Simplifying High-Performance Computing for the Scientific Community

Project Robbie, a platform designed to eliminate traditional IT and cloud complexities for scientific researchers by providing them with intuitive tools and robust performance to support their work, is the name of the fruit of Rao’s labor at Positron Networks. Its goal is to give researchers services that are compatible with the tools they already use, allowing them to hit the ground running and expedite improvements to their scientific research.

One of the most significant impacts that AI tools like Project Robbie will have on the field of scientific research is that they will democratize discovery. With the right AI tools, an individual researcher can have the capacity and power equivalent to that of an entire team. These tools are designed to help ease the burden on researchers, allowing them to focus more of their time and funding on what matters most: conducting the experiments and making the conclusions that will change the world.

The most common use of artificial intelligence tools is to automate repetitive tasks, which in scientific research would take the form of data entry and analysis. These steps could be particularly useful for researchers in the earliest stages of hypothesis formulation, allowing them to be more efficient when they reach the experiment stage.

For example, AI can be used to conduct a document review, helping researchers understand the existing research in their area and identify any knowledge gaps that their hypothesis could seek to solve. Researchers could also use AI models to conduct preliminary simulations, allowing them to optimize their experimental design before reaching the point of conducting costly experiments.

Artificial intelligence technology like Project Robbie also has the potential to enable greater collaboration between researchers. After all, collaboration is a crucial part of the research process — the majority of scientific research is peer reviewed — so ensuring researchers are on an equal playing field with their peers benefits everyone involved in the discovery process and allows researchers to make better quality discoveries.

One of the most powerful applications of Project Robbie’s technology, though, is its ability to simulate conditions that cannot be replicated cheaply or safely in a laboratory environment. Scientists have used AI models to replicate everything from outer space to the conditions of Earth in its earliest days of existence. These conditions would not be practical for most scientific researchers to explore, yet initiatives like Project Robbie are broadening the horizons on which innovative scientific minds can experiment.

However, this democratizing effect is, unfortunately, only the ideal of artificial intelligence — not how access often manifests itself in the real world. The truth of AI is that, like so many new technologies in innovation, access is primarily controlled by the few, not the many. The most powerful AI tools are owned by private companies that use them to serve their private interests, not to serve the greater good. Public research institutions often do not have access to these tools, at least not at the same level as their private counterparts.

Ultimately, the main factor causing this division in access is the cost and complexity of running advanced AI models, as using this technology requires extensive (and expensive) computing power. Frankly, many public institutions lack the budget to support the level of servers necessary to run the most powerful AI models. Because of this, researchers at these institutions are limited to more generalized, publicly available resources, while private institutions have the advantage of remaining on the cutting edge.

This disparity is why a profound need exists for more public private partnerships like Project Robbie. These initiatives help level the playing field by combining the power of the resources of private entities with the innovative thinking of researchers in public research institutions. Project Robbie takes advantage of infrastructure and servers operated by the Mass Open Cloud Alliance — a partnership between Boston University and Harvard University — to allow researchers at public institutions to access high-powered computing resources from consumer-level devices (such as laptops) through the cloud.

Partnerships like this make tools for scientific discovery more accessible and efficient than ever before. Although taking advantage of the power of artificial intelligence would typically require a researcher to invest tens or even hundreds of thousands of dollars into technology like servers and GPUs, tools like Project Robbie can provide them access to these resources without having to devote significant amounts of their precious funding to IT.

Indeed, considering the limited and competitive funding in the public research sector, researchers understandably want to use as much of their resources as possible in their experiments. With the help of tools like Project Robbie, researchers can divert their resources from expenses like IT management to genuine scientific discovery, allowing them to conduct experiments that could change the world or save lives.

Another benefit of the Project Robbie system is that the platform does not require researchers to have advanced software knowledge or coding experience. Researchers’ jobs are already challenging enough, and the goal of AI in scientific research is to make discovery easier, not harder. Adding another complicated tool they must then learn how to use negates the purpose of streamlining their process and trying to reduce their workload.

However, Project Robbie comprises a web application, a series of command lines, and Jupyter Notebook tools, allowing researchers to dispatch work from their local computers to advanced and highly capable scientific computing resources serviced by the Mass Open Cloud. In other words, even someone with a relatively underpowered laptop can take advantage of some of the most intricate and computationally intensive machine learning models.

With the help of these tools, scientific researchers can expect to significantly accelerate the rate of discovery for their experiments. According to estimates by the Positron Networks team, researchers can expect to conduct as many as five times the number of experiments when using Positron Network’s technology as they traditionally would. By reducing the time it takes to conduct experiments from days to minutes, Positron Networks empowers researchers to make more discoveries.

Indeed, the motivation behind Project Robbie is to return the focus of innovation in scientific research back to the needs of the people. In a world where too many powerful computing resources are used exclusively to serve the needs of corporate interests and make money, those like Sid Rao and Positron Networks, whose efforts are focused on improving access for public institutions, are poised to make a legitimate contribution not just to the scientific community but also to the world as a whole.

At the time of this article being written, Project Robbie is in a private beta with the University of Illinois at Urbana Champaign, and approximately 50 other labs, scientific institutions, and government research agencies have also gotten to see the power of Positron Networks’s technology in action. For institutions looking to join the revolution and become involved in future stages of Project Robbie’s beta, the best way to contact Positron Networks is through the company’s website or by emailing beta@positronnetworks.com. By helping Positron Networks explore the intersection of AI and scientific research, you can play a part in helping scientific researchers make world changing discoveries.