cd /news/machine-learning/phi-td-distributional-reinforcement-… · home › topics › machine-learning › article
[ARTICLE · art-173046] src=arxiv.org ↗ pub= topic=machine-learning verified=true sentiment=↑ positive

{\phi}TD: Distributional Reinforcement Learning using Characteristic Functions

A paper submitted to arXiv on 18 November 2026 proposes φTD, a distributional reinforcement learning method that computes the loss in the frequency domain using characteristic functions. The authors report that φTD unifies quantile and categorical approaches to distributional RL and enables learning of distributions previously impossible to capture, such as mixtures of continuous distributions. Across synthetic benchmarks and large-scale deep RL environments, the authors demonstrate that the flexibility from wider parametric families can lead to improved performance.

key takeaway

Computing the distributional RL loss in the frequency domain via characteristic functions lets φTD represent mixtures of continuous distributions that quantile and categorical methods cannot, and the authors report improved performance on synthetic benchmarks and large-scale deep RL environments.

by read1 min views1 publishedNov 20, 2026
{\phi}TD: Distributional Reinforcement Learning using Characteristic Functions
Image: source
  [Submitted on 18 Nov 2026]


[View PDF](https://arxiv.org/pdf/2611.25208)

[HTML (experimental)](https://arxiv.org/html/2611.25208v1)

Abstract:Distributional Reinforcement Learning methods aim to learn the entire distribution of returns, yet current algorithmic approaches are limited to a narrow class of distributions. We propose $\varphi$TD, a method to overcome this limitation by computing the loss in the frequency domain. We demonstrate that this further unifies quantile and categorical approaches to distributional RL, and enables the learning of distributions that were previously impossible to capture, such as mixtures of continuous distributions. We empirically study $\varphi$TD, across synthetic benchmarks and large-scale deep RL environments and demonstrate that the flexibility obtained from wider parametric families can lead to improved performance.

Additional Features

References & Citations

...

Bibliographic Explorer

(What is the Explorer?) Connected Papers

(What is Connected Papers?) Litmaps

(What is Litmaps?) scite Smart Citations

(What are Smart Citations?) alphaXiv

(What is alphaXiv?) CatalyzeX Code Finder for Papers

(What is CatalyzeX?) DagsHub

(What is DagsHub?) Gotit.pub

(What is GotitPub?) Hugging Face

(What is Huggingface?) ScienceCast

(What is ScienceCast?) Influence Flower

(What are Influence Flowers?) CORE Recommender

(What is CORE?) IArxiv Recommender

(What is IArxiv?) arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

── more in #machine-learning 4 stories · sorted by recency
── more on @arxiv 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/phi-td-distributiona…] indexed:0 read:1min 2026-11-20 · —