mabwiser
MABWiser: Parallelizable Contextual Multi-Armed Bandits Library
Decision gist · record as of 2026-08-14
Yes, if you are prototyping or researching multi-armed bandit algorithms in Python. The library is stable, well-documented, has no known vulnerabilities, and low install friction. However, dormant maintenance means no active bug fixes or feature development—suitable for research and learning, less ideal if you need ongoing support or expect breaking changes in your dependencies.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Low friction: pure Python wheel with standard scientific dependencies (joblib, numpy, pandas, scikit-learn, scipy, seaborn).
- Last commit was 2024-09-05, but maintenance is dormant—no active development expected.
License · maintenance · safety
permissive license (permissive) — Apache License 2.0 is permissive; you can use, modify, and distribute MABWiser freely in commercial and open-source projects with minimal restrictions.
last release 2024-08-30 (714 days) · last repo commit 2024-09-05 · 291 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 99,270 downloads/mo, #13,031 on PyPI
Alternatives
Verify before relying
from mabwiser.mab import MAB, LearningPolicy
arms = ['Arm1', 'Arm2']
mab = MAB(arms, LearningPolicy.UCB1(alpha=1.25))
mab.fit(['Arm1', 'Arm1', 'Arm2', 'Arm1'], [20, 17, 25, 9])
mab.predict()- Whether the library actively supports Python versions beyond 3.8 despite classifier only listing 3.8
- Current test coverage and whether heavy testing mentioned in description is still maintained
- Whether parallelization features are production-ready or primarily research-oriented
What it is and what it does
MABWiser is a research library for prototyping multi-armed bandit algorithms—a class of decision-making strategies that balance exploring new options against exploiting known good ones. It provides multiple learning policies (Epsilon Greedy, LinUCB, Thompson Sampling, UCB1, and others) and neighborhood policies (Clusters, K-Nearest, LSH Nearest, Radius, TreeBandit) to handle both simple arm-selection problems and contextual variants where decisions depend on feature data. The library follows scikit-learn conventions, includes a simulation utility for policy comparison and hyperparameter tuning, and supports parallelization across training and prediction.
The package is built on standard scientific Python dependencies (numpy, pandas, scikit-learn, scipy, joblib) and is maintained by Fidelity's AI Center of Excellence. It is used in downstream projects like Mab2Rec for personalized recommendations and ALNS for combinatorial optimization. The library is dormant—last release was 2024-08-30 with no active development—but remains stable for research and prototyping use.
Use it for
- Build a personalized recommendation system that learns which content to show each user by treating items as arms and user feedback as rewards.
- Implement A/B testing or multi-variant testing where you need to dynamically allocate traffic to better-performing variants while still exploring.
- Solve combinatorial optimization problems (e.g., vehicle routing, scheduling) by using bandits to select destroy-repair operations in a metaheuristic.
- Prototype and compare different exploration-exploitation strategies on a decision problem before deploying to production.
- Tune hyperparameters of bandit policies using the built-in simulation utility to find optimal settings for your reward structure.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you are prototyping or researching multi-armed bandit algorithms in Python.
The library is stable, well-documented, has no known vulnerabilities, and low install friction. However, dormant maintenance means no active bug fixes or feature development—suitable for research and learning, less ideal if you need ongoing support or expect breaking changes in your dependencies.
Install
mabwiser on PyPI
Before you install
Low friction: pure Python wheel with standard scientific dependencies (joblib, numpy, pandas, scikit-learn, scipy, seaborn). Last commit was 2024-09-05, but maintenance is dormant—no active development expected.
License in practice
Apache License 2.0 is permissive; you can use, modify, and distribute MABWiser freely in commercial and open-source projects with minimal restrictions.
Quickstart
from mabwiser.mab import MAB, LearningPolicy
arms = ['Arm1', 'Arm2']
mab = MAB(arms, LearningPolicy.UCB1(alpha=1.25))
mab.fit(['Arm1', 'Arm1', 'Arm2', 'Arm1'], [20, 17, 25, 9])
mab.predict()Verify before relying
- Whether the library actively supports Python versions beyond 3.8 despite classifier only listing 3.8
- Current test coverage and whether heavy testing mentioned in description is still maintained
- Whether parallelization features are production-ready or primarily research-oriented
Package facts
| License | permissive license permissive |
| Python support | Not specified |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 6 packagesjoblibnumpypandasscikit-learnscipyseaborn |
| Maintenance | Dormant 714 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 99,270 / month, #13,031 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | License :: OSI Approved :: Apache Software LicenseOperating System :: OS IndependentProgramming Language :: Python :: 3.8 |
Evidence: mabwiser-2.7.4-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “multi-armed bandit algorithms”
- mabwiserMABWiser implements multi-armed bandit algorithms for decision-making…
- bandit-sarif-formatterConverts Bandit security analysis output into SARIF 2.1.0 format for…
- flake8-banditIntegrates bandit security checks into flake8 linting, reporting…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also sb3-contrib · ax-platform · rsl-rl-lib · pymoo · implicit · skrl · open-spiel · bandit · proxsuite · emcee