# \#Reinforcement-Learning **URL:** https://forums.unrealengine.com/tag/reinforcement-learning/11469.md [Latest](https://forums.unrealengine.com/latest.md) · [Categories](https://forums.unrealengine.com/categories.md) · [Tags](https://forums.unrealengine.com/tags.md) --- ## [Reset recurrent memory at episode boundaries instead of masking the loss in Learning Agents](https://forums.unrealengine.com/t/reset-recurrent-memory-at-episode-boundaries-instead-of-masking-the-loss-in-learning-agents/2845440)
**Author:** [@MannyVindiola](https://forums.unrealengine.com/u/MannyVindiola)\ **Replies:** 1\ **Last updated:** [October 27, 2026, 2:37pm UTC](https://forums.unrealengine.com/t/reset-recurrent-memory-at-episode-boundaries-instead-of-masking-the-loss-in-learning-agents/2845440 "2026-10-27T14:37:35Z")
This one is not a bug report but rather, an alternative suggestion for your consideration. learning\_agents/ppo.py trains recurrent policies on fixed-length windows (PolicyWindowSize) taken from the flattened experience b… --- ## [Off-by-one errors in ppo.py affect the last episode of a gather during training](https://forums.unrealengine.com/t/off-by-one-errors-in-ppo-py-affect-the-last-episode-of-a-gather-during-training/2845303)
**Author:** [@MannyVindiola](https://forums.unrealengine.com/u/MannyVindiola)\ **Replies:** 1\ **Last updated:** [October 27, 2026, 1:39pm UTC](https://forums.unrealengine.com/t/off-by-one-errors-in-ppo-py-affect-the-last-episode-of-a-gather-during-training/2845303 "2026-10-27T13:39:19Z")
Hello, I think I have found two off-by-one errors in the LearningAgents PPO python trainer ppo.py While comparing the LearningAgents Python PPO trainer to the CleanRL version I found two off-by-one errors in loop bound… --- ## [Incorrect state when buffer fills mid-episode. ](https://forums.unrealengine.com/t/incorrect-state-when-buffer-fills-mid-episode/2835458)
**Author:** [@MannyVindiola](https://forums.unrealengine.com/u/MannyVindiola)\ **Replies:** 3\ **Last updated:** [October 22, 2026, 3:40pm UTC](https://forums.unrealengine.com/t/incorrect-state-when-buffer-fills-mid-episode/2835458 "2026-10-22T15:40:30Z")
Hello, I think there is an issue with filling the replay buffer when an episode is longer than the remaining steps. If you look here: https://github.com/EpicGames/UnrealEngine/blob/396c9f059903aed5fec78ecd3d437a40c641… --- ## [PPO eps clip is unbounded when advantage is negative ](https://forums.unrealengine.com/t/ppo-eps-clip-is-unbounded-when-advantage-is-negative/2835449)
**Author:** [@MannyVindiola](https://forums.unrealengine.com/u/MannyVindiola)\ **Replies:** 2\ **Last updated:** [October 22, 2026, 3:39pm UTC](https://forums.unrealengine.com/t/ppo-eps-clip-is-unbounded-when-advantage-is-negative/2835449 "2026-10-22T15:39:45Z")
Hello, I have found an issue with the trust region when adavantage is negative. Take a look at this code snippet. surr = torch.min(ratio, ratio.clamp(eps\_clip\_log\_min, eps\_clip\_log\_max)) loss\_surr += ... adv\[:,i\] \* -(s… --- ## [LearningAgentsInteractor::SetActionVector , size check uses wrong getter](https://forums.unrealengine.com/t/learningagentsinteractor-setactionvector-size-check-uses-wrong-getter/2719428)
**Author:** [@AdyTheQueen2020](https://forums.unrealengine.com/u/AdyTheQueen2020)\ **Replies:** 1\ **Last updated:** [May 1, 2026, 5:36pm UTC](https://forums.unrealengine.com/t/learningagentsinteractor-setactionvector-size-check-uses-wrong-getter/2719428 "2026-05-01T17:36:43Z")
Hi, Spotted what looks like a copy-paste bug in Epic’s LearningAgentsInteractor.cpp inside SetActionVector: https://github.com/EpicGames/UnrealEngine/blob/13b6acfd6224d13e37e5c75e94415544116ca945/Engine/Plugins/Experim… --- ## [Combining Learning Agents with Behavior Tree ? ](https://forums.unrealengine.com/t/combining-learning-agents-with-behavior-tree/2530347)
**Author:** [@Hoie](https://forums.unrealengine.com/u/Hoie)\ **Replies:** 4\ **Last updated:** [July 31, 2025, 1:51pm UTC](https://forums.unrealengine.com/t/combining-learning-agents-with-behavior-tree/2530347 "2025-07-31T13:51:30Z")
Hello. I’m a student working on an Action RPG game. I’m entirely new in LA, and I just have completed following example. Unreal Engine 5.4: Learning Agents for Character (Part 1) I want to implement the boss’s actions… --- ## [Bypassing obstacles and tracking enemies](https://forums.unrealengine.com/t/bypassing-obstacles-and-tracking-enemies/2547561)
**Author:** [@Hoie](https://forums.unrealengine.com/u/Hoie)\ **Replies:** 0\ **Last updated:** [June 9, 2025, 1:17am UTC](https://forums.unrealengine.com/t/bypassing-obstacles-and-tracking-enemies/2547561 "2025-06-09T01:17:35Z")
Sorry for the repost, I forgot to tag ‘Learning Agent’. I used raycast to store the distance from the wall. I penalized the agent when it is facing the wall and when it touches the wall. Here are the rewards Th… --- ## [Is real-time multi-agent RL with Transformers possible in UE5?](https://forums.unrealengine.com/t/is-real-time-multi-agent-rl-with-transformers-possible-in-ue5/2539266)
**Author:** [@SHAY2407](https://forums.unrealengine.com/u/SHAY2407)\ **Replies:** 0\ **Last updated:** [June 3, 2025, 4:04pm UTC](https://forums.unrealengine.com/t/is-real-time-multi-agent-rl-with-transformers-possible-in-ue5/2539266 "2025-06-03T16:04:44Z")
I’m primarily an AI/ML student with limited UE5 experience, so I’m honestly not sure how much of this is even technically feasible in a game engine context. I’m exploring a combat arena concept with 50+ AI agents learni… --- ## [Help with Learning Agents plugin UE 5.4](https://forums.unrealengine.com/t/help-with-learning-agents-plugin-ue-5-4/2525958)
**Author:** [@QYQYA](https://forums.unrealengine.com/u/QYQYA)\ **Replies:** 4\ **Last updated:** [May 28, 2025, 5:31pm UTC](https://forums.unrealengine.com/t/help-with-learning-agents-plugin-ue-5-4/2525958 "2025-05-28T17:31:17Z")
Is these all it takes to indicate the wall location, so learning agents’ bots can use it as cover? I assume I need to change something in bp trainer and add reward for reaching the wall, but I don’t know how --- ## [Q-Learning (Reinforcement Learning) in UE5.4.4 issue](https://forums.unrealengine.com/t/q-learning-reinforcement-learning-in-ue5-4-4-issue/2509119)
**Author:** [@QYQYA](https://forums.unrealengine.com/u/QYQYA)\ **Replies:** 0\ **Last updated:** [May 14, 2025, 1:14am UTC](https://forums.unrealengine.com/t/q-learning-reinforcement-learning-in-ue5-4-4-issue/2509119 "2025-05-14T01:14:49Z")
I’m trying to make AI Bot with 4 functions (Attack, Hide, Move to Player, Retreat) vs Player and I want to use Q-Learning method for the bot, but I can’t find any good tutorial for it and Learning Agents Plugin. Grok, Ch… --- ## [Learning Agents Project Doesn't Work as Expected](https://forums.unrealengine.com/t/learning-agents-project-doesnt-work-as-expected/2352999)
**Author:** [@enesQwerty3](https://forums.unrealengine.com/u/enesQwerty3)\ **Replies:** 1\ **Last updated:** [February 26, 2025, 4:42pm UTC](https://forums.unrealengine.com/t/learning-agents-project-doesnt-work-as-expected/2352999 "2025-02-26T16:42:47Z")
Hi, I’m working on a Learning Agents project in Unreal Engine 5.4 as my graduation project but I have a problem and I don’t know where I’m doing mistake. I need your help guys. I have to finish this project as soon as po… --- ## [UnrealMLAgents 1.0.0: Open-Source Deep Reinforcement Learning Framework!](https://forums.unrealengine.com/t/unrealmlagents-1-0-0-open-source-deep-reinforcement-learning-framework/2334035)
**Author:** [@AlanLaboratory](https://forums.unrealengine.com/u/AlanLaboratory)\ **Replies:** 0\ **Last updated:** [February 15, 2025, 11:43am UTC](https://forums.unrealengine.com/t/unrealmlagents-1-0-0-open-source-deep-reinforcement-learning-framework/2334035 "2025-02-15T11:43:59Z")
Hey everyone! I’ve been working on a Deep Reinforcement Learning framework for game engines, and I’m excited to share UnrealMLAgents! This project is an ongoing port of Unity ML-Agents to Unreal Engine, bringing similar… --- ## [Learning Agents Driving Tutorial + Obstacle avoidance](https://forums.unrealengine.com/t/learning-agents-driving-tutorial-obstacle-avoidance/2245470)
**Author:** [@OptimusSkooch](https://forums.unrealengine.com/u/OptimusSkooch)\ **Replies:** 5\ **Last updated:** [January 27, 2025, 2:48pm UTC](https://forums.unrealengine.com/t/learning-agents-driving-tutorial-obstacle-avoidance/2245470 "2025-01-27T14:48:59Z")
Hi I’m trying to make the driving agents also avoid obstacles on the track, but I can’t get it to work. After reading all the posts about this topic, such as: Learning Agents 5.4 Question, adding obstacles to the track … --- ## [DeepMimic-like System With Learning Agents Plugin](https://forums.unrealengine.com/t/deepmimic-like-system-with-learning-agents-plugin/1316035)
**Author:** [@ncmcclure](https://forums.unrealengine.com/u/ncmcclure)\ **Replies:** 12\ **Last updated:** [January 17, 2025, 1:11am UTC](https://forums.unrealengine.com/t/deepmimic-like-system-with-learning-agents-plugin/1316035 "2025-01-17T01:11:30Z")
I’m creating a new thread upon @Deathcalibur 's request based on a performance problem I’ve been encountering with my Learning Agents system. This should act as a place to discuss these issues and hopefully share finding… --- ## [How to use custom model in Learning Agents?](https://forums.unrealengine.com/t/how-to-use-custom-model-in-learning-agents/1947225)
**Author:** [@kromkh8](https://forums.unrealengine.com/u/kromkh8)\ **Replies:** 9\ **Last updated:** [December 18, 2024, 6:48pm UTC](https://forums.unrealengine.com/t/how-to-use-custom-model-in-learning-agents/1947225 "2024-12-18T18:48:49Z")
I am planning to conduct research on reinforcement learning and need to gather information and create a simple demo. I am currently planning to use ‘Learning Agents’ and would like to know how to apply a custom model. It… --- ## [Is there a way to determine when training starts in Learning Agents?](https://forums.unrealengine.com/t/is-there-a-way-to-determine-when-training-starts-in-learning-agents/2221562)
**Author:** [@kromkh8](https://forums.unrealengine.com/u/kromkh8)\ **Replies:** 1\ **Last updated:** [December 16, 2024, 2:26pm UTC](https://forums.unrealengine.com/t/is-there-a-way-to-determine-when-training-starts-in-learning-agents/2221562 "2024-12-16T14:26:53Z")
When using Learning Agents for training, is there a way to determine when training starts? I am conducting reinforcement learning using Learning Agents and have created a blueprint based on the ‘Learning to drive’ tutori… --- ## [Subprocess terminating before training can start](https://forums.unrealengine.com/t/subprocess-terminating-before-training-can-start/2143801)
**Author:** [@gwh0014](https://forums.unrealengine.com/u/gwh0014)\ **Replies:** 4\ **Last updated:** [November 22, 2024, 4:58pm UTC](https://forums.unrealengine.com/t/subprocess-terminating-before-training-can-start/2143801 "2024-11-22T16:58:09Z")
I am working through the tutorial learning to drive again but on 5.5 now that the update has come out. while i am reworking through the tutorial i am also using my own car model (i am not sure if that is relevant or not)… --- ## [Multiagents Reinforcement Learning Tool for Unreal Engine 5](https://forums.unrealengine.com/t/multiagents-reinforcement-learning-tool-for-unreal-engine-5/1240259)
**Author:** [@UnrayNS](https://forums.unrealengine.com/u/UnrayNS)\ **Replies:** 5\ **Last updated:** [June 5, 2024, 10:06pm UTC](https://forums.unrealengine.com/t/multiagents-reinforcement-learning-tool-for-unreal-engine-5/1240259 "2024-06-05T22:06:35Z")
Introducing “Unray” - Harnessing Reinforcement Learning for Unreal Engine! I am thrilled to offer a glimpse into the development of “Unray,” an RL tool crafted exclusively for Unreal Engine. Currently in the works, Unra… --- ## [\[PLUGIN\] NevarokML: Machine Learning for Unreal Engine](https://forums.unrealengine.com/t/plugin-nevarokml-machine-learning-for-unreal-engine/1318848)
**Author:** [@nevarok](https://forums.unrealengine.com/u/nevarok)\ **Replies:** 12\ **Last updated:** [March 7, 2024, 6:51am UTC](https://forums.unrealengine.com/t/plugin-nevarokml-machine-learning-for-unreal-engine/1318848 "2024-03-07T06:51:05Z")
I’m excited to share NevarokML Plugin, a project I’ve been grinding on since March. It’s tailored for Reinforcement Learning within Unreal Engine and supports both C++ and Blueprint users. Get started and explore: GitH…