Inequality in Congestion Games with Learning Agents

• URL: http://arxiv.org/abs/2601.20578v2
• Date: Fri, 30 Jan 2026 09:25:06 GMT
• Title: Inequality in Congestion Games with Learning Agents
• Authors: Dimitris Michailidis, Sennay Ghebreab, Fernando P. Santos,
• Abstract summary: We show that disparities arise not only from the structure of the network but also from differences in how commuters adapt to it.<n>To capture potential efficiency-fairness tradeoffs, we introduce the Price of Learning (PoL), a measure of inefficiency during learning.<n>Our simulations show that network expansions can simultaneously increase efficiency and amplify inequality.
• Score: 49.16654883862325
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Who benefits from expanding transport networks? While designed to improve mobility, such interventions can also create inequality. In this paper, we show that disparities arise not only from the structure of the network itself but also from differences in how commuters adapt to it. We model commuters as reinforcement learning agents who adapt their travel choices at different learning rates, reflecting unequal access to resources and information. To capture potential efficiency-fairness tradeoffs, we introduce the Price of Learning (PoL), a measure of inefficiency during learning. We analyze both a stylized network -- inspired in the well-known Braess's paradox, yet with two-source nodes -- and an abstraction of a real-world metro system (Amsterdam). Our simulations show that network expansions can simultaneously increase efficiency and amplify inequality, especially when faster learners disproportionately benefit from new routes before others adapt. These results highlight that transport policies must account not only for equilibrium outcomes but also for the heterogeneous ways commuters adapt, since both shape the balance between efficiency and fairness.

Related papers

• FAST: Similarity-based Knowledge Transfer for Efficient Policy Learning [57.4737157531239]
  Transfer Learning offers the potential to accelerate learning by transferring knowledge across tasks.<n>It faces critical challenges such as negative transfer, domain adaptation and inefficiency in selecting solid source policies.<n>In this work we challenge the key issues in TL to improve knowledge transfer, agents performance across tasks and reduce computational costs.
  arXiv  Detail & Related papers  (2025-07-27T22:21:53Z)
• Transfer Learning in Latent Contextual Bandits with Covariate Shift Through Causal Transportability [8.593124849405498]
  Transferring knowledge from one environment to another is an essential ability of intelligent systems.<n>In this paper, we consider transfer learning in latent contextual bandits, where the actual context is hidden.<n>We show that naively transferring all knowledge for classical bandit algorithms in this setting led to negative transfer.
  arXiv  Detail & Related papers  (2025-02-27T14:52:23Z)
• Real-time system optimal traffic routing under uncertainties -- Can physics models boost reinforcement learning? [2.298129181817085]
  Our paper presents TransRL, a novel algorithm that integrates reinforcement learning with physics models for enhanced performance, reliability, and interpretability. By leveraging the information from physics models, TransRL consistently outperforms state-of-the-art reinforcement learning algorithms.
  arXiv  Detail & Related papers  (2024-07-10T04:53:26Z)
• Common Knowledge Learning for Generating Transferable Adversarial Examples [60.1287733223249]
  This paper focuses on an important type of black-box attacks, where the adversary generates adversarial examples by a substitute (source) model. Existing methods tend to give unsatisfactory adversarial transferability when the source and target models are from different types of DNN architectures. We propose a common knowledge learning (CKL) framework to learn better network weights to generate adversarial examples.
  arXiv  Detail & Related papers  (2023-07-01T09:07:12Z)
• SocialLight: Distributed Cooperation Learning towards Network-Wide Traffic Signal Control [7.387226437589183]
  SocialLight is a new multi-agent reinforcement learning method for traffic signal control. It learns cooperative traffic control policies by estimating the individual marginal contribution of agents on their local neighborhood. We benchmark our trained network against state-of-the-art traffic signal control methods on standard benchmarks in two traffic simulators.
  arXiv  Detail & Related papers  (2023-04-20T12:41:25Z)
• Why Do Self-Supervised Models Transfer? Investigating the Impact of Invariance on Downstream Tasks [79.13089902898848]
  Self-supervised learning is a powerful paradigm for representation learning on unlabelled images. We show that different tasks in computer vision require features to encode different (in)variances.
  arXiv  Detail & Related papers  (2021-11-22T18:16:35Z)
• PsiPhi-Learning: Reinforcement Learning with Demonstrations using Successor Features and Inverse Temporal Difference Learning [102.36450942613091]
  We propose an inverse reinforcement learning algorithm, called emphinverse temporal difference learning (ITD) We show how to seamlessly integrate ITD with learning from online environment interactions, arriving at a novel algorithm for reinforcement learning with demonstrations, called $Psi Phi$-learning.
  arXiv  Detail & Related papers  (2021-02-24T21:12:09Z)
• Bridging the Imitation Gap by Adaptive Insubordination [88.35564081175642]
  We show that when the teaching agent makes decisions with access to privileged information, this information is marginalized during imitation learning. We propose 'Adaptive Insubordination' (ADVISOR) to address this gap. ADVISOR dynamically weights imitation and reward-based reinforcement learning losses during training, enabling on-the-fly switching between imitation and exploration.
  arXiv  Detail & Related papers  (2020-07-23T17:59:57Z)
• Feature-map-level Online Adversarial Knowledge Distillation [36.42703358752526]
  We propose an online knowledge distillation method that transfers not only the knowledge of the class probabilities but also that of the feature map. We train multiple networks simultaneously by employing discriminators to distinguish the feature map distributions of different networks. We show that our method performs better than the conventional direct alignment method such as L1 and is more suitable for online distillation.
  arXiv  Detail & Related papers  (2020-02-05T13:16:37Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.