eligible, to add card to eWallet

Is Intuit an acquisition target because of Quicken’s capabilities to provide users consistent reporting of transactions across all sources?  I just found this note in Wiki while writing this post:

Quicken is a personal finance management tool developed by Intuit, Inc. On March 3, 2016, Intuit announced plans to sell Quicken to H.I.G. Capital. Terms of the sale were not disclosed.[1]

For quite some time companies have attempted to tread in this space with mixed results, either through acquisition or build out of their existing platforms.  There seems to be significant opportunities within the services, software and infrastructure areas.  It will be interesting to see how it all plays out.

Inhibitors to enclosing a transaction within an end to end Transaction Management Solutions (TMS):

  • Higher level of risk (e.g. business, regulatory) expanding out service offerings
  • Stretching too thin, beyond core vision, and lose sight of vision.
  • Transforming tech  company to hybrid financial services
  • Automation, streamlining of processes, may derive efficiencies may lead to reduction in staff / workforce
  • Multiple platforms performing functions provides redundant capabilities, reduced risk, and more consumer choices

 Those inhibitors haven’t stopped these firms:

Payments Ecosystem
Payments Ecosystem

 

The Race Is On to Control Artificial Intelligence, and Tech’s Future

Amazon, Google, IBM and Microsoft are using high salaries and games pitting humans against computers to try to claim the standard on which all companies will build their A.I. technology.

In this fight — no doubt in its early stages — the big tech companies are engaged in tit-for-tat publicity stunts, circling the same start-ups that could provide the technology pieces they are missing and, perhaps most important, trying to hire the same brains.

For years, tech companies have used man-versus-machine competitions to show they are making progress on A.I. In 1997, an IBM computer beat the chess champion Garry Kasparov. Five years ago, IBM went even further when its Watson system won a three-day match on the television trivia show “Jeopardy!” Today, Watson is the centerpiece of IBM’s A.I. efforts.

Today, only about 1 percent of all software apps have A.I. features, IDC estimates. By 2018, IDC predicts, at least 50 percent of developers will include A.I. features in what they create.

Source: The Race Is On to Control Artificial Intelligence, and Tech’s Future – The New York Times

The next “tit-for-tat” publicity stunt should most definitely be a battle with robots, exactly like BattleBots, except…

  1. Use A.I. to consume vast amounts of video footage from previous bot battles, while identifying key elements of bot design that gave a bot the ‘upper hand’.  From a human cognition perspective, this exercise may be subjective. The BattleBot scoring process can play a factor in 1) conceiving designs, and 2) defining ‘rules’ of engagement.
  2. Use A.I. to produce BattleBot designs for humans to assemble.
  3. Autonomous battles, bot on bot, based on Artificial Intelligence battle ‘rules’ acquired from the input and analysis of video footage.

So Much Streaming Music, Just Not in One Place

In the old days, you never knew which CDs the record store would have in stock.  That limitation of physical media was supposed to be solved by digital. Back in the 1990s, technology evangelists and music fans alike began to talk about a “celestial jukebox” — a utopian ideal in which every song ever recorded would be available at a click.  In reality, even a celestial jukebox has gaps. Or more precisely, numerous jukeboxes have come along – iTunes, Pandora, Spotify, SoundCloud, YouTube – and each service has had gaps in its repertoire. And those gaps have been growing bigger and more complicated as artists have wielded more power in withholding their music from one outlet or another.

Source: So Much Streaming Music, Just Not in One Place – The New York Times

Additional Editorial:

Published music libraries are numerous, and have scattered artist coverage for one reason or another.  Music repositories may overlap, or lack completeness of coverage.

As expressed in “As a Data Deluge Grows, Companies Rethink Storage“, creating a system similar to the Internet Domain Name System for “Information Asset Libraries” would help in numerous ways.  Front end UIs may query these “Information Asset (object) libraries” to understand the availability of content across the Internet.

The Domain Name System (DNS) is a hierarchical decentralized naming system for computers, services, or any resource connected to the Internet or a private network.

Another opportunity would be to leverage the existing DNS platform for managing these “Information Asset Repositories”

In a relatively cost restrained implementation, a DNS type effort can be taken up by the music industry.  From artists to distribution channels, existing music repositories can be leveraged, and within months, a music aficionado may go to any participating platform, and search for an artist, title, album, or any other indexed meta data, and results across ‘Information Asset Repositories’ would be displayed to the user with a jump link to the registered information asset in the library.

Small independent artists need just populate a spreadsheet with rows that contain a row for each asset, and all the ‘advertised’ meta data.  Their Information Asset library may be a single flat file, i.e. XML, that conforms to a basic record/row structure.  The independent artist places this file on their web site, e.g. in their root folder, and informs their ISP of the address record type, and it’s location.  A new DNS record specification may need to be created, e.g. MX record.

As a Data Deluge Grows, Companies Rethink Storage

At Pure Storage, a device introduced on Monday holds five times as much data as a conventional unit.

  • IBM estimates that by 2020 we will have 44 zettabytes — the thousandfold number next up from exabytes — generated by all those devices. It is so much information that Big Blue is staking its future on so-called machine learning and artificial intelligence, two kinds of pattern-finding software built to cope with all that information.
  • Pure Storage chief executive, Scott Dietzen, “No one can look at all their data anymore; they need algorithms just to decide what to look at,”

Source: As a Data Deluge Grows, Companies Rethink Storage – The New York Times

Additional Editorial:

Pure Storage is looking to “compress” the amount of data that can be stored in a Storage Array using Flash Memory, “Flashblade”.   They are also tuning the capabilities of the solution for higher I/O throughput, and optimized, addressable storage.

Several companies with large and growing storage footprints have already begin to customize their storage solutions to accommodate the void in this space.

Building more storage arrays is a temporary measure while the masses of people, or fleets of cars turn on their IoT enabled devices.

Data is flooding the Internet, and innumerable, duplicate ‘objects’  of information, requiring redundant storage, are prevalent conditions. A registry, or public ‘records’ may be maintained.   Based on security measures, and the public’s appetite determine what “information objects” may be centrally located.  As intermediaries, registrars may build open source repositories, as an example, using Google Drive, or Microsoft Azure based on the data types of ‘Information Objects”

  • Information object registrars may contain all different types of objects, which indicate where data resides on the Internet.
    • vaguely similar to Domain name registrar hierarchy
    • another example, Domain Name System (DNS) is the best example of the registration process I am suggesting to clone and leverage for all types of data ranging from entertainment to medical records.
  • Medical “Records”, or Medical “Information Objects”
    • X-ray images, everything from dental to medical, and correlating to other medical information object(s),
  • Official ‘Education’ records from K-12 and beyond, e.g. degrees and certifications achieved;
  • Secure, easy access to ‘public’ ‘information objects’ by the owner, and creator.  Central portal(s) driving user traffic.  Enables ‘owner’ of records to take ‘ownership’ of their health, for example

Note: there are already ‘open’ platforms being developed and used for several industries including medical; with limed access.  However, the changes I’m proposing imposes a ‘registrar’ process whereby portals of information are registered, and are interwoven, linking to one another.

It’s an issue of excess weight upon the “Internet”, and not just the ‘weight’ of unnecessary storage, the throughput within a weaved set of networks as well.

Think of it in terms of opportunity cost.  First quantify what an ‘information object’, or ‘block of data’ equates to in cost.  It seems there must already be a measurement in existence, a medium amount to charge / cost per “information object”.  Finally, for each information object type, e.g. song, movie, news story, technical specifications, etc. identify how many times this exact object is perpetuated in the Internet.

Steps on reducing  data waste:

  • Without exception, each ‘information object’ contains an (XML) meta data file.
  • Each of the attributes describing information objects are built out as these assets are being used; e.g. proactive autopopulate search, and using an AI Induction engine
  • X out of Y metadata type and values are equivalent
    • the more attributes correlate to one or more objects, the more likely these objects are
      • related on some level, e.g. sibling, cousin
      • or identical objects, and may need meta relationship update
    • the metadata encapsulates the ‘information object’

Another opportunity to organize “Information Asset Objects” would be to leverage the existing DNS platform for managing “Information Asset Repositories”.   This additional Internet DNS structure would enable queries across information asset repositories.   Please see “So Much Streaming Music, Just Not in One Place”  for more details.

Aerial Photography Communities Aligned by Interest, Broadcast in Realtime

Although I fail to see the excitement and mass appeal of aerial drone use, the hobby has taken off on the tail end of military UAV.  Just like the stationary 24/7 webcams, and web sites that catalog these cams, the drone networks, or communities may spawn entirely new interest groups.

Do you have a drone with the ability to stream video in realtime?  You may drive a following to your stream based upon a multitude of reasons, e.g. location; subject(s) of focus.  Once airborne, your drone may broadcast to a web site that tracks your drone’s latitude and longitude, as well as dynamically tagging the feed with relevant frame data.  Object recognition may scan each frame, or a sampling for ‘objects of interest’.  Objects of interest may appear to a community of enthusiasts as a ‘tag cloud’.  Users may select a tag, and drill down to a list of active feeds.  Alternatively, users may bring up a map view to show the active drones flights.  The drones may also show ‘bread crumbs’ of a flight, maybe the last 1/2 hour,  the buffered video available.  Could be just an extension of YouTube, or a new platform designed entirely around Drone Realtime Streaming.

Building Apps Incorporating the AI Power of IBM Watson’s Cognative Computing Cloud

IBM Watson’s APIs are available today so teams may ramp up quickly and use IBM’s cognitive computing engine.  From IBM Watson’s site, it seems like anyone may build against their cognitive computing platform.  In addition,  your team may submit to be ‘Featured’ in their application Gallery.  Explore the library of featured applications produced by this partnership.  At the time of this writing, there were 14 applications.
Several of these apps have been created by IBM to showcase their technology.  IBM Watson APIs are categorized into ‘Services’ used:
  • Dialog
  • Natural Language Classifier
  • AlchemyData News
  • Personality Insights
  • Tradeoff Analytics
  • Speech to Text
  • Language Translation
  • Text to Speech
  • Visual Recognition
  • Concept Insights
  • Relationship Extraction

They sound like AI,BI comprehensive services, but in full disclosure, I’ve not read though the API docs available by IBM.  It can be found here, grouped by IBM Watson’s Cognitive Services.

One of the applications powered by IBM Watson in their gallery is a “News Explorer”, which leverages the Service ‘AlchemyData News’.
The app runs in a browser, and consists of 5 main User Interface components.  The centrally placed, “News Network” widget similar to a mind map, correlates articles, companies, organizations, and people.  Visually it displays these components and their relationships in groupings similar to a relationship tree.
 News Network
The left side of the screen has a table called ‘Details’, one column with short descriptions of the stories.  From the UI perspective, it enables users to follow the data from left to right, from details to graphical representations.
Details
The right most side of the screen contains a world map leveraged as a heat map in which all the News is derived.
Locations
Right under the ‘Locations’ widget, there is a ‘Topics’ tag cloud.
TopicsTags
I encourage you to check out the News application, click here.
In addition to UI drill down within the widgets, there is a comprehensive search capacity.
Are you ready to compete with Siri, or Cortana, or build your own Expert solution?  Looks like IBM is empowering you to do so!

Smart Solutions

Powered by "To Do List Member"
Skip to toolbar