BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Bayes Reading Group:  Nawaf Bou-Rabee (Rutgers\, Flatiron\, & Whar
 ton)
DTSTART:20251205T160000Z
DTEND:20251205T170000Z
DTSTAMP:20261009T013600Z
UID:indico-event-4110@indico.flatironinstitute.org
DESCRIPTION:Discussion lead: Nawaf Bou-Rabee (Rutgers\, Flatiron\, & Wha
 rton)\nTitle: When the Kinetic Energy Doesn’t Budge: Isokinetic Monte C
 arlo in Theory and Practice\nAbstract: Iso-kinetic HMC is an elegant and 
 somewhat overlooked variant of Hamiltonian Monte Carlo dating back to the 
 foundational work of Martyna\, Minary and Tuckerman in the early 2000s. Un
 like classical HMC\, iso-kinetic HMC evolves under a non-holonomic constra
 int that locks the kinetic energy to a fixed value. In other words\, inste
 ad of letting the instantaneous temperature fluctuate as in standard HMC\,
  the sampler “rides a geodesic” on a constant-kinetic-energy manifold.
  This constraint fundamentally alters the geometry of the dynamics\, and a
 s a result leads to markedly different statistical and numerical behavior.
 \nThe method resurfaced a decade later in the influential work of Fang\, S
 anz-Serna\, and Skeel\, who placed iso-kinetic dynamics within the broader
  framework of compressible generalized HMC. And in the last year\, it has 
 re-entered the spotlight in Bayesian computation\, in recent work by Robni
 k\, Cohn-Gordon\, and Seljak exploring its surprising empirical performanc
 e in high-dimensional inference problems.\nThis journal club talk will rev
 isit iso-kinetic HMC from first principles\, examining why the method is r
 eversible\, how the non-holonomic constraint interacts with ergodicity and
  volume preservation\, and whether the fixed-kinetic-energy geometry provi
 des a form of implicit preconditioning. We will also discuss what is known
  (and what remains mysterious) about its scaling in high dimension\, and h
 ow its tuning behavior compares with standard HMC and NUTS.\nSelected Refe
 rences:\n\n\nMartyna\, Minary & Tuckerman (2003a). Algorithms and novel a
 pplications based on the isokinetic ensemble. I. Biophysical and path inte
 gral molecular dynamics. J. Chem. Phys.\n\n\nMinary\, Martyna & Tuckerman
  (2003b). Algorithms and novel applications based on the isokinetic ensem
 ble. II. Ab initio molecular dynamics. J. Chem. Phys.\n\n\nTuckerman (201
 0). Statistical Mechanics: Theory and Molecular Simulation\, Chapter 4 \
 n\n\nFang\, Sanz-Serna & Skeel (2014). Compressible generalized hybrid Mo
 nte Carlo. J. Chem. Phys. \n\n\nRobnik\, Cohn-Gordon & Seljak (2025). M
 etropolis-adjusted Microcanonical Hamiltonian Monte Carlo. arXiv:2503.0170
 7.\n\nhttps://indico.flatironinstitute.org/event/4110/
LOCATION:3rd Floor Conference Room  (162 Fifth Avenue)
URL:https://indico.flatironinstitute.org/event/4110/
END:VEVENT
END:VCALENDAR