Wednesday, March 27, 2019

Nasscom NISC 2019: Joydeep Dam on Integrating AI into the Digital Enterprise

Joydeep Dam, Director of Algorithm Development & AI at BridgeI2I Corporation, talks about the distinguishing features of an AI-enabled Digital Enterprise.  He defines a digital enterprise as one with a well-characterized way of capturing data, plus a smart way of utilizing it, so that the result is a good way of basing decisions on it. The typical case is that a particular firm might have a lot of data, which then results in a lot of fancy dashboards or reports being created (Joydeep mentions a bank which had a database that created about 2000 [two thousand] different reports every single week.) But the question is, do the insights derived from that data actually flow into decision making?

He then goes on to describe the basics of the 'AI wrapper' one could put around the data generator/report generator stage:
  1. A Watchtower - a component that (figuratively) looks over data, identifies anomalies in it, and detects errant or deviant behavior from it or in it. A Watchtower could be conceived as a Digital Dashboard with an Intelligence Module overlaid, which one could perhaps call 'DDI'. Or one could have an 'SPI' (Signal Pool Intelligence) Watchtower, which might perform roughly the same function (but with signals of a more elemental form, for example, the outputs of an IoT array). Or one could have a BI Watchtower, i.e.,  a Watchtower with Business Intelligence, on which a structure is added that extracts insights or adds graphics and visuals, or even creates entire presentations based on the input, 
  2. A Recommender - which, based on data input of various kinds (e.g., browsing, personal, credit or purchase history of an individual or firm) creates a set of customized recommendations for either the sales or the purchase side of any business transaction. This could be based on a ranking such as 'most often first' or 'most recent first' or, if likelihood can somehow be estimated, 'most likely first'.
  3. An Optimizer - a module which runs either a mathematical or heuristics-based optimization routine under a set of given constraints, and objective function(s), and provides a unique answer if available, but could also yield a set of sensitivity analyses.
  4. A Converser - (The word in this form is mainly used in French, as an intransitive verb, similar to the English 'to converse'). However Joydeep uses it in the sense of  'that which converses or that which is conversed with'. This module has the function of turning the analysis into a form that the user can 'consume' - that is, it converses with the user. This could take many forms, starting with just more intelligent graphics or video, to conversational 'chat-bots' (which may analyze text and/or speech to generate probability distributions of word occurrence, and then, based on semantic and/or grammatical rules, output text and/or speech).
These four elements can be taken as the basic elements which, if present in a given software product or service, can be talked about as representing an 'AI-enabled Digital System'.

A question asked by a member of the audience after the talk deals with the issue of Data Paucity. He asks, what do you do if the AI algorithm needs a lot of data but you only have a little of it? This is particularly the case for smaller companies and early-stage startups, but even otherwise, is a very good quesition. The answer is, the solution depends on the situation. One strategy is to wait until there is more data, or try to acquire it from elsewhere if it already exists, or to acquire (or occasionally, even generate) what might be called 'proxy data' that can 'stand in' (as an alternative which may have much the same statistical characteristics). Many AI algorithms can 'learn continuously', i.e., you can train them with the data you have, and then continue training when more data becomes available.

Thursday, March 21, 2019

Nasscom International SME Conclave (NISC) 2019: Prof V Srinivasan IIMB Workshop

There are some things that you experience in the world that really make you go wow, Wow, WOW! This workshop, conducted by Prof Vasanthi Srinivasan of IIM-B at the NASSCOM International SME Conclave 2019 in Mumbai was for me, one such.

I was amazed first of all that Nasscom was organizing a workshop on such a theme, then to find that a Bangalore-based academic was running it, but most of all by the experience of how she ran it (with really amazing enthusiasm, enormous energy and an incredibly magnetic connection with her audience). She has the audience's rapt attention from the very beginning, and holds it till the very end; begins by saying she will be provocative, and makes sure that she is; is both entertaining and educative, and has a simply great sense of humour.  It would have been great to have attended the workshop in person, but I watched it on the video that Nasscom made available on youtube, and got much out of it nevertheless. This is often taken for granted, but I want to make sure that I also thank the cameraperson(s)/videographer(s) who have done an excellent job; one certainly gets the visual perspective from different angles as would befit an event, such as this one, in which the speaker is constantly in motion! My regret however is that the videographer was unable to capture the content of the slides, which, had they been available in the video, would have been very useful indeed. But the Professor makes this up by clearly enunciating, and repeating often, the sub-themes she covers over the course of the workshop.

The first thing Prof Srinivasan emphasizes is that there is no real consensus around the definition of the buzzword 'Digitalization' (or Digital Transformation) that has gripped the IT industry (everything from Robotic Process Automation [RPA] to Machine Learning to Sentiment Analysis, and a lot else besides, comes up when people are asked to define Digitalization. But the basic point is that no matter the definition one adopts or has become comfortable with, Digitalization is Different - it is not the same thing that IT companies were doing in the past. Much practical wisdom can emerge from surveys and studies of how SMEs are faring in developing and/or adopting products and services related to (what they think of as) Digitalization, even if the technical definition is set aside for a moment.

To frame the workshop she uses three questions (which also happen to be the questions around which she has run other focus groups for Nasscom in the recent past):

1. What is changing in your environment? (i.e. in client expectations, in the competitive milieu, etc)

2. How is the change impacting you and your business (i.e., who you are, and what you do)?

3. Are you ready for the change?

What emerges from focus groups she has organized in the past around these questions is that many of the issues brought up in answers are actually independent of the scale, size, geography and client base of the enterprise. And for many of the issues, there are no differences even among product and service oriented software companies.

To the question, who's driving the Digital Transformation, the majority of firms picked the choice 'The Customer!'. Also, the majority of firms felt that the Digital Transformation was happening NOW! Among those who felt that the Digital Transformation was four or more years away, SMEs were in the majority!

On queries related to the Barriers for Digital Transformation, the most common response was 'Lack of Overall Vision for the exercise', followed by the related 'Action Items are not clear!'. For the largest companies, the fact that business units were organized as verticals or existed in silos was mentioned prominently.

And then comes one of the Half Truths that the Professor has promised to bring out during the workshop - being, 'I need more customers!' This is only half-true, she emphasizes, because in this new environment, customers may place demands on the company that it is unable to fulfill! In the previous world, any customer was fine, but not now!

On the labour market, Prof Srinivasan points out that it is shifting in favor of SMEs. Overall, job growth in the formal sector is slowing, which means there are (or will be) more people from among whom SMEs can choose their future employees. In this context, 'Lack of (skilled) Employees' does not become a specifically 'Digital' challenge, but is just a (regular business) challenge.

In addressing challenges related to scaling up small businesses, Prof Srinivasan emphasized that often the 'Entrepreneur him/herself is the biggest stumbling block'. Entrepreneurs should remember that once they start to scale up, it is no longer about them. (And if they want it to be, they should go and start up a kirana shop, she half-joked!). A scaling business will need to have systems and processes in place (such as swiping ID cards at the entrance) and these should apply to everyone without exception. Similarly, when a reskilling or training opportunity is scheduled, then everyone including the entrepreneur-boss should go and 'get trained'. Often, even 'innovative entrepreneurs' lose their 'innovative mindset' when the company starts scaling up. Their emphasis shifts to 'management' and 'transactional efficiency' instead of remaining focused on creating value for the client, the basis of entrepreneurship!

Then the discussion moves on to the criticality of delegating functions to specialists or professionals in the organization - here the interesting issue comes up of the difference between professionals who turn entrepreneur versus the 'non-professional' entrepreneur in how they carry out the delegation task. Professional entrepreneurs oddly find it hard to delegate tasks they see as very important, even if it is in an area different from the one they specialize in. Many companies in the survey answered 'the entrepreneur/owner' when asked 'who is responsible for sales in your company?'. Sales are critical, and only companies that can successfully close on deals, repeatedly, are able to make the transition to scale. But entrepreneur-bosses must nevertheless learn to delegate this task. Because 'if you are running around trying to close a deal, then who runs the company?'. To which one respondent memorably said, 'But Professor, if I don't go, the client won't close!'. Which in turn is answered by: a) Why not? and b) This is a great way to transition from doing things 'best in class' to 'worst in class'. Entrepreneurs have to be good at selling, but at some point others can do this as well or even better than them. Another issue that comes up in delegating tasks is the time-sensitivity (or insensitivity). Beware of thought patterns where one assumes the delegate/subordinate can 'surely' do the task in 24 hours if one can do it in 6 onself. Because, the subordinate is an employee, and does not devote themselves to the business 24 hours a day, but can be expected to do so for at most 8! So,  the message is: delegate, but just not at the last minute.

Towards the end of the workshop, the Professor carefully spells out the distinction between 'Enterprise Competencies' and 'Entrepreneurial Competencies'. While the former seems generally well-understood, considerable confusion exists around the latter. One of the most important required competencies in an Entrepreneur is Sensing, which is the ability to pick up significant non-verbal cues from the environment. She noted that IT entrepreneurs often lack this ability. Another entrepreneurial competency is Agility, which one might try to define, based on the literal meaning, as quickness, but which must also include, especially in this context, the ability to try something quickly, and if it doesn't work, to also realize quickly that one has failed, and to move on to the next potential solution.

Saturday, March 9, 2019

Are Generalists or Specialists Better in a Data Science Value Chain?

How Data Bias in AI-based Healthcare can Harm Vulnerable Populations

The Limits of Ethical AI 

How Reliable are Surveys of AI Adoption by Companies?

Tuesday, March 5, 2019

The AI Hype and VC Funding for Startups

Friday, March 1, 2019

The Fusion Between Quantum Physics and Machine Learning