arXiv:2603.06656v1 Announce Type: cross Abstract: Human gameplay is a visually grounded interaction loop in which players act, reflect on failures, and watch tutorials to refine strategies. Can Vision-Language Models (VLMs) also learn from video-based reflection? We present GameVerse, a comprehensive video game benchmark that enables a reflective visual interaction loop. Moving beyond traditional fire-and-forget evaluations, it uses a novel reflect-and-retry paradigm to assess how VLMs internalize visual experience and improve policies. To facilitate systematic and scalable evaluation, we also introduce a cognitive hierarchical taxonomy spanning 15 globally popular games, dual action space for both semantic and GUI control, and milestone evaluation using advanced VLMs to quantify progress. Our experiments show that VLMs benefit from video-based reflection in varied settings, and perform best by combining failure trajectories and expert tutorials-a training-free analogue to reinforcement learning (RL) plus supervised fine-tuning (SFT).

arXiv:2603.08398v1 Announce Type: cross Abstract: In this work, we reveal that Large Language Models (LLMs) possess intrinsic behavioral plasticity-akin to chameleons adapting their coloration to environmental cues-that can be exposed through token-conditional generation and stabilized via reinforcement learning. Specifically, by conditioning generation on carefully selected token prefixes sampled from responses exhibiting desired behaviors, LLMs seamlessly adapt their behavioral modes at inference time (e.g., switching from step-by-step reasoning to direct answering) without retraining. Based on this insight, we propose Token-Conditioned Reinforcement Learning (ToCoRL), a principled framework that leverages RL to internalize this chameleon-like plasticity, transforming transient inference-time adaptations into stable and learnable behavioral patterns. ToCoRL guides exploration with token-conditional generation and keep enhancing exploitation, enabling emergence of appropriate behaviors. Extensive experiments show that ToCoRL enables precise behavioral control without capability degradation. Notably, we show that large reasoning models, while performing strongly on complex mathematics, can be effectively adapted to excel at factual question answering, which was a capability previously hindered by their step-by-step reasoning patterns.

arXiv:2509.26354v2 Announce Type: replace Abstract: Advances in Large Language Models (LLMs) have enabled a new class of self-evolving agents that autonomously improve through interaction with the environment, demonstrating strong capabilities. However, self-evolution also introduces novel risks overlooked by current safety research. In this work, we study the case where an agent's self-evolution deviates in unintended ways, leading to undesirable or even harmful outcomes. We refer to this as Misevolution. To provide a systematic investigation, we evaluate misevolution along four key evolutionary pathways: model, memory, tool, and workflow. Our empirical findings reveal that misevolution is a widespread risk, affecting agents built even on top-tier LLMs (e.g., Gemini-2.5-Pro). Different emergent risks are observed in the self-evolutionary process, such as the degradation of safety alignment after memory accumulation, or the unintended introduction of vulnerabilities in tool creation and reuse. To our knowledge, this is the first study to systematically conceptualize misevolution and provide empirical evidence of its occurrence, highlighting an urgent need for new safety paradigms for self-evolving agents. Finally, we discuss potential mitigation strategies to inspire further research on building safer and more trustworthy self-evolving agents. Our code and data are available at https://github.com/ShaoShuai0605/Misevolution . Warning: this paper includes examples that may be offensive or harmful in nature.

arXiv:2512.16301v3 Announce Type: replace Abstract: Large language model (LLM) agents are moving beyond prompting alone. ChatGPT marked the rise of general-purpose LLM assistants, DeepSeek showed that on-policy reinforcement learning with verifiable rewards can improve reasoning and tool use, and OpenClaw highlights a newer direction in which agents accumulate persistent memory and reusable skills. Yet the research landscape remains fragmented across post-training, retrieval, memory, and skill systems. This survey studies these developments under a single notion of \emph{adaptation}: improving an agent, its tools, or their interaction after pretraining. We organize the field with a four-paradigm framework spanning agent adaptation and tool adaptation. On the agent side, A1 (tool-execution-signaled) and A2 (agent-output-signaled) improve the agent itself through supervised fine-tuning, preference optimization, and reinforcement learning with verifiable rewards. On the tool side, T1 (agent-agnostic) provides reusable pre-trained modules any agent can call, while T2 (agent-supervised) uses the agent's outputs to train memory systems, skill libraries, or lightweight subagents. Using this framework, we review post-training methods, adaptive memory architectures, and agent skills; compare their trade-offs in cost, flexibility, and generalization; and summarize evaluation practices across deep research, software development, computer use, and drug discovery. We conclude by outlining open problems in agent-tool co-adaptation, continual learning, safety, and efficient deployment.

arXiv:2602.21534v2 Announce Type: replace Abstract: Agentic reinforcement learning (ARL) has rapidly gained attention as a promising paradigm for training agents to solve complex, multi-step interactive tasks. Despite encouraging early results, ARL remains highly unstable, often leading to training collapse. This instability limits scalability to larger environments and longer interaction horizons, and constrains systematic exploration of algorithmic design choices. In this paper, we first propose ARLArena, a stable training recipe and systematic analysis framework that examines training stability in a controlled and reproducible setting. ARLArena first constructs a clean and standardized testbed. Then, we decompose policy gradient into four core design dimensions and assess the performance and stability of each dimension. Through this fine-grained analysis, we distill a unified perspective on ARL and propose SAMPO, a stable agentic policy optimization method designed to mitigate the dominant sources of instability in ARL. Empirically, SAMPO achieves consistently stable training and strong performance across diverse agentic tasks. Overall, this study provides a unifying policy gradient perspective for ARL and offers practical guidance for building stable and reproducible LLM-based agent training pipelines.

arXiv:2201.05000v2 Announce Type: replace-cross Abstract: Reinforcement Learning and, recently, Deep Reinforcement Learning are popular methods for solving sequential decision-making problems modeled as Markov Decision Processes. RL modeling of a problem and selecting algorithms and hyper-parameters require careful consideration, as different configurations may entail completely different performances. These considerations are mainly the task of RL experts; however, RL is progressively becoming popular in other fields, such as combinatorial optimization, where researchers and system designers are not necessarily RL experts. Besides, many modeling decisions are typically made manually, such as defining state and action space, size of batches, batch update frequency, and time steps. For these reasons, automating different components of RL is of great importance, and it has attracted much attention in recent years. Automated RL provides a framework in which different components of RL, including MDP modeling, algorithm selection, and hyper-parameter optimization, are modeled and defined automatically. In this article, we present the literature on automated RL (AutoRL), including the recent large language model (LLM) based techniques. We also discuss the recent work on techniques that are not presently tailored for automated RL but hold promise for future integration into AutoRL. Furthermore, we discuss the challenges, open questions, and research directions in AutoRL.

arXiv:2505.19719v2 Announce Type: replace-cross Abstract: Common Neighbors (CNs) and their higher-order variants are important pairwise features widely used in state-of-the-art link prediction methods. However, existing methods often struggle with the repetition across different orders of CNs and fail to fully leverage their potential. We identify that these limitations stem from two key issues: redundancy and over-smoothing in high-order common neighbors. To address these challenges, we design orthogonalization to eliminate redundancy between different-order CNs and normalization to mitigate over-smoothing. By combining these two techniques, we propose Orthogonal Common Neighbor (OCN), a novel approach that significantly outperforms the strongest baselines by an average of 7.7\% on popular link prediction benchmarks. A thorough theoretical analysis is provided to support our method. Ablation studies also verify the effectiveness of our orthogonalization and normalization techniques. Code is available at: https://github.com/qingpingmo/OCN.

arXiv:2511.18152v2 Announce Type: replace-cross Abstract: Deep unfolding networks (DUNs) combine the interpretability of model-based methods with the learning ability of deep networks, yet remain limited for blind image restoration (BIR). Existing DUNs suffer from: (1) \textbf{Degradation-specific dependency}, as their optimization frameworks are tied to a known degradation model, making them unsuitable for BIR tasks; and (2) \textbf{Over-smoothing bias}, resulting from the direct feeding of gradient descent outputs, dominated by low-frequency content, into the proximal term, suppressing fine textures. To overcome these issues, we propose UnfoldLDM to integrate DUNs with latent diffusion model (LDM) for BIR. In each stage, UnfoldLDM employs a multi-granularity degradation-aware (MGDA) module as the gradient descent step. MGDA models BIR as an unknown degradation estimation problem and estimates both the holistic degradation matrix and its decomposed forms, enabling robust degradation removal. For the proximal step, we design a degradation-resistant LDM (DR-LDM) to extract compact degradation-invariant priors from the MGDA output. Guided by this prior, an over-smoothing correction transformer (OCFormer) explicitly recovers high-frequency components and enhances texture details. This unique combination ensures the final result is degradation-free and visually rich. Experiments show that our UnfoldLDM achieves a leading place on various BIR tasks and benefits downstream tasks. Moreover, our design is compatible with existing DUN-based methods, serving as a plug-and-play framework. Code will be released.

arXiv:2603.00040v2 Announce Type: replace-cross Abstract: Achieving reliable 4-bit attention is a prerequisite for end-to-end FP4 computation on emerging FP4-capable GPUs, yet attention remains the main obstacle due to FP4's tiny dynamic range and attention's heavy-tailed activations. This paper presents the first systematic study of 4-bit quantization-aware training (QAT) for attention. We find that "drop-in" QAT, which naively combines an FP4 forward pass with a high-precision Flash Attention (FA)-style backward pass, leads to training instability. We identify two key principles for stable FP4 attention: (1) matching low-precision recomputation of attention scores in the backward pass, and (2) resolving implicit precision assumptions in FA's gradient calculation. Based on these insights, we propose Attn-QAT and implement fused Triton kernels for training as well as FP4 inference kernels. Across diffusion and language models, Attn-QAT recovers the quality drop from FP4 attention without explicit outlier-mitigation heuristics used in prior FP4 attention, and delivers up to a 1.5x speedup on an RTX 5090. Video demos can be found at https://drive.google.com/drive/folders/190F6xbBDUF2kGQYIcXBt3ehSYij5jlim?usp=sharing.

arXiv:2603.00117v2 Announce Type: replace-cross Abstract: Living organisms exhibit persistent autonomy through internally generated goals and self-sustaining behavioral organization, yet current embodied agents remain driven by externally scripted objectives. This dependence on predefined task specifications limits their capacity for long-term deployment in dynamic, unstructured environments where continuous human intervention is impractical. We propose that personality traits provide an intrinsic organizational principle for achieving persistent autonomy. Analogous to genotypic biases shaping biological behavioral tendencies, personalities enable agents to autonomously generate goals and sustain behavioral evolution without external supervision. To realize this, we develop PEPA, a three-layer cognitive architecture that operates through three interacting systems: Sys3 autonomously synthesizes personality-aligned goals and refines them via episodic memory and daily self-reflection; Sys2 performs deliberative reasoning to translate goals into executable action plans; Sys1 grounds the agent in sensorimotor interaction, executing actions and recording experiences. We validate the framework through real-world deployment on a quadruped robot in a multi-floor office building. Operating without reliance on fixed task specifications, the robot autonomously arbitrates between user requests and personality-driven motivations, navigating elevators and exploring environments accordingly. Quantitative analysis across five distinct personality prototypes demonstrates stable, trait-aligned behaviors. The results confirm that personality-driven cognitive architectures enable sustained autonomous operation characteristic of persistent embodied systems. Code and demo videos are available at https://sites.google.com/view/pepa-persistent/.

arXiv:2603.04385v2 Announce Type: replace-cross Abstract: Feed-forward transformer models have driven rapid progress in 3D vision, but state-of-the-art methods such as VGGT and $\pi^3$ have a computational cost that scales quadratically with the number of input images, making them inefficient when applied to large image collections. Sequential-reconstruction approaches reduce this cost but sacrifice reconstruction quality. We introduce ZipMap, a stateful feed-forward model that achieves linear-time, bidirectional 3D reconstruction while matching or surpassing the accuracy of quadratic-time methods. ZipMap employs test-time training layers to zip an entire image collection into a compact hidden scene state in a single forward pass, enabling reconstruction of over 700 frames in under 10 seconds on a single H100 GPU, more than $20\times$ faster than state-of-the-art methods such as VGGT. Moreover, we demonstrate the benefits of having a stateful representation in real-time scene-state querying and its extension to sequential streaming reconstruction.

Nature, Published online: 10 March 2026; doi:10.1038/s41586-026-10351-8

Maximizing carrier extraction in hybrid back-contact silicon solar cells

Oncogene, Published online: 07 March 2026; doi:10.1038/s41388-026-03712-0

Arginine methylation-dependent stabilization of SUV39H1 promotes breast cancer growth

Oncogene, Published online: 05 March 2026; doi:10.1038/s41388-026-03713-z

KRAS-extrachromosomal DNA drives intratumoral heterogeneity in gastric cancer

arXiv:2603.03655v1 Announce Type: new Abstract: Tool-augmented large language model (LLM) agents promise to unify scientific reasoning with computation, yet their deployment in high-stakes domains like drug discovery is bottlenecked by two critical barriers: unconstrained tool-use governance and poor long-horizon reliability. In dependency-heavy pharmaceutical pipelines, autonomous agents often drift into irreproducible trajectories, where early-stage hallucinations multiplicatively compound into downstream failures. To overcome this, we present Mozi, a dual-layer architecture that bridges the flexibility of generative AI with the deterministic rigor of computational biology. Layer A (Control Plane) establishes a governed supervisor--worker hierarchy that enforces role-based tool isolation, limits execution to constrained action spaces, and drives reflection-based replanning. Layer B (Workflow Plane) operationalizes canonical drug discovery stages -- from Target Identification to Lead Optimization -- as stateful, composable skill graphs. This layer integrates strict data contracts and strategic human-in-the-loop (HITL) checkpoints to safeguard scientific validity at high-uncertainty decision boundaries. Operating on the design principle of ``free-form reasoning for safe tasks, structured execution for long-horizon pipelines,'' Mozi provides built-in robustness mechanisms and trace-level audibility to completely mitigate error accumulation. We evaluate Mozi on PharmaBench, a curated benchmark for biomedical agents, demonstrating superior orchestration accuracy over existing baselines. Furthermore, through end-to-end therapeutic case studies, we demonstrate Mozi's ability to navigate massive chemical spaces, enforce stringent toxicity filters, and generate highly competitive in silico candidates, effectively transforming the LLM from a fragile conversationalist into a reliable, governed co-scientist.

arXiv:2603.03637v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) integrate vision and text to power applications, but this integration introduces new vulnerabilities. We study Image-based Prompt Injection (IPI), a black-box attack in which adversarial instructions are embedded into natural images to override model behavior. Our end-to-end IPI pipeline incorporates segmentation-based region selection, adaptive font scaling, and background-aware rendering to conceal prompts from human perception while preserving model interpretability. Using the COCO dataset and GPT-4-turbo, we evaluate 12 adversarial prompt strategies and multiple embedding configurations. The results show that IPI can reliably manipulate the output of the model, with the most effective configuration achieving up to 64\% attack success under stealth constraints. These findings highlight IPI as a practical threat in black-box settings and underscore the need for defenses against multimodal prompt injection.

arXiv:2603.03805v1 Announce Type: cross Abstract: Relational Databases (RDBs) are the backbone of modern business, yet they lack foundation models comparable to those in text or vision. A key obstacle is that high-quality RDBs are private, scarce and structurally heterogeneous, making internet-scale pre-training infeasible. To overcome this data scarcity, We introduce $\textbf{RDB-PFN}$, the first relational foundation model trained purely via $\textbf{synthetic data}$. Inspired by Prior-Data Fitted Networks (PFNs) where synthetic data generated from Structural Causal Models (SCMs) enables reasoning on single tables, we design a $\textbf{Relational Prior Generator}$ to create an infinite stream of diverse RDBs from scratch. Pre-training on $\textbf{over 2 million}$ synthetic single-table and relational tasks, RDB-PFN learns to adapt to any new database instantly via genuine $\textbf{in-context learning}$. Experiments verify RDB-PFN achieves strong few-shot performance on 19 real-world relational prediction tasks, outperforming graph-based and single-table foundation-model baselines (given the same DFS-linearized inputs), while using a lightweight architecture and fast inference. The code is available at https://github.com/MuLabPKU/RDBPFN

arXiv:2603.03955v1 Announce Type: cross Abstract: Post-training with reinforcement learning (RL) has recently shown strong promise for advancing multimodal agents beyond supervised imitation. However, RL remains limited by poor data efficiency, particularly in settings where interaction data are scarce and quickly become outdated. To address this challenge, GIPO (Gaussian Importance sampling Policy Optimization) is proposed as a policy optimization objective based on truncated importance sampling, replacing hard clipping with a log-ratio-based Gaussian trust weight to softly damp extreme importance ratios while maintaining non-zero gradients. Theoretical analysis shows that GIPO introduces an implicit, tunable constraint on the update magnitude, while concentration bounds guarantee robustness and stability under finite-sample estimation. Experimental results show that GIPO achieves state-of-the-art performance among clipping-based baselines across a wide range of replay buffer sizes, from near on-policy to highly stale data, while exhibiting superior bias--variance trade-off, high training stability and improved sample efficiency.

arXiv:2603.04128v1 Announce Type: cross Abstract: Developing Audio-Visual Large Language Models (AV-LLMs) for unified scene understanding is pivotal in multimodal intelligence. While instruction tuning enables pre-trained models with multi-task abilities, we observe that conventional multi-task unification methods often suffer from severe negative transfer, where nearly 55% of tasks degrade compared to single-task training. We attribute this phenomenon to audio-visual task heterogeneity, characterized by disparate task granularity and divergent capability demands, which lead to negative interference under joint training. To tackle this, we present Crab$^{+}$, a scalable and unified audio-visual scene understanding model that addresses task heterogeneity through explicit cooperation from both data and model perspectives. On the data side, we introduce AV-UIE v2, a comprehensive Audio-Visual Unified Instruction-tuning dataset with Explicit reasoning processes. It contains approximately 222K samples spanning 17 datasets and 7 tasks, enabling the model to capture cross-task relationships at different levels of granularity. On the model side, we design a unified interface to align heterogeneous task formulations, and propose Interaction-aware LoRA (I-LoRA), which explicitly models inter-task relationships via dynamic routing to coordinate distinct audio-visual interaction patterns, mitigating parameter interference. Extensive experiments show Crab$^{+}$ covers broader tasks than existing unified models while outperforming specialized models on various benchmarks. We successfully reverse the negative transfer trend, achieving positive transfer where multi-task learning surpasses single-task baselines in nearly 88% of tasks. These results hold across diverse AV-LLM paradigms and are validated through in-depth visualization, positioning Crab$^{+}$ as a robust step towards holistic audio-visual scene understanding.

arXiv:2603.04222v1 Announce Type: cross Abstract: Multimodal perception enables robust autonomous driving but incurs unnecessary computational cost when all sensors remain active. This paper presents PRAM-R, a unified Perception-Reasoning-Action-Memory framework with LLM-Guided Modality Routing for adaptive autonomous driving. PRAM-R adopts an asynchronous dual-loop design: a fast reactive loop for perception and control, and a slow deliberative loop for reasoning-driven modality selection and memory updates. An LLM router selects and weights modalities using environmental context and sensor diagnostics, while a hierarchical memory module preserves temporal consistency and supports long-term adaptation. We conduct a two-stage evaluation: (1) synthetic stress tests for stability analysis and (2) real-world validation on the nuScenes dataset. Synthetic stress tests confirm 87.2% reduction in routing oscillations via hysteresis-based stabilization. Real-world validation on nuScenes shows 6.22% modality reduction with 20% memory recall while maintaining comparable trajectory accuracy to full-modality baselines in complex urban scenarios. Our work demonstrates that LLM-augmented architectures with hierarchical memory achieve efficient, adaptive multimodal perception in autonomous driving.