Uncertainty and Adaptive Systems

file

Visible to the public Formal Verification of C Programs with Floating-Point Computations: Certified Error Bounds for Signal Processing

ABSTRACT
In this technical talk, I will present our results related to the certification of floating-point error bounds in C implementations of signal processing algorithms. In particular, this relates to the topic of reasoning about the uncertainty caused by the noise arising from floating-point rounding errors and approximate computations in C programs.

file

Visible to the public Programming Uncertain <T>hings

ABSTRACT
Innovation flourishes with good abstractions. For instance, codification of the IEEE Floating Point standard in 1985 was critical to the subsequent success of scientific computing. Programming languages currently lack appropriate abstractions for uncertain data. Applications already use estimates from sensors, machine learning, big data, humans, and approximate algorithms, but most programming languages do not help developers address correctness, programmability, and optimization problems due to estimates.

file

Visible to the public Safety-Constrained Reinforcement Learning for MDPs

ABSTRACT
Many formal system models are inherently stochastic, consider for instance randomized distributed algorithms (where randomization breaks the symmetry between processes), security (e.g., key generation at encryption), systems biology (where species randomly react depending on their concentration), or embedded systems (interacting with unknown and varying environments).

file

Visible to the public Wigmore: A Constrain-Based Language for Reasoning About Evidence and Uncertainty

ABSTRACT
Historically, probability theory has proven to be very useful in dealing with uncertainty, especially when it can be quantified by statistical means. This is why the literature on the subject often distinguishes between risk, which applies to situations where uncertainty can be captured by a probability, and ambiguity, when there exists uncertainty without meaningful probabilities.