Aditya Grover is co-founder and CTO of Inception and an assistant professor of computer science at UCLA. His research connects generative models and reinforcement learning with applications in science; earlier work includes node2vec for learning representations of networks.[1][3]
Start with node2vec →
Education
Links
Website
X LinkedIn Google Scholar Assistant professor, Computer Science · University of California, Los Angeles
Co-founder and chief technology officer · Inception
Career
Grover studied at IIT Delhi and Stanford, then conducted postdoctoral research at UC Berkeley in 2020–2021. He joined UCLA in 2021 and leads its Machine Intelligence group. His current website also describes a year as a research scientist at Meta FAIR and his role developing diffusion language models at Inception.[2][1]
Contributions
node2vec, coauthored with Jure Leskovec, turns a graph into vectors that can be used by other learning methods. It samples biased random walks to create node neighborhoods, then learns representations that preserve those neighborhoods. Adjusting the walk balances exploration of nearby connections with more distant structural relationships. The paper evaluates classification and link-prediction tasks.[3]
Ideas
Grover’s research group studies how systems can reason and interact with limited supervision. His projects connect generative modeling to physical-world understanding and scientific applications, including weather forecasting. That combination gives practitioners a way to follow model research beyond conversational benchmarks.[1]
Tools, writing & talks
Writing, research & talks
Start here
node2vecarxiv.orgBiased random walks produce graph neighborhoods for learning node representations.
Start with the random-walk diagram and the two controls on exploration. Basic graph and vector vocabulary helps; compare the downstream tasks before assuming one representation suits every network.
Co-author[3]