SMC2 #1775
SMC2 #1775
Conversation
tensorflow_probability/python/experimental/mcmc/sequential_monte_carlo_kernel.py
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter_test.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
| @@ -670,6 +1122,8 @@ def _compute_observation_log_weights(step, | |||
| observation = tf.nest.map_structure( | |||
| lambda x, step=step: tf.gather(x, observation_idx), observations) | |||
|
|
|||
| if particles_dim == 1: | |||
| observation = tf.expand_dims(observation, axis=0) | |||
There was a problem hiding this comment.
I don't think this added code should be necessary. The lines right below this should already handle any necessary expand_dims for particles_dim.
There was a problem hiding this comment.
I think it is necessary due to the dimensions mismatch
There was a problem hiding this comment.
With these two lines included, ParticleFilterTestFloat32.test_batch_of_filters_particles_dim_1 fails when I run it. If I delete these two lines, then that tests passes.
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
tensorflow_probability/python/experimental/mcmc/particle_filter.py
Outdated
Show resolved
Hide resolved
| @@ -670,6 +1122,8 @@ def _compute_observation_log_weights(step, | |||
| observation = tf.nest.map_structure( | |||
| lambda x, step=step: tf.gather(x, observation_idx), observations) | |||
|
|
|||
| if particles_dim == 1: | |||
| observation = tf.expand_dims(observation, axis=0) | |||
There was a problem hiding this comment.
With these two lines included, ParticleFilterTestFloat32.test_batch_of_filters_particles_dim_1 fails when I run it. If I delete these two lines, then that tests passes.
| inner_initial_state_prior=lambda _, params: | ||
| mvn_diag.MultivariateNormalDiag( | ||
| loc=loc, scale_diag=scale_diag | ||
| ), |
There was a problem hiding this comment.
I think this should be:
inner_initial_state_prior=lambda _, params: mvn_diag.MultivariateNormalDiag(
loc=tf.broadcast_to([0., 0.], params.shape + [2]),
scale_diag=tf.broadcast_to([0.01, 0.01], params.shape + [2])
Because:
params.shapeis just the batch shape (i.e.,[num_outer_particles])- Which will give
inner_inital_state_prior(0, params)the batch_shape of[num_outer_particles]and the correct event_shape of[2]. (And the SMC^2 code is then responsible for drawingnum_inner_particlessamples from this distribution and re-arranging the dimensions to get initial particles of shape[num_outer_particles, num_inner_particles, 2].)
| broadcasted_params = tf.broadcast_to(reshaped_params, | ||
| previous_state.shape) | ||
| reshaped_dist = independent.Independent( | ||
| normal.Normal(previous_state + broadcasted_params + 1, 0.1), |
There was a problem hiding this comment.
params is has shape [num_outer_particles] and previous_state has shape [num_outer_particles, num_inner_particles, 2], so here it should be sufficient to just do:
previous_state + params[..., tf.newaxis, tf.newaxis] + 1
| tf.equal(tf.math.mod(step, tf.constant(2)), tf.constant(0)), | ||
| tf.not_equal(state.extra[0], tf.constant(0)) | ||
| ) | ||
| return tf.cond(cond, lambda: tf.constant(True), |
There was a problem hiding this comment.
It should work to just return cond, instead of using tf.cond(cond, ...).
| self.evaluate( | ||
| tf.reduce_sum(tf.exp(log_weights) * | ||
| particles['position'], axis=2)), | ||
| observed_positions, |
There was a problem hiding this comment.
I would hope this assertion would pass if we changed this line to observed_positions[..., tf.newaxis, tf.newaxis].
But if that didn't work, we could also do: tf.broadcast_to(observation_positions, [num_timesteps] + batch_shape).
Added SMC2 from https://arxiv.org/abs/1101.1528