Today, InVivo AI becomes Valence, the next step in our journey to unlock the true potential of deep learning in drug design
After spending our PhDs building machine learning tools for big pharma, it was clear that traditional deep learning methods are woefully under-equipped for the challenges inherent to drug design.
Existing approaches—data hungry with limited ability to move beyond the distribution they were trained on—are at odds with the data constraints of a typical discovery campaign and the enormity of molecular search space.
InVivo AI—the company we’ve been quietly building out of Mila these past two years—was born on the understanding that new approaches custom-built for the problem space, combined with new enabling infrastructure to make these technologies more broadly accessible, are needed to unlock the true potential of deep learning in drug design.
Rethinking drug design to design better drugs
For anyone that’s known us as InVivo AI, you know of our commitment to empowering our partners with the latest advances in AI-enabled drug design. You know of our dedication to research, ensuring we have the best tools available for the job, and you know of the emphasis we place on seamlessly integrating into drug discovery workflows, ensuring our technologies are accessible to R&D organizations of all sizes.
But in the midst of a global pandemic, our mission took on a new meaning, and is more important now than ever. Although technology platforms like Moderna’s proved essential against COVID-19, enabling the design of a vaccine candidate in a matter of days, no similar platforms exist for therapeutics.
While there’s a long road ahead, this is why we’re building Valence.
Introducing Valence, the new standard in AI-enabled drug design
As we’ve evolved as a company—expanding our platform for molecular design, partnering with some of the industry’s leading chemistry groups, and growing our team with exceptional interdisciplinary talent—so should our brand.
In chemistry, valency represents the combining power of an element. Valence therefore reflects the power of combining our expertise in AI-enabled drug design with our partners’ expertise in all aspects of drug discovery and development. Together, we're stronger than the sum of our parts.
Beyond the new name, our focus at Valence remains the same:
1. Solving fundamental technical challenges through research
Our team is working tirelessly on the challenges we regularly encounter in our discovery programs: how to improve predictions from small and noisy datasets? How to expand the chemical space in which model predictions are reliable? How to account for the dynamic and flexible nature of molecules? How to guarantee that AI-designed molecules are of high medicinal chemistry quality and readily synthesizable?
2. Supporting our machine learning technologies with industry-leading infrastructure
Valence is about more than just algorithms. Our engineering team is building enabling infrastructure—from APIs and databases, to no-code machine learning workflows for biologists and chemists, to intuitive interfaces that enable human-in-the-loop active learning—to ensure seamless integration into R&D organizations of all sizes.
What’s next for Valence?
We’re on a mission to level the playing field for all drug discovery scientists, ensuring every drug discovery team—not just those in the largest pharmas—has access to the latest advances in AI-enabled drug design.
But for now, welcome to Valence—we’re just getting started.