DR SANDEEP REDDY
  • Home
  • Qualifications
  • Projects
  • Publications
  • Media
  • Conferences/Presentations
  • Blog
  • Consultancy
  • Home
  • Qualifications
  • Projects
  • Publications
  • Media
  • Conferences/Presentations
  • Blog
  • Consultancy

Why you shouldn't do a PhD that focuses on a specific AI algorithmic approach?

2/3/2019

0 Comments

 
In recent years, there has been a great deal of coverage about the dearth of PhD qualified AI data-scientists and the level of salaries qualified candidates can gain. One such piece can be found here: NYTimes article. Then you have universities complaining how their PhD qualified AI scientists are being poached by the industry thus demonstrating the demand for PhD qualified AI scientists: Guardian Article. Also, you have many universities opening numerous funded AI PhD positions such as this university: Leeds University Isn't it then obvious, a PhD in AI technology should be on all data scientists to do list. Well, as one who contemplated briefly to do a second PhD (focusing on swarm intelligence and multi-agent system in Healthcare) and who spent some time researching the necessity of completing a PhD to be across AI, I found it detrimental to undertake a PhD focusing on a specific AI algorithmic approach. Let me explain why?
  1. The first and foremost reason why a PhD (relying on a particular AI algorithmic approach) is futile to pursue is the pace at which the mathematical and algorithmic translation of intelligence is occurring. We are quickly realizing the limitations of deep learning in replicating human intelligence. Deep learning, which was even as late as last year heralded by some as a greater invention than electricity. Also, newer AI algorithmic approaches are emerging that are better suited to address real world problems than traditional AI algorithms pushed by certain AI academics/data-scientists (who likely based their PhD's on these techniques). More importantly, it is very likely we will see authorities and industry seeking approaches that can achieve general intelligence rather than narrow intelligence that current algorithmic methods are capable of. This move will require a causal deterministic approach, which current neural network algorithms seem to be incapable of delivering. Therefore, spending three years of time (at the least) studying and applying a particular algorithmic approach that can quickly become outdated seems a surefire way of washing your time and money down the drain.
  2. The second reason is the changing higher education landscape with disruptors like MOOCs and short courses delivering condensed educational content to a student cohort that is time-conscious and not willing to spend years pursuing a doctoral degree. Increasingly, the technical industry having a scarcity of PhD qualified AI data scientists has become open to accepting candidates who have completed machine learning/deep learning/AI courses through one of the many popular online education platforms. Many research scientists in big technological firms and start-up founders,without PhD's but with expertise in AI, have gone to pursue successful AI careers. With a fast changing AI technological field, a short course covering current AI technical approaches would make more practical sense.
  3. The third reason is the opportunity costs in undertaking a PhD as opposed to a shorter version like a Masters or a short course. While PhD does impart on you a expert status and the ability to rely on an university framework to pursue an in-depth education in a sought after domain, it also significantly limits your earning and social possibilities during the course of your PhD studies. The industry is hungry for AI professionals but do all the open positions need PhD qualified AI data scientists or would someone who has the expertise suffice? I suspect the latter.
This article is not meant to signal an obituary for PhD's focusing AI , but a cautionary note for someone contemplating a PhD focusing on a specific AI technique. On the other hand, if one was to choose a broad PhD topic that isn't reliant on a particular AI algorithmic approach while keeping oneself abreast with latest trends (through short technical courses during the course of their PhD studies), they will be future proofing themselves. They will also be not wedded to a particular statistical/mathematical approach and perhaps even be contributing to the development of AI-General Intelligence! Another option is to pursue a Masters course in AI that are increasingly being offered by academic institutions and online education platforms. This path would not only give you the academic and technical credibility in AI but also limit the opportunity costs that one may incur with a PhD route. So if you are thinking of undertaking a PhD in AI, think carefully as to what would suit you best!


0 Comments

    Author

    Health System Academic
    ​&
    Program Evaluation Specialist

    Archives

    March 2023
    January 2020
    April 2019
    March 2019
    February 2019
    December 2018
    October 2018
    August 2018
    March 2018
    February 2018
    January 2018
    November 2017
    October 2017
    March 2017
    December 2016
    November 2016
    May 2016
    January 2016
    January 2015

    Categories

    All

    RSS Feed