Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 150 for “"Reward functions"”.

  1. Loss and Reward Functions for Generative Question Answering Systems

    Recent advancements in AI, mainly through Large Language Models (LLMs), have transformed the field, driving both industrial and academic progress. These models are typically trained using causal language modeling tasks, which, while effective, face certain limitations compared to traditional …

    trento Repository record for Loss and Reward Functions for Generative Question Answering Systems (opens in a new tab)

  2. Reinforcement Learning-Guided Generative Design of Surfactant Molecules - Exploring Reward Functions for Multi-Objective Molecular Optimisation

    … finding is that balanced multi-objective scoring functions both outperform single-objective weightings and better preserve chemical diver sity throughout training. However, a recurring failure mode termed here the hydrocarbon trap is identified, in which the optimisation converges on structurally …

    chalmers Repository record for Reinforcement Learning-Guided Generative Design of Surfactant Molecules - Exploring Reward Functions for Multi-Objective Molecular Optimisation (opens in a new tab)

  3. A flexible approach for optimal rewards in multi-agent reinforcement learning problems

    Defining a reward function that, when optimized, results in rapid acquisition of an optimal policy, is one of the most challenging tasks involved in applying reinforcement learning algorithms. The behavior learned by agents is directly related to the reward function they are using. Existing work on …

    brazil-ufrgs Repository record for A flexible approach for optimal rewards in multi-agent reinforcement learning problems (opens in a new tab)

  4. Emulation and imitation via perceptual goal specifications

    … representations for specifying domain-specific reward functions for reinforcement learning. The works outlined in this document aim to validate the following thesis statement: Employing perceptual goal specifications for goal-directed tasks: is as straightforward as specifying domain-specific …

    gatech Repository record for Emulation and imitation via perceptual goal specifications (opens in a new tab)

  5. Learning Heterogeneous Resource-Constrained Task Allocation Using Concurrent Multi-Task Bandits

    … often assume that task requirements or reward functions are known and explicitly specified by the user in advance. In this thesis, we explore the challenge of forming effective coalitions for a given heterogeneous multi-robot team when task reward functions are unknown. To tackle this …

    gatech Repository record for Learning Heterogeneous Resource-Constrained Task Allocation Using Concurrent Multi-Task Bandits (opens in a new tab)

  6. Foundation Reward Models for Robot Learning

    … in imitation learning or carefully engineered reward functions in reinforcement learning. This dependency limits scalability, as it is infeasible to collect demonstrations or design reward functions for every possible task and environment. This thesis proposes a viable path towards scaling …

    penn Repository record for Foundation Reward Models for Robot Learning (opens in a new tab)

  7. Empirical Analysis On Multi-Agent Reinforcement Learning Frameworks

    … learning is often difficult because reward functions can be challenging to specify accurately. Inverse reinforcement learning addresses this problem by inferring reward functions directly from expert demonstrations rather than relying on manually engineered objectives. This thesis …

    rowan Repository record for Empirical Analysis On Multi-Agent Reinforcement Learning Frameworks (opens in a new tab)

  8. Learning Safe and Efficient Driving Policy with Proximal Policy Optimization (PPO) using Rule-Based Rewards and Real-World Offline Data

    Designing an effective reward function is essential yet challenging for developing robust autonomous driving policies. Many existing approaches focus on specific tasks or rely on implicit rule learning, which may hinder generalization. This work introduces explicit, rule-based reward functions to …

    carleton Repository record for Learning Safe and Efficient Driving Policy with Proximal Policy Optimization (PPO) using Rule-Based Rewards and Real-World Offline Data (opens in a new tab)

  9. Robust and Flexible Reward Modeling for LLM Alignment

    … ways to align language models is by designing a reward function that measures alignment with human values, and training the language model to maximize this reward. In this thesis, we focus on two approaches reward design: reward design from external feedback signals and reward learning from human …

    gatech Repository record for Robust and Flexible Reward Modeling for LLM Alignment (opens in a new tab)

  10. Application and Assessment of Cognitive Algorithms for Locating Pollutant Source in Oceanic and Atmospheric Turbulence Modeled by Large Eddy Simulation

    … this problem by using information-theoretic reward functions to guide a mobile sensor toward the source through sequential decision-making under uncertainty. Relying on sparse and noisy measurements, these algorithms aim to optimally balance exploration of the domain with exploitation of …

    houston Repository record for Application and Assessment of Cognitive Algorithms for Locating Pollutant Source in Oceanic and Atmospheric Turbulence Modeled by Large Eddy Simulation (opens in a new tab) Embargoed until 2027-12-01

  11. Reinforcement Learning With High-Level Task Specifications

    … Existing RL algorithms typically optimize reward-based surrogates rather than the task performance itself. Therefore, they suffer from several shortcomings in providing guarantees for the task performance of the learned policies: An optimal policy for a surrogate objective may not have …

    penn Repository record for Reinforcement Learning With High-Level Task Specifications (opens in a new tab)

  12. Application of martingale methods to a change set problem

    … change set in an optimal way, according to some reward function. Given the change set, the random points are assumed to be distributed according to a Poisson process N with intensities mu0 on xic and mu1 on xi. Optimal solutions are found for various reward functions and applied to some examples.

    ottawa-retro Repository record for Application of martingale methods to a change set problem (opens in a new tab)

  13. Nonparametric Estimation and Model Combination in a Bandit Problem with Covariates

    … tradeoff to achieve optimal total reward. Motivated from industrial applications such as online advertising and clinical research, we consider a setting where the rewards of bandit machines are associated with covariates. Under a flexible problem setup, we focus on a sequential …

    umn Repository record for Nonparametric Estimation and Model Combination in a Bandit Problem with Covariates (opens in a new tab)

  14. Multi-modality Data Analysis Using Deep Reinforcement Learning

    … to design a good state space, action space and reward function for reinforcement learning models to tackle a specific task. In this thesis, we propose a series of novel techniques to cope with challenging multi-modality data based on the idea of deep reinforcement learning. We first proposed a …

    adelaide Repository record for Multi-modality Data Analysis Using Deep Reinforcement Learning (opens in a new tab)

  15. ALGORITHMS FOR CONVEX SWITCHING SYSTEMS

    … The main feature is that it contains convex reward functions and the underlying states follow linear dynamics. Many problems in engineering, finance and statistics can be modelled under the scheme of convex switching systems. Mathematically, an exact solution for such problems can be …

    nus Repository record for ALGORITHMS FOR CONVEX SWITCHING SYSTEMS (opens in a new tab) Download file for ALGORITHMS FOR CONVEX SWITCHING SYSTEMS (opens in a new tab)

  16. Application of Reinforcement Learning methods in solving a Lunar Landing game

    … Q-Learning για την επίλυση του παιχνιδιού, για 2 reward functions, και συγκρίνουμε τις αποδόσεις που πετυχαίνει ο καθένας τους. Για την ανάπυξη του περιβάλλοντος εκμάθησης, χρησιμοποιήσαμε τη βιβλιοθήκη Gym του OpenAI. Tέλος, χρησιμοποιώντας το ίδιο Gym περιβάλλον, κατασκευάζουμε και ρυθμίζουμε …

    athens Repository record for Application of Reinforcement Learning methods in solving a Lunar Landing game (opens in a new tab)

  17. Using reinforcement learning to control advanced life support systems

    … of the state table and the design of reward functions. Distributed controllers prove to be superior to centralized controllers both in terms of speed and performance of the controller. Generalization helps to speed convergence, but the performance of the policy derived is dependent on …

    rice Repository record for Using reinforcement learning to control advanced life support systems (opens in a new tab)

  18. Using natural language to aid task specification in sequential decision making problems

    … to the learning agent, i.e. designing effective reward functions for reinforcement learning and providing demonstrations for imitation learning, are often cumbersome and time-consuming. Further, designing reward functions and providing a set of demonstrations that sufficiently disambiguates the …

    texas Repository record for Using natural language to aid task specification in sequential decision making problems (opens in a new tab)

  19. Development of a multi-agent based simulation model for cyclist-pedestrian interactions in shared spaces

    … road users' behaviour and their interactions as reward-based intelligent agents. This thesis bridges the gap in modeling road user interactions by accounting for their rationality, intelligence, and sequential decision-making process by implementing the Markov Decision Process (MDP) modeling …

    ubc Repository record for Development of a multi-agent based simulation model for cyclist-pedestrian interactions in shared spaces (opens in a new tab) Download PDF for Development of a multi-agent based simulation model for cyclist-pedestrian interactions in shared spaces (opens in a new tab)

Page 1 of 8