Singing Voice Synthesis (SVS) remains constrained in practical deployment due to its strong dependence on accurate phoneme-level alignment and manually annotated melody contours, requirements that are resource-intensive and hinder scalability. To overcome these limitations, we propose a melody-driven SVS framework capable of synthesizing arbitrary lyrics following any reference melody, without relying on phoneme-level alignment. Our method builds on a Diffusion Transformer (DiT) architecture, enhanced with a dedicated melody extraction module that derives melody representations directly from reference audio. To ensure robust melody encoding, we employ a teacher model to guide the optimization of the melody extractor, alongside an implicit alignment mechanism that enforces similarity distribution constraints for improved melodic stability and coherence. Additionally, we refine duration modeling using weakly annotated song data and introduce a Flow-GRPO reinforcement learning strategy with a multi-objective reward function to jointly enhance pronunciation clarity and melodic fidelity. Experiments show that our model achieves superior performance over existing approaches in both objective measures and subjective listening tests, especially in zero-shot and lyric adaptation settings, while maintaining high audio quality without manual annotation. This work offers a practical and scalable solution for advancing data-efficient singing voice synthesis. To support reproducibility, we release our inference code and model checkpoints.

Drones are becoming indispensable in many application domains. In data-driven missions, besides sensing, the drone must process the collected data at runtime to decide whether additional action must be taken on the spot, before moving to the next point of interest. If processing does not reveal an event or situation that requires such an action, the drone has waited in vain instead of moving to the next point. If, however, the drone starts moving to the next point and it turns out that a follow-up action is needed at the previous point, it must spend time to fly-back. To take this decision, we propose different machine-learning methods based on branch prediction and reinforcement learning. We evaluate these methods for a wide range of scenarios where the probability of event occurrence changes with time. Our results show that the proposed methods consistently outperform the regression-based method proposed in the literature and can significantly improve the worst-case mission time by up to 4.1x. Also, the achieved median mission time is very close, merely up to 2.7% higher, to that of a method with perfect knowledge of the current underlying event probability at each point of interest.

Traffic congestion is a major and complex challenge for cities worldwide with the rapid growth of urbanization and vehicle ownership. Longer commute times, excessive fuel consumption, and elevated air pollution levels are direct consequences of over-saturated roads. For instance, according to the 2024 INRIX Global Traffic Scorecard, individual commuters in Istanbul, New York City, and Chicago experienced total annual delay of about 105, 102, and 102 hours, respectively, underscoring the magnitude of intersection-driven delays in major metros (INRIX). Within urban networks, signalized intersections are the dominant bottlenecks: the policies implemented at these intersections allocate scarce space-time among competing traffic streams and therefore largely determine corridor-level delay, queues, and emissions. Reinforcement learning (RL) has become a standard practice for adaptive traffic signal control (ATSC), controlling phase selection and timing as a sequential decision problem that optimizes long-horizon objectives such as delay, throughput, and emissions under nonstationary demand (Yau et al., 2017). Deep RL (DRL) extends this by using function approximation to digest rich state representations--from detector queues to trajectories and graph-structured networks--enabling policies that generalize across varying traffic flows and topologies (Zhao et al., 2024). Collectively, this body of work motivates moving beyond single-intersection controllers toward coordinated, network-level solutions and setting the stage for multi-agent formulations.

Automating disassembly of critical components from end-of-life (EoL) desktops, such as high-value items like RAM modules and CPUs, as well as sensitive parts like hard disk drives, remains challenging due to the inherent variability and uncertainty of these products. Moreover, their disassembly requires sequential, precise, and dexterous operations, further increasing the complexity of automation. Current robotic disassembly processes are typically divided into several stages: perception, sequence planning, task planning, motion planning, and manipulation. Each stage requires explicit modeling, which limits generalization to unfamiliar scenarios. Recent development of vision-language-action (VLA) models has presented an end-to-end approach for general robotic manipulation tasks. Although VLAs have demonstrated promising performance on simple tasks, the feasibility of applying such models to complex disassembly remains largely unexplored. In this paper, we collected a customized dataset for robotic RAM and CPU disassembly and used it to fine-tune two well-established VLA approaches, OpenVLA and OpenVLA-OFT, as a case study. We divided the whole disassembly task into several small steps, and our preliminary experimental results indicate that the fine-tuned VLA models can faithfully complete multiple early steps but struggle with certain critical subtasks, leading to task failure. However, we observed that a simple hybrid strategy that combines VLA with a rule-based controller can successfully perform the entire disassembly operation. These findings highlight the current limitations of VLA models in handling the dexterity and precision required for robotic EoL product disassembly. By offering a detailed analysis of the observed results, this study provides insights that may inform future research to address current challenges and advance end-to-end robotic automated disassembly.

Our work presents a novel reinforcement learning (RL) based framework to optimize heuristic selection within the conflict-driven clause learning (CDCL) process, improving the efficiency of Boolean satisfia-bility (SAT) solving. The proposed system, LangSAT, bridges the gap between natural language inputs and propositional logic by converting English descriptions into Conjunctive Normal Form (CNF) expressions and solving them using an RL-enhanced CDCL SAT solver. Unlike existing SAT-solving platforms that require CNF as input, LangSAT enables users to input standard English descriptions, making SAT-solving more accessible. The framework comprises two key components: Lang2Logic, which translates English sentences into CNF expressions, and SmartSAT, an RL-based SAT solver. SmartSAT encodes clause-variable relationships as structured graph representations and extracts global features specific to the SAT problem. This implementation provides the RL agent with deeper contextual information, enabling SAT problems to be solved more efficiently. Lang2Logic was evaluated on diverse natural language inputs, processing descriptions up to 450 words. The generated CNFs were solved by SmartSAT, which demonstrated comparable performance to traditional CDCL heuristics with respect to solving time. The combined LangSAT framework offers a more accessible and scalable solution for SAT-solving tasks across reasoning, formal verification, and debugging.

Contemporary DevSecOps pipelines have to deal with the evolution of security in an ever-continuously integrated and deployed environment. Existing methods,such as rule-based intrusion detection and static vulnerability scanning, are inadequate and unreceptive to changes in the system, causing longer response times and organization needs exposure to emerging attack vectors. In light of the previous constraints, we introduce AutoGuard to the DevSecOps ecosystem, a reinforcement learning (RL)-powered self-healing security framework built to pre-emptively protect DevSecOps environments. AutoGuard is a self-securing security environment that continuously observes pipeline activities for potential anomalies while preemptively remediating the environment. The model observes and reacts based on a policy that is continually learned dynamically over time. The RL agent improves each action over time through reward-based learning aimed at improving the agent's ability to prevent, detect and respond to a security incident in real-time. Testing using simulated ContinuousIntegration / Continuous Deployment (CI/CD) environments showed AutoGuard to successfully improve threat detection accuracy by 22%, reduce mean time torecovery (MTTR) for incidents by 38% and increase overall resilience to incidents as compared to traditional methods. Keywords- DevSecOps, Reinforcement Learning, Self- Healing Security, Continuous Integration, Automated Threat Mitigation

Finding meaningful and accurate dense rewards is a fundamental task in the field of reinforcement learning (RL) that enables agents to explore environments more efficiently. In traditional RL settings, agents learn optimal policies through interactions with an environment guided by reward signals. However, when these signals are sparse, delayed, or poorly aligned with the intended task objectives, agents often struggle to learn effectively. Dense reward functions, which provide informative feedback at every step or state transition, offer a potential solution by shaping agent behavior and accelerating learning. Despite their benefits, poorly crafted reward functions can lead to unintended behaviors, reward hacking, or inefficient exploration. This problem is particularly acute in complex or high-dimensional environments where handcrafted rewards are difficult to specify and validate. To address this, recent research has explored a variety of approaches, including inverse reinforcement learning, reward modeling from human preferences, and self-supervised learning of intrinsic rewards. While these methods offer promising directions, they often involve trade-offs between generality, scalability, and alignment with human intent. This proposal explores several approaches to dealing with these unsolved problems and enhancing the effectiveness and reliability of dense reward construction in different RL applications.

Post-training with reinforcement learning (RL) typically optimizes a single scalar objective and ignores structure in how solutions are produced. We ask whether a scalar hint toward a canonical solver ordering, used only during RL post-training, improves performance even when fine-tuned on randomized solution sequences. On Sudoku, we train a Transformer with standard fine-tuning on randomized solving orders, then post-train it with Group Relative Policy Optimization (GRPO) with two rewards: cell accuracy and an ordering reward that increases when the model's emission order aligns with the solver order. To compare signals cleanly, we combine them via fixed mixtures and use a simple bootstrapped scaling to equalize component magnitudes at initialization. Mixed rewards generally outperform cell-only optimization--the best mixture yields substantially higher test accuracy than the fine-tuned-only model trained on random-order and approaches the fine-tuned-only model trained on solver-order sequences in accuracy. These results suggest that coarse ordering signals can steer RL post-training toward solver-order trajectories without modifying supervised data or architecture.

This paper critiques common patterns in machine ethics for Reinforcement Learning (RL) and argues for a virtue focused alternative. We highlight two recurring limitations in much of the current literature: (i) rule based (deontological) methods that encode duties as constraints or shields often struggle under ambiguity and nonstationarity and do not cultivate lasting habits, and (ii) many reward based approaches, especially single objective RL, implicitly compress diverse moral considerations into a single scalar signal, which can obscure trade offs and invite proxy gaming in practice. We instead treat ethics as policy level dispositions, that is, relatively stable habits that hold up when incentives, partners, or contexts change. This shifts evaluation beyond rule checks or scalar returns toward trait summaries, durability under interventions, and explicit reporting of moral trade offs. Our roadmap combines four components: (1) social learning in multi agent RL to acquire virtue like patterns from imperfect but normatively informed exemplars; (2) multi objective and constrained formulations that preserve value conflicts and incorporate risk aware criteria to guard against harm; (3) affinity based regularization toward updateable virtue priors that support trait like stability under distribution shift while allowing norms to evolve; and (4) operationalizing diverse ethical traditions as practical control signals, making explicit the value and cultural assumptions that shape ethical RL benchmarks.

Recent advances in Omni models have enabled unified multimodal perception and generation. However, most existing systems still exhibit rigid reasoning behaviors, either overthinking simple problems or failing to reason when necessary. To address this limitation, we propose Omni-AutoThink, a novel adaptive reasoning framework that dynamically adjusts the model's reasoning depth according to task difficulty. Our framework comprises two stages: (1) an Adaptive Supervised Fine-Tuning (Adaptive SFT) stage, which endows the Omni model with fundamental reasoning capability using large-scale reasoning-augmented data, and (2) an Adaptive Reinforcement Learning (Adaptive GRPO) stage, which optimizes reasoning behaviors based on task complexity and reward feedback. We further construct a comprehensive adaptive reasoning benchmark that spans text-only, text-audio, text-visual, and text-audio-visual modalities, providing both training and evaluation splits for multimodal reasoning assessment. Experimental results demonstrate that our proposed framework significantly improves adaptive reasoning performance compared to previous baselines. All benchmark data and code will be publicly released.