My Vision for Artificial Life

Posted October 6, 2015 by Charles Ofria in Information / 1 Comment

Recently, I was nominated for the Board of Directors for the International Society for Artificial Life (ISAL).  After serious consideration about whether I could devote the necessary time and whether I felt I would be able to make a difference, I’ve decided to run.  Normally, I try to say “no” to all such requests, but in this case I care quite a bit about the society and the field as a whole, and I want to play my part.  Below is the 250-word statement I submitted with my acceptance of the nomination, followed by some more details that I couldn’t fit in the limited space.

I am a Professor of Computer Science at Michigan State University and a dedicated member of the Artificial Life community.  I have developed the Avida digital evolution software since 1993 and published results in top-tier biology journals including Science, Nature, PNAS, and PLoS Biology.  I consistently advocate using Artificial Life techniques in biology and in 2008 I united a group of computer scientists and evolutionary biologists to form the BEACON Center for the Study of Evolution in Action, now funded by the US National Science Foundation at $47.5 million.  In 2012, I chaired ALIFE XIII, linking it with the annual BEACON Congress and substantially increased participation of biologists.

My research involves computational experiments that parallel traditional wet-lab studies to understand fundamental evolutionary dynamics.  I am passionate about understanding the evolution of biological information and complexity, including the evolution of intelligence, cooperation, and major transitions.

If elected my priorities will be to:

  1. Involve more biologists in the Artificial Life community, promote ALife studies alongside traditional wetlab experiments, and cultivate respect at funding agencies worldwide.
  2. Highlight the value of ALife for engaging the public and improving scientific understanding through art, games, and societal impacts.
  3. Push open science by exploiting the ease of experimental replication, methods validation, and transparent data collection in artificial life systems.
  4. Develop materials to streamline organizing ALIFE and ECAL conferences, local ALife-based workshops, and student organizations.
  5. Foster stronger ties and jointly-sponsored activities with the BEACON Center.
  6. Expand ISAL’s membership, diversity, and funding.

We’ll see if this statement garners me enough votes to join the board; I know most of the other candidates running and I quite respect all of them, so I won’t feel at all slighted if I don’t succeed.  That said, I’d like to take a bit more space to flesh out my six agenda items above, incorporating some important nuance.

1. I want to improve how biologists view Artificial Life, to the point where Alife systems are commonly used in concert with other experimental techniques.

Why? Our field is heavily populated with computer scientists, but biologists are less aware of the informative power of artificial systems.  What these biologists fail to realize is that under controlled conditions, ALife provides us with a rapid, finely controlled, and transparent method to learn about how living systems function.  For evolutionary studies in particular, this case was best made by the great John Maynard Smith:  “So far, we have been able to study only one evolving system and we cannot wait for interstellar flight to provide us with a second. If we want to discover generalizations about evolving systems, we will have to look at artificial ones.”  I’ve also previously made this case, showing how artificial life systems fill a critical niche between simple simulations and experiments.

How? Most biologists who take the time to learn about artificial life are quickly sold as to its value, so our best tool is education.  We must heavily advertise our conferences to biologists, provide workshops at biology conferences, encourage how-to and methods papers targeting biologists, and acknowledge those in the field who bring more members in.

2. I want Artificial Life techniques to be used frequently in schools and for public outreach.

Why? Given the shockingly poor scientific understanding of the public in many countries and the active opposition to evolution in the US (and growing elsewhere), engaging the public in science is ever more critical.  Artificial Life provides compelling vehicles to experience the wonders of biology, often accessible at any computer.  Part of the challenge many people have with understanding evolution is that it is such a gradual process in nature, so it can’t be observed, only learned about in theory.  Frustratingly, this process is like trying to master chess from just reading the rules, but never playing a game.  Artificial life brings that game into the hands of anyone interested.  From my experience, those of us who study evolution in artificial systems have a much greater intuitive grasp of evolutionary dynamics than most experts in the field.  And those students given access to these systems instinctively start engaging in the scientific method, trying to learn how they work.

How? Both our journal and conferences should push hard to publish education-focused work, and software projects should be heavily publicized.   We should also consider an additional annual award from the society for those who most effectively use artificial life systems in education or outreach.

3. I want open science to be the norm, and for us to lead the way.

Why? As the world has become more connected and we live in an information overload, most of science is only now beginning to crawl out of a 19th century mindset. Publications are making the transition to being widely available and software is more frequently being open-sourced, but expectations for easily replicable science are lagging behind.  With many Artificial Life studies, our entire experiments are done virtually, so not only can the source code be released, but everything needed for a colleague to repeat the exact experiments.  Many studies have shown the overwhelming number of scientific papers that are false due to overlooked statistical issues, but the best way to combat this phenomenon is to make sure we provide all of the material necessary to verify our results and to actually work together on understanding the generalities of our work.

How? Require all software used in published papers to be publicly released, along with all relevant configuration files.  Compel authors to include detailed methods sections (or supplementary material) that allow for easy replication.  Expect strong statistics to back important results.  Most of all, encourage the submission and publication of papers focused on verification of results, ideally across independent systems.  The Artificial Life journal is free to authors by default, but the papers are then closed access.  We should change that default to expecting authors to pay a $1000 fee to make all papers open access (and do what we can to bring that cost down, especially for ISAL members).

4. I want our community to come together frequently and easily.

Why? Having chaired an ALIFE conference a few years ago, I’m painfully aware of how complex the organization process can be, and how expensive it is.  In many cases we needed to reinvent the wheel to figure out what processes to use.  To a lesser extent, similar challenges come from running smaller workshops or student clubs.

How? For conferences and workshops, we should produce “best practices” guides that are updated by annually.  These guides would cover details including how to get papers reviewed, how to work with publishers, how to effectively schedule events, and generally how to run a high-end conference on a reasonable budget (at ALife XII, we kept fees down while providing over 30 fellowships covering 100% of room, board, and registration fees for students, and still returned a $6000 profit to ISAL).  For student groups, we can provide lists of activities for meetings, interesting topics for discussion, potential speakers to invite, and pointers to web resources.

5. I want ISAL and BEACON to work closely together to reach common goals.

Why? The BEACON Center for the Study of Evolution in Action is a Science & Technology Center, half-way through its 10 years of funding from the U.S. National Science Foundation.  Our goals of studying real-time evolution in both computational and biological systems mean that many of our members are core researchers in Artificial Life or Evolutionary Computation fields.  As such, many of the goals of BEACON and ISAL closely align on both the research and education fronts.  Research by BEACON members extend into many areas of artificial life beyond evolutionary topics, including ecology, behavior, robotics, and synthetic biology.

How? BEACON and ISAL need to open regular lines of dialog.  Regardless of whether I get elected to the ISAL board, I would be happy to be the point person within BEACON.

6. I want ISAL’s membership, diversity, and funding to steadily grow.

Why? Clearly we have a thriving community, but there are many others who naturally fit in with performing artificial-life-related research.  However, we should strive to not just increase our numbers but to improve our diversity of backgrounds, ideas, ethnicity, and especially gender.  Of course, along with any growth in numbers, comes a growth in dues collected and more opportunities for obtaining external resources.

How?  Bringing more biologists into the fold will naturally increase both total numbers and the number of women that are part of ISAL.  Furthermore, involving more women in the ISAL leadership will reduce the sense of the society being an “old boys network.” (I’ll note that both Alexandra Penn and Katie Bentley seem like excellent candidates this time around).  A secondary problem is the number of de facto members who only pay dues on years when they go to conferences (if that).  An easy way of solving this problem is to require membership (possibly for multiple years) to be eligible for awards.  Along with this fix should come many more awards: for education and outreach; for being a successful ALife ambassador in other fields; for the most cited papers.  The key is that as long as there are always highly-worthy recipients, we will only strengthen the field by having more awards given out.  And they will, in turn, encourage more people to join the society and maintain continuous membership.

Will I push toward most of these goals regardless of whether I am elected to the board of directors of ISAL?  Absolutely.  If nothing else, I hope that these ideas inspire current and future board members.  But being on the board myself will allow me to coordinate more smoothly with the community and have credentials that others will respect.

I also want to encourage all of you to join ISAL if you’re not already a member.  It’s a fantastic organization, well worth supporting.  For membership, you just need to subscribe to the Artificial Life journal.

EDIT: I updated this article after seeing the full slate of candidates.

Charles Ofria

Charles Ofria is a professor of Computer Science at Michigan State University, director of the Devolab, and deputy director of the BEACON Center for the Study of Evolution in Action.

More Posts - Website

One response to “My Vision for Artificial Life

  1. Travis Collier

    On open access publishing… I think moving to (or creating a new) overlay journal would be an excellent approach.

    The recently announce Discrete Analysis apparently manages a cost of $10 per article by using Scholastica. If you haven’t seen it, there is a nice writeup here:

    It also seems like the society would be a good place to host a sort of ‘meta journal’. I’m thinking of something along the lines of a blog talking about recent papers relevant to ALife, regardless of where they are published. Not only is that very useful, it also promotes outreach to other disciplines… “Hey, these ALife guys think my work is relevant to them” is a pretty persuasive way to hook in new folks 😉

Leave a Reply