The Wolfram Physics Project: A New Kind of (Doing) Science

The Wolfram Physics Project is an example of how open science can be done. The world of theoretical physics is not full of crazy billionaires with a vision to advance humanity. Fortunately, there is Stephen Wolfram, founder and CEO of Wolfram Research, designer of Mathematica and the Wolfram Alpha answer engine, the brain behind the Wolfram Language and author of the 1000+ pages scientific book A New Kind of Science. I suggest reading his vita to learn of an eccentric, terribly interesting man who has polarized the science community for decades.

On April 14th, he has launched his newest venture, called the Wolfram Physics Project, to find the fundamental theory of physics. In a long blog post, Wolfram describes his ideas to model reality by complex hypergraphs exhibiting emerging properties like space and matter, based only on basic initial conditions and a series of simple update events. For the lazy reader, watch the online colloquium he recently gave to the Harvard Black Hole Initiative.

Fig. 2: Visual summary of the Wolfram Physics Project (Source: Wolfram Physics Project)

The work of Wolfram and his colleagues is certainly interesting and intellectually highly stimulating. I suggest you read the documentation provided and enjoy the brilliantly clear ideas. While it cannot be said yet if they are on the right track — the theory is still in its infancy — it is definitely clear that the open, collaborative approach the Wolfram Physics Project uses is groundbreaking. Here are three things that the team is doing right.

Open access documentation

It is not difficult to find info on the research of Wolfram. There is heaps of documentation available online and openly accessible. Everyone is invited to join discussions and materials/tools are openly shared. Find below a long (but not exhaustive) list of things you will find on the project website:

• a long and detailed blog post describing the high-level ideas of the project in a non-technical way
• the backstory of the project
• hundreds of pages of technical exposition
• technical papers describing their findings to a scientific audience
• regular project updates
• a Q&A sectionarchives of previous work dating back to 1994
• regular working sessions streamed on Youtube, Twitch and Facebook where interested people can jump into the discussion
• the Stephen Wolfram Podcast

Being an outsider of academia himself, Wolfram is following his own path and opts for openness and transparency in his research. His inclusive approach may make up for his missing ties to the scientific community and could encourage young scientist to jump on this topic. I certainly really enjoyed exploring the project page.

A focus on visualization

Theoretical physics papers can be a bit hard to digest at times and are usually very formula-laden. To get a feeling on how information is usually conveyed, please have a look at the two of the most cited high-energy physics papers by the legendary physicists Edward Witten and Juan Malcadena, picked from a summary on inspirehep.net.

Some researchers, however, do have a rather visual understanding of science. Great examples are the spin networks developed by Roger Penrose or Feynman diagrams popularized by Richard Feynman (who also happened to be on the thesis committee of the 20-year-old Wolfram at Caltech).

For Wolfram, visualizations seem to be an integral part of the work and ideation process. Making use of the powerful Wolfram language, he and his coworkers have built an extensive toolset to visualize their models. See Fig. 3 for example visualization of some exemplary toy hypergraphs.

Fig. 3: Possible universes and their update rules (Source: Wolfram Physics Visual Gallery)

The mathematics behind all of this is still very complicated, but visualizing such complex ideas makes thinking about fundamental theories of physics more approachable for non-experts and experts alike.

Shareable code and software tools

In 2015, I had a conversation with fellow physicist Martin Kavalar who just had founded his company nextjournal. He shared with me his vision on how to make scientific research both more reproducible and more accessible by sharing computational notebooks everyone can validate, copy and remix as he or she feels fit. Being aware of both the growing impact of computational models in physics and the missing level of programming experience in the science community, I was immediately sold on his ideas.

The Wolfram Physics Project follows the same concept. All software and visualization tools are based on the Wolfram Language and freely usable. You can duplicate, change and run the code. You will also find tutorials on the project teaching you the basic ideas and guiding you on how to use the computational tools provided. See Fig. 4 for some example work that one can use and rework. I suggest to try and change some parameters in these notebooks to immediately see the effects.

Fig. 4: Examples of remixable notebooks (Source: Wolfram Physics Archives)

To illustrate how accessible this really is, I updated the code in the provided notebooks to create my own universe creation rule which led to the hypergraph shown in Fig. 5. With only basic knowledge of the Wolfram Language, I was able to create the koi-shaped universe below within minutes.

Fig. 5: My very own universe bears an eerie resemblance to a koi fish from above

Inspiration from an outsider

Not every scientist can have the same resources at his or her disposal as Stephen Wolfram. But still, there are a couple of things to learn from how he approaches his search for a fundamental theory.

To build a community around your field of research, it is essential to share knowledge and provide open access to your publications. Even old and outdated material can be helpful and a big plus would be to provide information about outdated approaches and previous mistakes. Note that this does not remove the need for proper peer review.

Thinking of visualization of your ideas provides a deep level of understanding and helps to communicate your theories. This is usually not a trivial task and needs some effort, but will pay off in the long run.

I love the approach of sharing and adapting notebooks. Everyone can reuse code and build on the ideas provided. This makes scientific exploration fast and efficient and should be a standard collaboration tool. You may use Wolfram’s platform, but there are also alternatives like shared Jupyter notebooks or nextjournal. Although less interactive, even making your source code available on GitHub is much preferred to letting it go mouldy on some university server.

In summary, the Wolfram Physics Project is following an extremely interesting path to hunting for a fundamental theory of physics, redefining how we think about time and space. It is also following an interesting approach on how to build and enable a community of researchers that hopefully finds imitators across the science community and has the potential to leading to a new way of doing science.