Definition
hy·per·pri·or
/ˈhaɪ.pəˌpraɪ.ə/ n. ハイパープライアー
- 1. stats. A belief spread over the priors one might start from.
- 2. A small, privately funded research group in London. We build environments and graders for learning agents, study complex adaptive systems, and selectively work with firms facing difficult technical decisions and challenges.
Origin From the Greek hyper, over, and the Latin prior, earlier. A prior on a prior, and so a way of writing down how sure one is about how sure one is.
See also prior, posterior, calibration, hyperparameter
prior stats. A probability distribution over an unknown quantity, written down before the data are seen.
posterior stats. The distribution over an unknown quantity after the data are seen, which is the prior updated by the evidence.
calibration stats. The agreement between how confident forecasters say they are and how often their forecasts prove right.
hyperparameter stats. A parameter of a prior distribution, which a hyperprior in turn treats as uncertain.
The question
How sure are you about how sure you are?
Asked of a model, it is the confidence behind each prediction and how often that confidence is earned. Asked of an agent, it is the balance between acting on a belief and stopping to test it. Asked of a human, it is the set of implicit or explicit beliefs that underpin our actions and feelings.
Research agenda
We build environments and graders, and most of our research is on reinforcement learning, world models and program synthesis. We study what emerges when many learners interact and adapt to one another in synthetic and complex adaptive environments.
Research interests subject to change without notice. Beliefs likewise.
Who we are
We are a small, interdisciplinary team of researchers and engineers, drawn from trading firms and AI labs, who like to understand and explore complex systems.
We work across machine learning, statistics, economics and computer science. Our expertise is broad, and exceptionally deep in global markets. We know the plumbing of finance from the inside out, and our expertise in shared state machines is regularly called upon by major financial institutions and technology companies.
HyperPrior Research was founded in London in 2022 and is privately funded, with no outside money. It is led by Patrick Steele, who read computer science at the University of Auckland, holds an MBA from Cambridge and spent more than a decade in systematic and quantitative trading, most recently at DRW.
We work from Shoreditch, a short walk from Bunhill Fields, where Thomas Bayes is buried.
Positions lists three open roles. Please take a look if you are interested in collaborating.
Contact
We would be happy to hear from you, for questions or simply to connect.
We are currently recruiting exceptional collaborators. Please check out our open positions.
