{"url":"/method/noisy-net","slug":"noisy-net","name":"Noisy Linear Layer","full_name":"Noisy Linear Layer","full_name_withheld":false,"description_markdown":"A **Noisy Linear Layer** is a [linear layer](https://paperswithcode.com/method/linear-layer) with parametric noise added to the weights. This induced stochasticity can be used in reinforcement learning networks for the agent's policy to aid efficient exploration. The parameters of the noise are learned with gradient descent along with any other remaining network weights. Factorized Gaussian noise is the type of noise usually employed.\r\n\r\nThe noisy linear layer takes the form:\r\n\r\n$$y = \\left(b + Wx\\right) + \\left(b\\_{noisy}\\odot\\epsilon^{b}+\\left(W\\_{noisy}\\odot\\epsilon^{w}\\right)x\\right) $$\r\n\r\nwhere $\\epsilon^{b}$ and $\\epsilon^{w}$ are random variables.","description_state":"present","introduced_year":null,"introduced_by":{"title":"Noisy Networks for Exploration","paper":"/paper/noisy-networks-for-exploration","first_author":"Meire Fortunato","n_authors":12,"url_abs":null,"archive_paper_url":"https://paperswithcode.com/paper/noisy-networks-for-exploration"},"source":{"url":"https://arxiv.org/abs/1706.10295v3","title":"Noisy Networks for Exploration","url_on_a_paper_host":true},"code_snippet_url":null,"code_snippet_url_on_a_code_host":false,"categories":[{"area":"Reinforcement Learning","area_id":"reinforcement-learning","collection":"Randomized Value Functions","url":"/methods/category/randomized-value-functions","pwc_aliases":[]}],"n_papers_tagged":11,"archive_num_papers":11,"papers_newest_first":[{"paper":"/paper/beyond-the-rainbow-high-performance-deep","title":"Beyond The Rainbow: High Performance Deep Reinforcement Learning on a Desktop PC","date":"2024-11-06","arxiv_id":"2411.03820","n_code_links":3,"syntology":{"ran":13,"of":25,"unverified":12,"pointer_only":21}},{"paper":"/paper/reducing-variance-in-temporal-difference","title":"Reducing Variance in Temporal-Difference Value Estimation via Ensemble of Deep Networks","date":"2022-09-16","arxiv_id":"2209.07670","n_code_links":1,"syntology":null},{"paper":"/paper/dna-proximal-policy-optimization-with-a-dual","title":"DNA: Proximal Policy Optimization with a Dual Network Architecture","date":"2022-06-20","arxiv_id":"2206.10027","n_code_links":1,"syntology":null},{"paper":"/paper/deep-reinforcement-learning-at-the-edge-of","title":"Deep Reinforcement Learning at the Edge of the Statistical Precipice","date":"2021-08-30","arxiv_id":"2108.13264","n_code_links":3,"syntology":{"ran":5,"of":5,"unverified":0,"pointer_only":0}},{"paper":null,"title":"Weighted Bellman Backups for Improved Signal-to-Noise in Q-Updates","date":"2021-01-01","arxiv_id":null,"n_code_links":0,"syntology":null},{"paper":null,"title":"A New Approach for Tactical Decision Making in Lane Changing: Sample Efficient Deep Q Learning with a Safety Feedback Reward","date":"2020-09-24","arxiv_id":"2009.11905","n_code_links":0,"syntology":null},{"paper":"/paper/sunrise-a-simple-unified-framework-for","title":"SUNRISE: A Simple Unified Framework for Ensemble Learning in Deep Reinforcement Learning","date":"2020-07-09","arxiv_id":"2007.04938","n_code_links":1,"syntology":null},{"paper":null,"title":"NROWAN-DQN: A Stable Noisy Network with Noise Reduction and Online Weight Adjustment for Exploration","date":"2020-06-19","arxiv_id":"2006.10980","n_code_links":0,"syntology":null},{"paper":null,"title":"Generative Adversarial Imagination for Sample Efficient Deep Reinforcement Learning","date":"2019-04-30","arxiv_id":"1904.13255","n_code_links":0,"syntology":null},{"paper":"/paper/rainbow-combining-improvements-in-deep","title":"Rainbow: Combining Improvements in Deep Reinforcement Learning","date":"2017-10-06","arxiv_id":"1710.02298","n_code_links":34,"syntology":{"ran":2,"of":6,"unverified":4,"pointer_only":1}},{"paper":"/paper/noisy-networks-for-exploration","title":"Noisy Networks for Exploration","date":"2017-06-30","arxiv_id":"1706.10295","n_code_links":15,"syntology":{"ran":1,"of":3,"unverified":2,"pointer_only":3}}],"papers_shown":11,"tasks":[{"task":"/task/reinforcement-learning-1","name":"Reinforcement Learning (RL)","papers":9},{"task":"/task/deep-reinforcement-learning","name":"Deep Reinforcement Learning","papers":8},{"task":"/task/reinforcement-learning-2","name":"reinforcement-learning","papers":6},{"task":"/task/reinforcement-learning","name":"Reinforcement Learning","papers":5},{"task":"/task/atari-games","name":"Atari Games","papers":3},{"task":"/task/q-learning","name":"Q-Learning","papers":3},{"task":"/task/efficient-exploration","name":"Efficient Exploration","papers":2},{"task":"/task/computational-efficiency","name":"Computational Efficiency","papers":1},{"task":"/task/decision-making","name":"Decision Making","papers":1},{"task":"/task/diversity","name":"Diversity","papers":1},{"task":"/task/ensemble-learning","name":"Ensemble Learning","papers":1},{"task":"/task/game-of-go","name":"Game of Go","papers":1},{"task":"/task/montezumas-revenge","name":"Montezuma's Revenge","papers":1}],"tasks_shown":13,"n_tasks":13,"usage_by_year":[{"year":"2017","papers":2},{"year":"2019","papers":1},{"year":"2020","papers":3},{"year":"2021","papers":2},{"year":"2022","papers":2},{"year":"2024","papers":1}],"row_source":"methods_table","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/noisy-net"},"syntology_read_at":"2026-09-24T18:15:14+00:00"}