IJCNN 2019 Panels


Panel 1: Funding Opportunities in Neural Networks and Biologically Inspired AI Research

Panel Chair: Robert Kozma

Panelists (tentative list): Nandini Iyer, AFOSR; Anthony Kuh, NSF; Hava Siegelmann, DARPA; Wlodek Duch, INCF, EU, more TBA.

Panel 1 Abstract:

This panel addresses novel avenues to support neural network research. Representatives of funding agencies and leading experts in the field will describe research challenges and funding opportunities. Which cutting edge areas are at the focus of new funding initiatives? The panel will provide a forum for thorough discussions on these topics between the panelists. It is expected to have an intensive questions and answers section with the audience.

Panel 2: NSF Career Award Winners in Intelligent and Adaptive Systems

Panel Co-Chairs: Anthony Kuh, NSF; Robi Polikar, Rowan University; Haibo He, University of Rhode Island

Moderator: Anthony Kuh, NSF

Panelists: Yiran Chen, Duke University; Silvia Ferrari, Cornell University:  Haibo He, University of Rhode Island; Robi Polikar, Rowan Universit

Panel 2 Abstract:

This panel will feature past NSF Career Award winners that received awards from the National Science Foundation (NSF) in the Electrical, Communications, and Cyber Systems Division in the Intelligent and Adaptive Systems Area.  The panel will take approximately two hours. First the panelists will give short presentations (10-12 minutes apiece) about what they did for their NSF Career Award and how it advanced their careers. This will be followed by a question and answer session with the audience (40- 60 minutes).  We anticipate having 5 to 6 past NSF Career Award winners.  We have listed four confirmed panelists.

There are currently many lucrative career opportunities for researchers in AI, data science and machine learning in large companies (e.g. Amazon, Google, Facebook, Apple, Microsoft) and also in numerous startup firms. We want to showcase successful academic careers.  We will have former NSF Career Award winners in intelligent and adaptive systems discuss their NSF Career Awards and how it helped them launch their academic careers.  This should be of great interest to all participants, but especially to junior faculty, postdocs, and graduate students.  There will be significant time for questions and answers so that the audience can ask panelists questions ranging from how the panelists got their NSF Career award (including tips for writing proposals) to how they used their Career Award to achieve success in research and academia.

Panel 3: Deep Learning: Hype or Hallelujah?

Panel Chair: Vladimir Cherkassky, University of Minnesota, USA 

Panelists: in progress

Panel 3 Abstract:

In the last 3-5 years there have been tremendous interest in the so-called Deep Learning Networks (DLN). Unfortunately, there is little theoretical understanding of DLNs and many claims about their superior capabilities often represent technical marketing. These are 3 main types of arguments made by supporters of DLNs: (1) automatic feature selection by DLNs; (2) biological flavor of DLN learning; (3) their competitive generalization performance on several large real-life application data sets, such as image recognition, etc. One may adopt more cautious and skeptical viewpoint about DLNs arguing that:

  • There is no theoretical reason for DLNs to perform better than other methods. So their superior performance (on some application data) is simply due to good match between statistical characteristics of the data at hand and DLN parameterization.
  • All existing empirical results using DLN on large data sets effectively implement Empirical Risk Minimization (ERM) inductive principle (under VC-theoretical framework).
  • In spite of all hype and publicity, here have been no systematic empirical comparison studies using synthetic data sets under (under small size setting). 
  • Claims about biological motivation behind DL are rather naive (especially since such claims are made by computer scientists and engineers, not neuroscientists).

The panel will present opposing views on DL, followed by questions from the audience. The panel starts with a critical view by panel chair, continues with responses from panelists, some follow up questions, and questions from the audience.