Deep Learning for Reinforcement Learning
Abstract
We address the challenge of coordinating multiple robots in narrow and confined environments, where congestion and interference often hinder collective task performance. Drawing inspiration from insect colonies, which achieve robust coordination through stigmergy -- modifying and interpreting environmental traces -- we propose a Stigmergic Multi-Agent Deep Reinforcement Learning (S-MADRL) framework that leverages virtual pheromones to model local and social interactions, enabling decentralized emergent coordination without explicit communication. To overcome the convergence and scalability limitations of existing algorithms such as MADQN, MADDPG, and MAPPO, we leverage curriculum learning, which decomposes complex tasks into progressively harder sub-problems. Simulation results show that our framework achieves the most effective coordination of up to eight agents, where robots self-organize into asymmetric workload distributions that reduce congestion and modulate group performance. This emergent behavior, analogous to strategies observed in nature, demonstrates a scalable solution for decentralized multi-agent coordination in crowded environments with communication constraints.
We address the challenge of coordinating multiple robots in narrow and confined environments, where congestion and interference often hinder collective task performance. Drawing inspiration from insect colonies, which achieve robust coordination through stigmergy -- modifying and interpreting environmental traces -- we propose a Stigmergic Multi-Agent Deep Reinforcement Learning (S-MADRL) framework that leverages virtual pheromones to model local and social interactions, enabling decentralized emergent coordination without explicit communication. To overcome the convergence and scalability limitations of existing algorithms such as MADQN, MADDPG, and MAPPO, we leverage curriculum learning, which decomposes complex tasks into progressively harder sub-problems. Simulation results show that our framework achieves the most effective coordination of up to eight agents, where robots self-organize into asymmetric workload distributions that reduce congestion and modulate group performance. This emergent behavior, analogous to strategies observed in nature, demonstrates a scalable solution for decentralized multi-agent coordination in crowded environments with communication constraints.
Technical University of D
Abstract
Sample efficiency is a central property of effective deep reinforcement learning algorithms. Recent work has improved this through added complexity, such as larger models, exotic network architectures, and more complex algorithms, which are typically motivated purely by empirical performance. We take a more principled approach by focusing on the optimization landscape of the critic network. Using the eigenspectrum and condition number of the critic's Hessian, we systematically investigate the impact of common architectural design decisions on training dynamics. Our analysis reveals that a novel combination of batch normalization (BN), weight normalization (WN), and a distributional cross-entropy (CE) loss produces condition numbers orders of magnitude smaller than baselines. This combination also naturally bounds gradient norms, a property critical for maintaining a stable effective learning rate under non-stationary targets and bootstrapping. Based on these insights, we introduce XQC: a well-motivated, sample-efficient deep actor-critic algorithm built upon soft actor-critic that embodies these optimization-aware principles. We achieve state-of-the-art sample efficiency across 55 proprioception and 15 vision-based continuous control tasks, all while using significantly fewer parameters than competing methods.
Sample efficiency is a central property of effective deep reinforcement learning algorithms. Recent work has improved this through added complexity, such as larger models, exotic network architectures, and more complex algorithms, which are typically motivated purely by empirical performance. We take a more principled approach by focusing on the optimization landscape of the critic network. Using the eigenspectrum and condition number of the critic's Hessian, we systematically investigate the impact of common architectural design decisions on training dynamics. Our analysis reveals that a novel combination of batch normalization (BN), weight normalization (WN), and a distributional cross-entropy (CE) loss produces condition numbers orders of magnitude smaller than baselines. This combination also naturally bounds gradient norms, a property critical for maintaining a stable effective learning rate under non-stationary targets and bootstrapping. Based on these insights, we introduce XQC: a well-motivated, sample-efficient deep actor-critic algorithm built upon soft actor-critic that embodies these optimization-aware principles. We achieve state-of-the-art sample efficiency across 55 proprioception and 15 vision-based continuous control tasks, all while using significantly fewer parameters than competing methods.
Agentic RL
Zhejiang University, The
Abstract
The rise of LLM-powered agents is driving a fundamental transformation in services computing: from static, request-response functions to dynamic, goal-oriented, and autonomous multi-agent ecosystems. In response to this shift, we introduce Agentic Service Computing (ASC), a new paradigm that reimagines services as intelligent, self-adaptive, and socially embedded entities. This comprehensive survey presents a lifecycle-driven framework for ASC, structured around four core phases: Design, Deployment, Operation, and Evolution. We systematically analyze ASC through four foundational research dimensions: (1) Perception, Context, and Environment Modeling, (2) Autonomous Decision-Making and Task Execution, (3) Multi-Agent Collaboration and Organization, and (4) Evaluation, Value Alignment, and Trustworthiness. We examine how these dimensions are instantiated, integrated, and continuously adapted across the service lifecycle. Our synthesis reveals that agentic services are not merely assembled but orchestrated: contextual awareness enables robust deployment; autonomous reasoning supports real-time operation; collaborative structures emerge and evolve through interaction; and trustworthiness must be upheld as a cross-cutting, lifelong imperative. We further identify and discuss emerging trends shaping the future of ASC. By integrating classical principles of services computing with advances in LLM-based multi-agent systems, this work establishes a holistic and forward-looking foundation for ASC. It provides a unified reference for researchers and practitioners aiming to develop adaptive, accountable, and human-centered intelligent services.
The rise of LLM-powered agents is driving a fundamental transformation in services computing: from static, request-response functions to dynamic, goal-oriented, and autonomous multi-agent ecosystems. In response to this shift, we introduce Agentic Service Computing (ASC), a new paradigm that reimagines services as intelligent, self-adaptive, and socially embedded entities. This comprehensive survey presents a lifecycle-driven framework for ASC, structured around four core phases: Design, Deployment, Operation, and Evolution. We systematically analyze ASC through four foundational research dimensions: (1) Perception, Context, and Environment Modeling, (2) Autonomous Decision-Making and Task Execution, (3) Multi-Agent Collaboration and Organization, and (4) Evaluation, Value Alignment, and Trustworthiness. We examine how these dimensions are instantiated, integrated, and continuously adapted across the service lifecycle. Our synthesis reveals that agentic services are not merely assembled but orchestrated: contextual awareness enables robust deployment; autonomous reasoning supports real-time operation; collaborative structures emerge and evolve through interaction; and trustworthiness must be upheld as a cross-cutting, lifelong imperative. We further identify and discuss emerging trends shaping the future of ASC. By integrating classical principles of services computing with advances in LLM-based multi-agent systems, this work establishes a holistic and forward-looking foundation for ASC. It provides a unified reference for researchers and practitioners aiming to develop adaptive, accountable, and human-centered intelligent services.
AI Insights
- Federated learning enables privacy‑preserving on‑device updates for agentic services.
- Formal verification can guarantee safety of autonomous decision modules in multi‑agent ecosystems.
- Dynamic resource schedulers adapt to workload shifts, preserving QoS in agentic clusters.
- OpenAPI extensions for agentic interactions standardize cross‑domain collaboration.
- Benchmarks that score explainability, latency, and trust guide agentic framework comparison.
- Human‑in‑the‑loop UIs let users steer agentic goals while preserving autonomy.
- Edge‑centric deployments cut latency and boost resilience for distributed agentic services.
AXA Group Operations, EPF
Abstract
When used in high-stakes settings, AI systems are expected to produce decisions that are transparent, interpretable, and auditable, a requirement increasingly expected by regulations. Decision trees such as CART provide clear and verifiable rules, but they are restricted to structured tabular data and cannot operate directly on unstructured inputs such as text. In practice, large language models (LLMs) are widely used for such data, yet prompting strategies such as chain-of-thought or prompt optimization still rely on free-form reasoning, limiting their ability to ensure trustworthy behaviors. We present the Agentic Classification Tree (ACT), which extends decision-tree methodology to unstructured inputs by formulating each split as a natural-language question, refined through impurity-based evaluation and LLM feedback via TextGrad. Experiments on text benchmarks show that ACT matches or surpasses prompting-based baselines while producing transparent and interpretable decision paths.
When used in high-stakes settings, AI systems are expected to produce decisions that are transparent, interpretable, and auditable, a requirement increasingly expected by regulations. Decision trees such as CART provide clear and verifiable rules, but they are restricted to structured tabular data and cannot operate directly on unstructured inputs such as text. In practice, large language models (LLMs) are widely used for such data, yet prompting strategies such as chain-of-thought or prompt optimization still rely on free-form reasoning, limiting their ability to ensure trustworthy behaviors. We present the Agentic Classification Tree (ACT), which extends decision-tree methodology to unstructured inputs by formulating each split as a natural-language question, refined through impurity-based evaluation and LLM feedback via TextGrad. Experiments on text benchmarks show that ACT matches or surpasses prompting-based baselines while producing transparent and interpretable decision paths.
AI Insights
- ACT uses iterative prompt refinement guided by impurity metrics to craft discriminative questions.
- Experiments on DIAGNO, SPAM, and JAILBREAK datasets demonstrate ACT's competitive accuracy.
- Qualitative inspection shows ACT-generated questions align with human intuition, enhancing trust.
- The tree structure is fully auditable, allowing end-users to trace every decision path.
- Performance hinges on LLM quality; biases in the underlying model can propagate to the tree.
- The refinement loop can be resource-intensive, suggesting a trade-off between accuracy and cost.
- For deeper understanding, consult BERT and RoBERTa papers, which underpin many LLMs used in ACT.
Reinforcement Learning
University of Notre Dame
Abstract
The development of state-of-the-art large language models is commonly understood as a two-stage process involving pre-training and post-training. We point out the need for an additional intermediate stage called reinforcement mid-training with potential for strong performance gains. In this paper, we formally define the problem and identify three key challenges: (1) inefficient training due to excessive reasoning steps, (2) disregard of the imbalanced token entropy distribution, and (3) underutilization of token information. To address these challenges, we propose RMT, a framework for efficient, adaptive, and unified reinforcement mid-training with various innovative components. In particular, we first introduce a dynamic token budget mechanism that constrains unnecessary reasoning steps and mitigates model overthinking. Next, we design a curriculum-based adaptive sampling method that fosters a progressive learning trajectory from easy to hard tokens. Finally, we present a dual training strategy that combines reinforcement learning with next-token prediction, ensuring targeted learning on key tokens and full exploitation of all token information. Extensive experiments demonstrate the superiority of RMT over state-of-the-art methods, achieving up to +64.91% performance improvement with only 21% of the reasoning length in language modeling. We also show that checkpoints obtained after reinforcement mid-training can benefit the subsequent post-training, yielding up to +18.76% improvement in the mathematical domain.
The development of state-of-the-art large language models is commonly understood as a two-stage process involving pre-training and post-training. We point out the need for an additional intermediate stage called reinforcement mid-training with potential for strong performance gains. In this paper, we formally define the problem and identify three key challenges: (1) inefficient training due to excessive reasoning steps, (2) disregard of the imbalanced token entropy distribution, and (3) underutilization of token information. To address these challenges, we propose RMT, a framework for efficient, adaptive, and unified reinforcement mid-training with various innovative components. In particular, we first introduce a dynamic token budget mechanism that constrains unnecessary reasoning steps and mitigates model overthinking. Next, we design a curriculum-based adaptive sampling method that fosters a progressive learning trajectory from easy to hard tokens. Finally, we present a dual training strategy that combines reinforcement learning with next-token prediction, ensuring targeted learning on key tokens and full exploitation of all token information. Extensive experiments demonstrate the superiority of RMT over state-of-the-art methods, achieving up to +64.91% performance improvement with only 21% of the reasoning length in language modeling. We also show that checkpoints obtained after reinforcement mid-training can benefit the subsequent post-training, yielding up to +18.76% improvement in the mathematical domain.
University of Toronto and
Abstract
We study reinforcement learning problems where state observations are stochastically triggered by actions, a constraint common in many real-world applications. This framework is formulated as Action-Triggered Sporadically Traceable Markov Decision Processes (ATST-MDPs), where each action has a specified probability of triggering a state observation. We derive tailored Bellman optimality equations for this framework and introduce the action-sequence learning paradigm in which agents commit to executing a sequence of actions until the next observation arrives. Under the linear MDP assumption, value-functions are shown to admit linear representations in an induced action-sequence feature map. Leveraging this structure, we propose off-policy estimators with statistical error guarantees for such feature maps and introduce ST-LSVI-UCB, a variant of LSVI-UCB adapted for action-triggered settings. ST-LSVI-UCB achieves regret $\widetilde O(\sqrt{Kd^3(1-\gamma)^{-3}})$, where $K$ is the number of episodes, $d$ the feature dimension, and $\gamma$ the discount factor (per-step episode non-termination probability). Crucially, this work establishes the theoretical foundation for learning with sporadic, action-triggered observations while demonstrating that efficient learning remains feasible under such observation constraints.
We study reinforcement learning problems where state observations are stochastically triggered by actions, a constraint common in many real-world applications. This framework is formulated as Action-Triggered Sporadically Traceable Markov Decision Processes (ATST-MDPs), where each action has a specified probability of triggering a state observation. We derive tailored Bellman optimality equations for this framework and introduce the action-sequence learning paradigm in which agents commit to executing a sequence of actions until the next observation arrives. Under the linear MDP assumption, value-functions are shown to admit linear representations in an induced action-sequence feature map. Leveraging this structure, we propose off-policy estimators with statistical error guarantees for such feature maps and introduce ST-LSVI-UCB, a variant of LSVI-UCB adapted for action-triggered settings. ST-LSVI-UCB achieves regret $\widetilde O(\sqrt{Kd^3(1-\gamma)^{-3}})$, where $K$ is the number of episodes, $d$ the feature dimension, and $\gamma$ the discount factor (per-step episode non-termination probability). Crucially, this work establishes the theoretical foundation for learning with sporadic, action-triggered observations while demonstrating that efficient learning remains feasible under such observation constraints.
Help us improve your experience!
This project is on its early stages your feedback can be pivotal on the future of the project.
Let us know what you think about this week's papers and suggestions!
Give Feedback