Add nslc_cma_imp to sphere example (#482)

## Description

<!-- Provide a brief description of the PR's purpose here. -->

An experimental algorithm that combines a ProximityArchive with local
competition with an EvolutionStrategyEmitter.

## Status

- [x] I have read the guidelines in

[CONTRIBUTING.md](https://github.com/icaros-usc/pyribs/blob/master/CONTRIBUTING.md)
- [x] I have formatted my code using `yapf`
- [x] I have tested my code by running `pytest`
- [x] I have linted my code with `pylint`
- [x] I have added a one-line description of my change to the changelog
in
      `HISTORY.md`
- [x] This PR is ready to go

HISTORY.md

@@ -19,7 +19,7 @@
 
 #### Documentation
 
-- Add novelty search with CMA-ES to sphere example ({pr}`478`)
+- Add novelty search with CMA-ES to sphere example ({pr}`478`, {pr}`482`)
 
 #### Improvements
 

examples/sphere.py

@@ -47,6 +47,9 @@
   GradientArborescenceEmitter using ImprovementRanker.
 - `ns_cma`: Novelty Search with CMA-ES; implemented using a ProximityArchive
   with EvolutionStrategyEmitter. Results are stored in a passive GridArchive.
+- `nslc_cma_imp`: EvolutionStrategyEmitter with a ProximityArchive with local
+  competition turned on. Thus, the archive returns two-stage improvement
+  information that is fed to the EvolutionStrategyEmitter just like in CMA-ME.
 
 The parameters for each algorithm are stored in CONFIG. The parameters
 reproduce the experiments presented in the paper in which each algorithm is
@@ -608,6 +611,8 @@
         "archive": {
             "class": ProximityArchive,
             "kwargs": {
+                # Hyperparameters from Density Descent paper:
+                # https://arxiv.org/abs/2312.11331
                 "k_neighbors": 15,
                 "novelty_threshold": 0.037 * 512,
             }
@@ -627,6 +632,37 @@
             "kwargs": {}
         }
     },
+    "nslc_cma_imp": {
+        "dim": 100,
+        "iters": 5000,
+        "archive_dims": (100, 100),
+        "use_result_archive": True,
+        "is_dqd": False,
+        "batch_size": 36,
+        "archive": {
+            "class": ProximityArchive,
+            "kwargs": {
+                "k_neighbors": 15,
+                # Note: This is untuned.
+                "novelty_threshold": 0.037 * 100,
+                "local_competition": True,
+            }
+        },
+        "emitters": [{
+            "class": EvolutionStrategyEmitter,
+            "kwargs": {
+                "sigma0": 0.5,
+                "ranker": "2imp",
+                "selection_rule": "filter",
+                "restart_rule": "no_improvement"
+            },
+            "num_emitters": 15
+        }],
+        "scheduler": {
+            "class": Scheduler,
+            "kwargs": {}
+        }
+    },
 }
 
 

tests/examples.sh

@@ -49,6 +49,7 @@ python examples/sphere.py cma_mega_adam --dim 20 --itrs 10 --outdir "${SPHERE_OU
 python examples/sphere.py cma_mae --dim 20 --itrs 10 --learning_rate 0.01 --outdir "${SPHERE_OUTPUT}"
 python examples/sphere.py cma_maega --dim 20 --itrs 10 --learning_rate 0.01 --outdir "${SPHERE_OUTPUT}"
 python examples/sphere.py ns_cma --dim 20 --itrs 10 --outdir "${SPHERE_OUTPUT}"
+python examples/sphere.py nslc_cma_imp --dim 20 --itrs 10 --outdir "${SPHERE_OUTPUT}"
 
 # lunar_lander.py
 install_deps examples/lunar_lander.py