Tuesday, September 10, 2013

Three Years after my PhD Defense: the Gap between Academic Research and Industry Needs

It has been three full years since I defended my dissertation. There has been a lot going on during these three years. I would like to put some of my random thoughts into a few blog posts starting from this one:

The Gap between Academic Research and Industry Needs.

Three years ago, I wrote the core methodology of my PhD work into two load forecasting papers and submitted them to a top journal. With negative comments from 10 reviewers, both papers were rejected. I looked at the comments and felt that they were mostly nonsense. I guessed the reviewers must be inexperienced graduate students. As a young guy with a full confidence of the quality of my research, I wrote a complain letter to the Editor-in-Chief. Surprisingly, he kindly replied to me and said that his reviewers were competent. I felt that it must be a joke, so I replied back and told him how nonsense these reviewers' comments are. I never heard a word back since then. In fact, the nonsense review comments were not totally useless to me. I used them as counterexamples while teaching the "Frequently Made Mistakes" chapter of my load forecasting course. Some of them will show up in my upcoming book as well.

I was fortunate enough that all of my committee members were top experts in their field and recognized the quality of my work - they let me graduate without a journal publication. Now looking at their decision three years ago, I think they did the right thing. The work being rejected by that journal was soon being deployed to many other utilities. Many trading firms and utilities started using my methodology in their operations environment. Recently, it was commercialized by the largest analytics vendor in the world. During the past three years, many people have read my dissertation and offered me very nice comments. Some of them even bought the paper copy while the PDF is freely accessible through NCSU library. My load forecasting course, which was primarily based on my dissertation, was one of the most popular courses among hundreds of SAS training courses...

On the other hand, I've been getting emails from graduate students during these three years asking for help with their job hunting. Many of them have a long list of papers published in top journals, but they have a hard time launching a job.

What is going on?

My papers that were proven to be highly valuable to the industry were rejected by a top journal. Many students with extensive research experience couldn't find a job.
Here are my two cents:
  1. Too many irresponsible authors. Many people are publishing papers for the sake of getting promoted or graduated. They don't care about the quality of the paper and whether it's useful or not. "It's none of my business," they think, "because I will never read these papers again." They will submit the paper to one journal after another until it gets published. They will publish papers on with similar contents again and again to increase the publication count.
  2. Too many irresponsible reviewers. Receiving so many papers every year, the journals have to find reviewers to review them. Since there are only limited amount of experts in each field, it is impossible for them to review all the papers. So a lot of journals use graduate students to review papers. Many graduate students have no idea about what the industry needs. All they know is based on the papers they read from those journals, which were mostly written by the irresponsible authors I mentioned above.
With such a negative feedback loop, the overall "usefulness" of the academic papers is getting lower and lower. In fact, today I just received another email from a researcher sharing with me his observations: utilities are not paying much attention to the scientific papers. He has several papers that demonstrated very significant contribution, but the utility people are not reading them. My answer can be simple: even if your paper is good, it is buried by thousands of other bad ones. The utility guys have no time reading 1000 garbage papers to find a good paper. (Tao's Recommended Reading List for Energy Forecasters)

How to fix this?

I observed this problem three years ago when my papers were rejected. After discussing with my mentor, some of my professors, and some other colleagues in the industry, I decided to do something for the industry. With the help from many people, I started up a group, IEEE Working Group on Energy Forecasting. Our mission is to bridge the gap between academic research and industry needs by promoting practical work to the community. Over the past three years, we organized 5 panel sessions with more than 30 practical talks, an energy forecasting competition with hundreds of participants worldwide, an IEEE tutorial on energy forecasting, and a special issue on forecasting of IEEE Transactions on Smart Grid. With increasing number of attendance to our sessions at IEEE conferences, we recognize that this research field is getting better organized than before. And the industry people are getting more and more interested and involved in what university professors are doing.

There is still a long way to go to bridge the gap. If you have been having the same frustration, no matter you are on the industry side or academic side, and you are willing to devote some of your time to improving the forecasting practice of the utility industry, we now have a group for you. Please leave your comments below or send an email to me with your thoughts.

No comments:

Post a Comment

Note that you may link to your LinkedIn profile if you choose Name/URL option.