survival-rl 0.0.1__tar.gz

survival_rl-0.0.1/.gitignore

@@ -0,0 +1,220 @@
+*.sh
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#   Usually these files are written by a python script from a template
+#   before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+# Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+# uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+# poetry.lock
+# poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+# pdm.lock
+# pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+# pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# Redis
+*.rdb
+*.aof
+*.pid
+
+# RabbitMQ
+mnesia/
+rabbitmq/
+rabbitmq-data/
+
+# ActiveMQ
+activemq-data/
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#   JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#   be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#   and can be added to the global gitignore or merged into this file.  For a more nuclear
+#   option (not recommended) you can uncomment the following to ignore the entire idea folder.
+# .idea/
+
+# Abstra
+#   Abstra is an AI-powered process automation framework.
+#   Ignore directories containing user credentials, local state, and settings.
+#   Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#   Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#   that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#   and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#   you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+# Temporary file for partial code execution
+tempCodeRunnerFile.py
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Streamlit
+.streamlit/secrets.toml

survival_rl-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Phil Wang
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

survival_rl-0.0.1/PKG-INFO

@@ -0,0 +1,72 @@
+Metadata-Version: 2.5
+Name: survival-rl
+Version: 0.0.1
+Summary: Explorations into Survival Reinforcement Learning
+Project-URL: Homepage, https://pypi.org/project/survival-rl/
+Project-URL: Repository, https://github.com/lucidrains/survival-rl
+Author-email: Phil Wang <lucidrains@gmail.com>
+License: MIT License
+        
+        Copyright (c) 2026 Phil Wang
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+License-File: LICENSE
+Keywords: artificial intelligence,deep learning,reinforcement learning,survival analysis
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Requires-Dist: einops>=0.8.2
+Requires-Dist: einx>=0.4.3
+Requires-Dist: torch-einops-utils>=0.1.31
+Requires-Dist: torch>=2.5
+Requires-Dist: x-mlps-pytorch>=0.6.4
+Provides-Extra: examples
+Provides-Extra: test
+Requires-Dist: pytest; extra == 'test'
+Description-Content-Type: text/markdown
+
+## Survival RL (wip)
+
+Explorations into an application of survival analysis to RL, proposed by Tiofack et al. earlier this year
+
+## Citations
+
+```bibtex
+@inproceedings{Tiofack2026SVL,
+    title     = {SVL: Goal-Conditioned Reinforcement Learning as Survival Learning},
+    author    = {Franki Nguimatsia Tiofack and Fabian Schramm and Th{\'e}otime Le Hellard and Justin Carpentier},
+    booktitle = {International Conference on Machine Learning (ICML)},
+    year      = {2026},
+    url       = {https://arxiv.org/abs/2604.17551}
+}
+```
+
+```bibtex
+@article{Tiofack2026SRL,
+    title   = {Survival Reinforcement Learning: Toward Scalable Self-Supervised RL},
+    author  = {Franki Nguimatsia-Tiofack and Fabian Schramm and Th{\'e}otime Le Hellard and Justin Carpentier},
+    journal = {arXiv preprint arXiv:2605.31273},
+    year    = {2026},
+    note    = {19th European Workshop on Reinforcement Learning (EWRL), Oral},
+    url     = {https://arxiv.org/abs/2605.31273}
+}
+```

survival_rl-0.0.1/README.md

@@ -0,0 +1,26 @@
+## Survival RL (wip)
+
+Explorations into an application of survival analysis to RL, proposed by Tiofack et al. earlier this year
+
+## Citations
+
+```bibtex
+@inproceedings{Tiofack2026SVL,
+    title     = {SVL: Goal-Conditioned Reinforcement Learning as Survival Learning},
+    author    = {Franki Nguimatsia Tiofack and Fabian Schramm and Th{\'e}otime Le Hellard and Justin Carpentier},
+    booktitle = {International Conference on Machine Learning (ICML)},
+    year      = {2026},
+    url       = {https://arxiv.org/abs/2604.17551}
+}
+```
+
+```bibtex
+@article{Tiofack2026SRL,
+    title   = {Survival Reinforcement Learning: Toward Scalable Self-Supervised RL},
+    author  = {Franki Nguimatsia-Tiofack and Fabian Schramm and Th{\'e}otime Le Hellard and Justin Carpentier},
+    journal = {arXiv preprint arXiv:2605.31273},
+    year    = {2026},
+    note    = {19th European Workshop on Reinforcement Learning (EWRL), Oral},
+    url     = {https://arxiv.org/abs/2605.31273}
+}
+```

survival_rl-0.0.1/pyproject.toml

@@ -0,0 +1,69 @@
+[project]
+name = "survival-rl"
+version = "0.0.1"
+description = "Explorations into Survival Reinforcement Learning"
+authors = [
+    { name = "Phil Wang", email = "lucidrains@gmail.com" }
+]
+readme = "README.md"
+requires-python = ">= 3.10"
+license = { file = "LICENSE" }
+keywords = [
+    'artificial intelligence',
+    'deep learning',
+    'reinforcement learning',
+    'survival analysis'
+]
+
+classifiers=[
+    'Development Status :: 4 - Beta',
+    'Intended Audience :: Developers',
+    'Topic :: Scientific/Engineering :: Artificial Intelligence',
+    'License :: OSI Approved :: MIT License',
+    'Programming Language :: Python :: 3.10',
+]
+
+dependencies = [
+    "einx>=0.4.3",
+    "einops>=0.8.2",
+    "torch>=2.5",
+    "x-mlps-pytorch>=0.6.4",
+    "torch-einops-utils>=0.1.31"
+]
+
+[project.urls]
+Homepage = "https://pypi.org/project/survival-rl/"
+Repository = "https://github.com/lucidrains/survival-rl"
+
+[project.optional-dependencies]
+examples = []
+test = [
+    "pytest"
+]
+
+[tool.pytest.ini_options]
+pythonpath = [
+    "."
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.rye]
+managed = true
+dev-dependencies = []
+
+[tool.hatch.metadata]
+allow-direct-references = true
+
+[tool.hatch.build.targets.wheel]
+packages = ["survival_rl"]
+
+[tool.hatch.build.targets.sdist]
+include = [
+    "/survival_rl",
+    "/pyproject.toml",
+    "/README.md",
+    "/LICENSE"
+]

survival_rl-0.0.1/survival_rl/__init__.py

@@ -0,0 +1,7 @@
+from survival_rl.survival_rl import (
+    HazardCritic
+)
+
+__all__ = [
+    'HazardCritic'
+]

survival_rl-0.0.1/survival_rl/survival_rl.py

@@ -0,0 +1,135 @@
+from __future__ import annotations
+from math import log
+
+import torch
+from torch.nn.functional import (
+    logsigmoid,
+    binary_cross_entropy_with_logits
+)
+from torch.nn import Module, Linear
+
+import einx
+
+from torch_einops_utils import (
+    lens_to_mask,
+    masked_sum
+)
+
+from x_mlps_pytorch import (
+    MLP,
+    ResidualNormedMLP,
+    AttnResidualNormedMLP
+)
+
+# helpers
+
+def exists(v):
+    return v is not None
+
+def default(v, d):
+    return v if exists(v) else d
+
+# classes
+
+class HazardCritic(Module):
+    def __init__(
+        self,
+        *,
+        dim,
+        depth,
+        dim_state,
+        num_actions,
+        pred_time_bins,
+        dim_action = None,
+        dim_goal = None,
+        attn_residual = True,
+        discount_factor = 0.99
+    ):
+        super().__init__()
+
+        mlp_klass = AttnResidualNormedMLP if attn_residual else ResidualNormedMLP
+
+        dim_goal = default(dim_goal, dim_state)
+        dim_action = default(dim_action, dim // 2)
+
+        self.action_encoder = MLP(num_actions, dim_action, dim_action)
+
+        mlp_kwargs = dict(
+            dim_in = dim_state + dim_goal + dim_action,
+            dim = dim,
+            depth = depth
+        )
+
+        if not attn_residual:
+            mlp_kwargs['residual_every'] = next(e for e in (4, 3, 2, 1) if depth % e == 0)
+
+        self.mlp = mlp_klass(**mlp_kwargs)
+
+        self.pred_time_bins = Linear(dim, pred_time_bins)
+
+        self.num_time_bins = pred_time_bins
+
+        self.register_buffer('times', torch.arange(pred_time_bins))
+
+        # calculate log discount factors upfront
+
+        self.register_buffer('log_discounts', torch.arange(pred_time_bins) * log(discount_factor))
+
+    def extract_policy(
+        self,
+        actor: Module,
+        state,
+        goal,
+    ):
+        actions = actor(state, goal)
+        values = self.forward(state, actions, goal, return_values = True)
+        return -values
+
+    def forward(
+        self,
+        state,
+        actions,
+        goal,
+        reach_goal_index = None, # -1 or >= horizon_cutoff treated as right-censored
+        horizon_cutoff = None,
+        return_values = False
+    ):
+        horizon_cutoff = default(horizon_cutoff, self.num_time_bins)
+
+        # encode actions, embed with deep mlp and predict bins
+
+        encoded_action = self.action_encoder(actions)
+
+        embed = self.mlp((state, encoded_action, goal))
+
+        time_bin_logits = self.pred_time_bins(embed)
+
+        # log survival
+
+        log_one_minus_hazard = logsigmoid(-time_bin_logits)
+
+        log_survival = log_one_minus_hazard.cumsum(dim = -1)
+
+        # returning values, as discounted sum of survival probabilities
+
+        if return_values:
+            return -(log_survival + self.log_discounts).exp().sum(dim = -1)
+
+        # if no reach goal index given, return the log survival function
+
+        if not exists(reach_goal_index):
+            return log_survival
+
+        # bce loss on logits
+
+        reached_goal = reach_goal_index < horizon_cutoff
+        reached_goal_bin = einx.equal('b, t -> b t', reach_goal_index, self.times) & reached_goal[..., None]
+
+        losses = binary_cross_entropy_with_logits(time_bin_logits, reached_goal_bin.to(time_bin_logits.dtype), reduction = 'none')
+
+        # nll loss - time after cutoff is masked out
+
+        lens = torch.where(reach_goal_index >= 0, reach_goal_index + 1, horizon_cutoff).clamp(max = horizon_cutoff)
+        mask = lens_to_mask(lens, self.num_time_bins)
+
+        return masked_sum(losses, mask, dim = -1).mean()