README.md

@@ -65,7 +65,9 @@ pytest tests/test_projections.py
 *Note*: The test suite verifies:
 
 1. All projections run without errors and maintain basic properties (shapes, positive definiteness)
-2. KL bounds are actually (approximately) met for true KL projection (both diagonal and full covariance)
+2. KL bounds are actually (approximately) met for:
+   - KL projection (both diagonal and full covariance)
+   - Wasserstein projection (diagonal covariance only)
 3. Gradients can be computed through all projections:
    - Both through projection operation and trust region loss
    - Gradients have correct shapes and are finite

itpal_jax/frobenius_projection.py

@@ -1,7 +1,6 @@
 import jax.numpy as jnp
 from .base_projection import BaseProjection
 from typing import Dict
-import jax
 
 class FrobeniusProjection(BaseProjection):
     def __init__(self, trust_region_coeff: float = 1.0, mean_bound: float = 0.01,
@@ -65,26 +64,12 @@ class FrobeniusProjection(BaseProjection):
         old_mean, old_cov = q
 
         if self.scale_prec:
-            # Mahalanobis distance for mean
-            diff = mean - old_mean
-            if old_cov.ndim == mean.ndim:  # diagonal case
-                mean_part = jnp.sum(jnp.square(diff / old_cov), axis=-1)
-            else:
-                solved = jax.scipy.linalg.solve_triangular(
-                    old_cov, diff[..., None], lower=True
-                )
-                mean_part = jnp.sum(jnp.square(solved.squeeze(-1)), axis=-1)
+            prec_old = jnp.linalg.inv(old_cov)
+            mean_part = jnp.sum(jnp.matmul(mean - old_mean, prec_old) * (mean - old_mean), axis=-1)
+            cov_part = jnp.sum(prec_old * cov, axis=(-2, -1)) - jnp.log(jnp.linalg.det(jnp.matmul(prec_old, cov))) - mean.shape[-1]
         else:
             mean_part = jnp.sum(jnp.square(mean - old_mean), axis=-1)
-
-        # Frobenius norm for covariance
-        if cov.ndim == mean.ndim:  # diagonal case
-            diff = old_cov - cov
-            cov_part = jnp.sum(jnp.square(diff), axis=-1)
-        else:
-            diff = jnp.matmul(old_cov, jnp.swapaxes(old_cov, -1, -2)) - \
-                   jnp.matmul(cov, jnp.swapaxes(cov, -1, -2))
-            cov_part = jnp.sum(jnp.square(diff), axis=(-2, -1))
+            cov_part = jnp.sum(jnp.square(cov - old_cov), axis=(-2, -1))
 
         return mean_part, cov_part
 
@@ -98,20 +83,22 @@ class FrobeniusProjection(BaseProjection):
 
     def _cov_projection(self, cov: jnp.ndarray, old_cov: jnp.ndarray, 
                        cov_part: jnp.ndarray) -> jnp.ndarray:
-        batch_shape = cov.shape[:-2] if cov.ndim > 2 else cov.shape[:-1]
+        batch_shape = cov.shape[:-2]
         cov_mask = cov_part > self.cov_bound
 
         eta = jnp.ones(batch_shape, dtype=cov.dtype)
         eta = jnp.where(cov_mask, 
-                        jnp.sqrt(cov_part / self.cov_bound) - 1.,
-                        eta)
+                       jnp.sqrt(cov_part / self.cov_bound) - 1.,
+                       eta)
         eta = jnp.maximum(-eta, eta)
 
         if self.full_cov:
-            new_cov = (cov + jnp.einsum('...,...ij->...ij', eta, old_cov)) / \
-                      (1. + eta + 1e-16)[..., None, None]
-            proj_cov = jnp.where(cov_mask[..., None, None], new_cov, cov)
-            return jnp.linalg.cholesky(proj_cov)
+            new_cov = (cov + jnp.einsum('...,...ij->...ij', eta, old_cov)) / (1. + eta + 1e-16)[..., None, None]
         else:
+            # For diagonal case, simple broadcasting
             new_cov = (cov + eta[..., None] * old_cov) / (1. + eta + 1e-16)[..., None]
-            return jnp.where(cov_mask[..., None], jnp.sqrt(new_cov), cov)
+            
+        proj_cov = jnp.where(cov_mask[..., None] if not self.full_cov else cov_mask[..., None, None], 
+                             new_cov, cov)
+
+        return proj_cov

itpal_jax/kl_projection.py

@@ -13,24 +13,6 @@ from .exception_projection import makeExceptionProjection
 
 MAX_EVAL = 1000
 
-# Cache for projection operators
-_diag_proj_op = None
-_full_proj_op = None
-
-def _get_diag_proj_op(batch_shape, dim):
-    global _diag_proj_op
-    if _diag_proj_op is None:
-        _diag_proj_op = cpp_projection.BatchedDiagCovOnlyProjection(
-            batch_shape, dim, max_eval=MAX_EVAL)
-    return _diag_proj_op
-
-def _get_full_proj_op(batch_shape, dim):
-    global _full_proj_op
-    if _full_proj_op is None:
-        _full_proj_op = cpp_projection.BatchedCovOnlyProjection(
-            batch_shape, dim, max_eval=MAX_EVAL)
-    return _full_proj_op
-
 class KLProjection(BaseProjection):
     """KL divergence-based projection for Gaussian policies.
     
@@ -184,49 +166,6 @@ class KLProjection(BaseProjection):
             if key not in policy_params or key not in old_policy_params:
                 raise KeyError(f"Missing required key '{key}' in policy parameters")
 
-@partial(jax.custom_vjp, nondiff_argnums=(2,))
-def project_diag_covariance(cov, old_cov, eps_cov):
-    """JAX wrapper for C++ diagonal covariance projection"""
-    batch_shape = cov.shape[0]
-    dim = cov.shape[-1]
-    
-    cov_np = np.asarray(cov)
-    old_cov_np = np.asarray(old_cov)
-    eps = eps_cov * np.ones(batch_shape, dtype=old_cov_np.dtype)
-    
-    p_op = _get_diag_proj_op(batch_shape, dim)
-    
-    try:
-        proj_cov = p_op.forward(eps, old_cov_np, cov_np)
-    except:
-        proj_cov = cov_np  # Return input on failure
-        
-    return jnp.array(proj_cov)
-
-def project_diag_covariance_fwd(cov, old_cov, eps_cov):
-    y = project_diag_covariance(cov, old_cov, eps_cov)
-    return y, (cov, old_cov)
-
-def project_diag_covariance_bwd(eps_cov, res, g):
-    cov, old_cov = res
-    
-    # Convert to numpy for C++ backward pass
-    g_np = np.asarray(g)
-    batch_shape = g_np.shape[0]
-    dim = g_np.shape[-1]
-    
-    # Get C++ projection operator
-    p_op = _get_diag_proj_op(batch_shape, dim)
-    
-    # Run C++ backward pass
-    grad_cov = p_op.backward(g_np)
-    
-    # Convert back to JAX array
-    return jnp.array(grad_cov), None
-
-# Register VJP rule for diagonal covariance projection
-project_diag_covariance.defvjp(project_diag_covariance_fwd, project_diag_covariance_bwd)
-
 @partial(jax.custom_vjp, nondiff_argnums=(3,))
 def project_full_covariance(cov, chol, old_chol, eps_cov):
     """JAX wrapper for C++ full covariance projection"""
@@ -240,7 +179,7 @@ def project_full_covariance(cov, chol, old_chol, eps_cov):
         eps = eps_cov * np.ones(batch_shape)
 
         # Create C++ projection operator directly
-        p_op = _get_full_proj_op(batch_shape, dim)
+        p_op = cpp_projection.BatchedCovOnlyProjection(batch_shape, dim, max_eval=MAX_EVAL)
         
         # Run C++ projection
         proj_cov = p_op.forward(eps, old_chol_np, chol_np, cov_np)
@@ -264,7 +203,7 @@ def project_full_covariance_bwd(eps_cov, res, g):
     dim = g_np.shape[-1]
     
     # Get C++ projection operator
-    p_op = _get_full_proj_op(batch_shape, dim)
+    p_op = cpp_projection.BatchedCovOnlyProjection(batch_shape, dim, max_eval=MAX_EVAL)
     
     # Run C++ backward pass
     grad_cov = p_op.backward(g_np)
@@ -275,5 +214,48 @@ def project_full_covariance_bwd(eps_cov, res, g):
 # Register VJP rule for full covariance projection
 project_full_covariance.defvjp(project_full_covariance_fwd, project_full_covariance_bwd)
 
+@partial(jax.custom_vjp, nondiff_argnums=(2,))
+def project_diag_covariance(cov, old_cov, eps_cov):
+    """JAX wrapper for C++ diagonal covariance projection"""
+    # Convert JAX arrays to numpy for C++ function
+    cov_np = np.asarray(cov)
+    old_cov_np = np.asarray(old_cov)
+    batch_shape = cov_np.shape[0]
+    dim = cov_np.shape[-1]
+    eps = eps_cov * np.ones(batch_shape)
+
+    # Create C++ projection operator directly
+    p_op = cpp_projection.BatchedDiagCovOnlyProjection(batch_shape, dim, max_eval=MAX_EVAL)
+    
+    # Run C++ projection
+    proj_cov = p_op.forward(eps, old_cov_np, cov_np)
+    
+    # Convert back to JAX array
+    return jnp.array(proj_cov)
+
+def project_diag_covariance_fwd(cov, old_cov, eps_cov):
+    y = project_diag_covariance(cov, old_cov, eps_cov)
+    return y, (cov, old_cov)
+
+def project_diag_covariance_bwd(eps_cov, res, g):
+    cov, old_cov = res
+    
+    # Convert to numpy for C++ backward pass
+    g_np = np.asarray(g)
+    batch_shape = g_np.shape[0]
+    dim = g_np.shape[-1]
+    
+    # Get C++ projection operator
+    p_op = cpp_projection.BatchedDiagCovOnlyProjection(batch_shape, dim, max_eval=MAX_EVAL)
+    
+    # Run C++ backward pass
+    grad_cov = p_op.backward(g_np)
+    
+    # Convert back to JAX array
+    return jnp.array(grad_cov), None
+
+# Register VJP rule for diagonal covariance projection
+project_diag_covariance.defvjp(project_diag_covariance_fwd, project_diag_covariance_bwd)
+
 if not cpp_projection_available:
     KLProjection = makeExceptionProjection("ITPAL (C++ library) is not available. Please install the C++ library to use this projection.")

itpal_jax/wasserstein_projection.py

@@ -1,7 +1,6 @@
 import jax.numpy as jnp
 from .base_projection import BaseProjection
 from typing import Dict, Tuple
-import jax
 
 def scale_tril_to_sqrt(scale_tril: jnp.ndarray) -> jnp.ndarray:
     """
@@ -23,8 +22,7 @@ class WassersteinProjection(BaseProjection):
 
     def project(self, policy_params: Dict[str, jnp.ndarray], 
                 old_policy_params: Dict[str, jnp.ndarray]) -> Dict[str, jnp.ndarray]:
-        if self.full_cov:
-            print("Warning: Wasserstein projection with full covariance is wip, we recommend using diagonal covariance instead.")
+        assert not self.full_cov, "Wasserstein projection only supports diagonal covariance"
         
         mean = policy_params["loc"]  # shape: (batch_size, dim)
         old_mean = old_policy_params["loc"]
@@ -75,47 +73,27 @@ class WassersteinProjection(BaseProjection):
 
     def _scale_projection(self, scale: jnp.ndarray, old_scale: jnp.ndarray, 
                          scale_part: jnp.ndarray) -> jnp.ndarray:
-        """Project scale parameters using multiplicative update.
+        """Project scale parameters (standard deviations for diagonal case)"""
+        diff = scale - old_scale
+        norm = jnp.sqrt(scale_part)
         
-        Args:
-            scale: Current scale/sqrt of covariance
-            old_scale: Previous scale/sqrt of covariance
-            scale_part: W2 distance between scales
+        if scale.ndim == 2:  # Batched scale
+            norm = norm[..., None]
         
-        Returns:
-            Projected scale that satisfies the trust region constraint
-        """
-        # Check if projection needed
-        cov_mask = scale_part > self.cov_bound
-        
-        # Compute eta (multiplier for the update)
-        batch_shape = scale.shape[:-2] if scale.ndim > 2 else scale.shape[:-1]
-        eta = jnp.ones(batch_shape, dtype=scale.dtype)
-        eta = jnp.where(cov_mask,
-                        jnp.sqrt(scale_part / self.cov_bound) - 1.,
-                        eta)
-        eta = jnp.maximum(-eta, eta)
-        
-        # Multiplicative update with correct broadcasting
-        if scale.ndim > 2:  # Full covariance case
-            new_scale = (scale + jnp.einsum('...,...ij->...ij', eta, old_scale)) / \
-                       (1. + eta + 1e-16)[..., None, None]
-            mask = cov_mask[..., None, None]
-        else:  # Diagonal case
-            new_scale = (scale + eta[..., None] * old_scale) / \
-                       (1. + eta + 1e-16)[..., None]
-            mask = cov_mask[..., None]
-        
-        return jnp.where(mask, new_scale, scale)
+        return jnp.where(norm > self.cov_bound,
+                        old_scale + diff * self.cov_bound / norm,
+                        scale)
 
     def _gaussian_wasserstein(self, p, q):
         mean, scale = p
         mean_other, scale_other = q
         
-        # Euclidean distance for mean part (we're in diagonal case)
-        mean_part = jnp.sum(jnp.square(mean - mean_other), axis=-1)
+        # Keep batch dimension by only summing over feature dimension
+        mean_part = jnp.sum(jnp.square(mean - mean_other), axis=-1)  # -> (batch_size,)
         
-        # Standard W2 objective for covariance (diagonal case)
-        cov_part = jnp.sum(scale_other**2 + scale**2 - 2 * scale_other * scale, axis=-1)
+        if scale.ndim == mean.ndim:  # Batched scale
+            cov_part = jnp.sum(scale_other**2 + scale**2 - 2 * scale_other * scale, axis=-1)
+        else:  # Non-contextual scale (single scale for all batches)
+            cov_part = jnp.sum(scale_other**2 + scale**2 - 2 * scale_other * scale)
         
         return mean_part, cov_part

tests/test_projections.py

@@ -94,8 +94,8 @@ def test_diagonal_projection(ProjectionClass, needs_cpp, gaussian_params):
     assert jnp.all(jnp.isfinite(proj_params["scale"]))
     assert jnp.all(proj_params["scale"] > 0)
     
-    # Only check KL bounds for KL projection
-    if ProjectionClass in [KLProjection]:
+    # Only check KL bounds for KL projection (and W2, which should approx hold as well)
+    if ProjectionClass in [KLProjection, WassersteinProjection]:
         kl = compute_gaussian_kl(proj_params, gaussian_params["old_params"])
         max_kl = (mean_bound + cov_bound) * 1.1  # Allow 10% margin