Policy optimisation, reward modelling and RLHF — how models are trained by feedback rather than by labels.
Training LLM agents commonly relies on supervised fine-tuning from expert trajectories or online reinforcement learning over human-specified tasks with handcrafted verifiers. Though effective, both remain bottlenecked by externally specified tasks and supervision signals, limiting the scalability and diversity of agent training.
Large language model agents are commonly trained through reinforcement learning with sparse trajectory-level rewards, which offer limited guidance on how strongly individual tokens should be updated. On-Policy Self-Distillation (OPSD) addresses this by re-scoring generated tokens under a privileged replay view to obtain dense, token-level supervision.
Incentivized advertising allocates monetary or virtual rewards to drive user engagement, where a key challenge is optimizing continuous incentive magnitudes under strict global constraints. This problem is complicated by high-frequency interactions, delayed feedback, and non-Markovian user dynamics such as fatigue, which limit the effectiveness of existing uplift modeling and constrained reinforcement learning approaches.
arXiv:2608. 02958v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) policies trained by behavior cloning fail silently: from the action stream alone, a collapsing rollout looks much like one making clean progress, because imitation supplies no notion of progress.
arXiv:2608. 03437v1 Announce Type: cross Abstract: While human evaluation is the gold standard in many NLP tasks, it suffers from prohibitive costs and poor scalability.
arXiv:2608. 03382v1 Announce Type: cross Abstract: Multi-armed bandit algorithms, especially Thompson sampling, are widely used in online recommendation.
arXiv:2608. 03527v1 Announce Type: cross Abstract: Retrieval systems help deep research agents generate high-quality answers by providing relevant documents.
arXiv:2606. 20880v2 Announce Type: replace-cross Abstract: Decision-making under partial or adversarial observability requires accurate inference of the environment's latent state and its associated uncertainty.
arXiv:2608. 03327v1 Announce Type: new Abstract: Hybrid computer-use agents can act through screenshots or call text tools.
arXiv:2608. 03483v1 Announce Type: cross Abstract: Existing chunk-based Vision-Language-Action (VLA) models execute a fixed number of actions (i.
arXiv:2608. 03468v1 Announce Type: new Abstract: Historical tool-use trajectories provide valuable experience for large language model (LLM) agents to plan and coordinate tool usage.
arXiv:2608. 03682v1 Announce Type: new Abstract: Physical AI policies require inference throughout their lifecycle, including model evaluation, cloud reinforcement learning rollout, edge GPU serving, and onboard deployment.
arXiv:2510. 11834v3 Announce Type: replace Abstract: Generative models are increasingly paired with safety classifiers that filter harmful or undesirable outputs.
arXiv:2608. 03119v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) improves LLM reasoning but typically relies on ground-truth (GT) answers, limiting scalability.
arXiv:2602. 06932v5 Announce Type: replace Abstract: Speculative decoding can significantly accelerate LLM serving, yet most deployments today disentangle speculator training from serving, treating speculator training as a standalone offline modeling problem.
arXiv:2608. 03036v1 Announce Type: cross Abstract: Large Language Models (LLMs) are integrated into software systems and AI services, making efficient LLM serving a concern for software engineering.
arXiv:2608. 03562v1 Announce Type: new Abstract: Reinforcement learning (RL) with general utility extends classic RL by optimizing an arbitrary utility functional of the policy-induced occupancy measure, thereby enabling a broader range of applications.
arXiv:2608. 03215v1 Announce Type: cross Abstract: Reinforcement learning for flow-matching text-to-speech is complicated by deterministic ODE sampling: trajectory-level policy-gradient methods typically convert the ODE into an SDE and track per-step likelihood ratios, introducing stochastic perturbations and substantial overhead.
arXiv:2608. 03467v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) com- monly optimizes each correct completion as an independent learning signal.
arXiv:2608. 03875v1 Announce Type: cross Abstract: Designing effective reward functions remains a major bottleneck in Reinforcement Learning (RL).