Josh Fass
Josh Fass
The stored pseudoeigenvalues in SparseTICA are different from those computed for GMRQ scores, so `stica.score_ != stica.score(X_train)`. I think this is because `stica.eigenvalues_` [are computed using deflated versions of the...
Hi! If I ask a `MarkovStateModel` to score discrete trajectories that never enter its largest strongly connected component, MSMBuilder raises an error instead of returning `0`. ``` python import numpy...
## Description Intention: Efficient implementations of force factories for splitting solvent-solvent interactions from other interactions, for use in multiple-timestepping schemes. ## Status Incomplete. Currently the splitter using exceptions and the...
- [x] Add a "brute-force" method for resampling the coordinates of dummy atoms, conditioned on the coordinates of anchor atoms - [x] Add a basic test (asserting that only the...
Introduces a new submodule `timemachine.maps` - [x] Add and test utilities for computing mapped works or mapped `u_kn` matrix, incorporating invertible `map_fxns` - [x] Add and test utilities for mapping...
* Add `HMCMove` * Make misc. small improvements to `moves.py` * Test `HMCMove` * Test stability etc. on vacuum ligand (using Jax potentials) * Test stability etc. on solvated DHFR...
Addresses https://github.com/proteneer/timemachine/pull/776#discussion_r932660565 by simplifying `.multiple_steps` to return both positions and velocities "on-step", so that it is now consistent with `.step`. This may be helpful if we wish to use the...
* Disables passing `symmetrize=True` to the AM1 routine (since this appears to average over stereo-oblivious atom classes, and occasionally introduces very small asymmetries in cases like benzene) * Symmetrizes by...
Migrating from this thread: https://github.com/proteneer/timemachine/pull/712#discussion_r847399620 Problem: [`validate_lambda_schedule`](https://github.com/proteneer/timemachine/blob/56d9933d208b07bfa0524376086f27a26a468f82/timemachine/fe/lambda_schedule.py#L32-L37) fails assertion that endpoints exactly == 0.0 or 1.0, for some lambda schedules generated with num_windows < 5 . Extent: Locally it looks...
**Context** We would like to define a Jax function referencing the potential energy functions of many systems at once. (E.g. to define a differentiable loss function in terms of a...