What is the compelling question or challenge?
Through a popular, open world game with NSF highlight challenges, gamers learn science while aiding scientists as human intelligence in simulations that typically use machine learning or artificial intelligence (AI).
What do we know now about this Big Idea and what are the key research questions we need to address?
What do we know now about this Big Idea and what are the key research questions we need to address?
You are in a silent school room filled with children seemingly playing an advanced and complex game that simulates battles where the children are generals. However, these children are not playing at all, but instead managing a real battle with real people. This is the situation in “Ender’s Game.” What if I told you that this scenario isn’t just science fiction? We can use games to help scientists solve problems. A Danish group has done just this with a game called “Quantum Moves” in which the gamers intuit approximate wavefunctions which algorithms fine tune (https://www.scienceathome.org/games/quantum-moves/results/). This study showed that humans combined with an algorithm are much more powerful than each alone. Many National Science Foundation (NSF) highlights feature machine learning to solve complex problems. An alternative to algorithms is to use the nimble minds of gamers or human intelligence (HI). An NSF-sponsored open world game giving challenges based on NSF highlights can use gamers as HI in a mixed computer-human hybrid system. Challenges can also teach gamers science in a simplified system. The challenges will show gamers the myriad of sciences that are available to them. Gamers will earn badges which can be posted on social media, both advertising the game and NSF sponsored research. Through an NSF sponsored open world game, scientists will help gamers learn science and gamers will help scientists with their research serving as willing HI.
Before NSF can create this open world game which simultaneously teaches gamers and uses them as HI, we must address several questions. What is the best software paradigm for reaching a wide audience while allowing scientists to contribute modules to share their expertise and use the gamers as HI at the same time? Most scientists code in FORTRAN, C/C++, or Python without the graphics that gamers expect. Furthermore, in open world games gamers expect to have the ability to explore freely and interact with a simulated world. Can a platform be developed that would allow scientists to contribute their NSF highlight ideas and HI projects in their native language but with gamer quality graphics? Gamers seek new challenges and also a corresponding reward. Can challenges and rewards be made that meet gamers’ expectations? What are the best ways to lure gamers either individually or in teams from a psychological or educational perspective? Is giving in-game and social media recognition enough? Do pro-gamers receive greater challenges, or new entertainment opportunities like better gear or in-game money?
This open-world game provides a gateway between the virtual and real world. Three examples for modules follow. All program areas of the NSF can also contribute to the modules. In the questions above, currency may allow gamers to create their own system of economics. This environment can be used to study the development of economic systems. Materials chemistry modules would not just teach gamers about science, but use gamers to extrapolate and hypothesize new materials as HI. A module implementing the physics of fracture modules will encourage the gamers to use new and better materials to build their base and to think about how and where to place the materials in order build an effective fortification. Additionally, gamers would be better prepared to understand and enter science as a career. A growing number of NSF projects use machine learning. Investigating ways to use this HI/algorithm hybrid effectively would be critical to success in future research. New algorithms and computer paradigms would need development. Ensuring gamer satisfaction and advertisement of badges in social media will be important in attracting more players and therefore increasing the HI computing power. Not only is this new tool a way of involving gamers in research as inspired by “Ender’s Game”, but it provides a completely new research area and resource in the interaction of machine and HI.
Why does it matter? What scientific discoveries, innovations, and desired societal outcomes might result from investment in this area?
The core objective of the gaming platform is to create an environment to solve scientific puzzles more efficiently and quickly than before, where citizen science complements contributions by scientists and their algorithms to NSF highlight challenges. Citizen scientists can contribute data, ideas and solutions to projects in a guided environment, through simulations and teamwork. Wide-scale participation among US citizens in this gaming platform will result in a more-informed understanding of science and an increased exposure to scientific concepts and processes. Specific societal outcomes for youth and the scientific community are anticipated.
At least of ¼ of the gamer community is under 18 years old putting them in an impressionable time period for making career choices. Another 1/3 is between 18 and 35. These are the people who are actively making career choices. Helping these young adults and adults refine their problem-solving skills to include more of the scientific process in an enjoyable environment will create a more informed view of science and lead to increased exposure of a wide variety of science areas with a focus on NSF highlights. The wider exposure should help them choose their careers or educate them to be informed citizens.
All gamers can help scientists in areas where HI can be used to complement algorithms. HI is unique in its ability to recognize and roughly forecast patterns in complex and obscure data. This resource would be extremely helpful especially when working with AIs, because many algorithms are designed to be objective and consider all data as being potentially relevant. Unguided AIs are computationally expensive and can result in exploration of many unproductive solutions. Imperfect, fast answers can be used to allow algorithms to work efficiently at refining solutions.
Gamers may provide essential data to psychologists, sociologists and economists about group dynamics, teamwork and value systems. Psychologists can explore individual behavior patterns of a committed group of subjects working on complex problems, as trainees in highly structured environments or as contributors at the “pro” level and identify factors that lead to positive adaptive behaviors of individuals. Sociologists can design and study how team combinations and rewards affect the productivity and self-organization of groups. Economists can study artificial economies derived from challenges, contributions and rewards. Synthetic economies can be designed and tested; for example, determining the importance of traceability of currency as an improvement of value.
The points presented above are just a sample of how the platform might be used by NSF grant holders. Usage of the platform will improve the quality of research by HI and algorithm interaction. In the long run, the platform will result in more and better scientists.
If we invest in this area, what would success look like?
Initial success would mean that 1) the game is popular among gamers and 2) scientists begin to use the HI and algorithms to solve scientific problems. Initial success would lead gamers to spend time on more scientific challenges, getting better at solving these challenges and spreading the word. Gamers would experience a wider variety of sciences and develop a curiosity toward them. The gamers would begin working in teams and helping each other, which teaches them team building, an important life skill. Additionally, this would be a gender- and race-free platform, which limits the prejudice that can be brought to the game.
Scientists will be able to use the HI in their game modules to solve complex problems and then refine those solutions using their algorithms, this combination provides answers more reliably than just using an algorithm. In addition, social scientists and psychologists would be able to use the game as a laboratory where they can study economies and team forming. Longer term success will mean more students will consider science as a career and will be better informed of the many areas of science and in particular the hot topics at NSF. If only 1% of all gamers in the United States were to participate we could recruit 1 million new participants in science.
Why is this the right time to invest in this area?
The growing number of NSF initiatives on machine learning highlights its current usefulness to many different kinds of science. Machine learning algorithms may be more efficient when incorporated into an HI/algorithm machine. The Denmark group’s success using the gamers’ intuition to get approximate solutions to a quantum mechanical problem, which were then refined using an algorithm, shows that now is the right time.
The US needs more scientists and the best way to make sure more people will become scientists is to show the younger generation that science is a viable and fun career choice for them. Almost 70% of the US is made up of people who play video games and this number is growing. Just a decade ago the gamer population was only 27% of the United States. The age distribution of gamers is a perfect target for youth contemplating their career options before and during college.
References
Reference #1
Show MoreShow Less