Title: A whole new woRLD of Fisher information Speaker:  Sujay Kazi (Duke University) Date & Time:  June 12, 2026, 2:00pm Where to Attend:  ATL 3100A and Virtual Via Zoom: To be announced
In the field of information geometry, Chentsov's theorem states that Fisher information is the unique (up to normalization) distance metric on the space of probability distributions that is non-increasing under stochastic maps. Naturally extending Chentsov's theorem to quantum states and quantum channels yields a whole family of metrics, which we may call quantum Fisher information (QFI) metrics. However, although these metrics were classified thirty years ago, the full family of QFI metrics is still unfamiliar to most quantum information theorists, and surprisingly little is known about which QFI metrics are useful and where.
In this talk, I will introduce the theory of both classical and quantum Fisher information. I will then highlight the underappreciated significance of a specific QFI metric known as the right logarithmic derivative (RLD) Fisher information.
I will conclude by showing how RLD Fisher information finds novel operational interpretations from two deceptively simple qubit conversion problems. The first is the problem of coherence distillation, i.e., to find the maximum fidelity with which a large number of noisy qubit coherent states can be converted into one pure qubit coherent state while preserving the phase (longitude) information. The second is the problem of linear-rate conversion, i.e., to find the maximum asymptotic rate at which qubits at one purity level can be converted into qubits at another purity level while preserving their direction (latitude and longitude).
*We strongly encourage attendees to use their full name (and if possible, their UMD credentials) to join the zoom session.*