Background: Family caregivers of children with chronic health conditions experience substantial physical and mental health burdens, including burnout, anxiety, depression, fatigue, and sleep disturbances. Despite this need, validated digital mental health tools tailored to family caregivers remain limited. AI-powered conversational agents offer a promising approach for delivering on-demand, personalized mental health support, yet development and evaluation frameworks for this population are lacking. Objective: This paper describes the iterative development and formative evaluation of COCO (Caring of Caregivers Online), a conversational agent designed for family caregivers of children with chronic health conditions. COCO integrates problem-solving therapy (PST) and motivational interviewing (MI) within a human-in-the-loop development framework that progressed from rule-based interactions to a large language model (LLM)–powered conversational agent. Methods: COCO was developed across four phases: (1) caregiver persona and dialogue development based on PST and MI; (2) usability testing of a low-fidelity prototype with standardized patients in a single session of PST; (3) usability testing of a high-fidelity prototype with caregivers in a single session of PST (n=38); (4) integration of an LLM into COCO. The Wizard-of-Oz method was used across phases 2 and 3 to collect naturalistic dialogues and refine COCO’s conversational design. In phase 3, usability of COCO was assessed using the System Usability Scale (SUS). Caregiver emotions were measured before and after the session using 6 subscales of the PANAS-X. In phase 4, GPT-4 was integrated into COCO with few-shot learning and evaluated by research team members using the caregiver personas. Descriptive statistics were used to summarize quantitative measures. The MI principles and techniques used by COCO across the 4 phases were coded using the . Results: In phase 1, 4 gold-standard dialogues were developed using caregiver personas. In phase 2, standardized patients described COCO as validating and identified its problem-solving and on-demand support as helpful for caregivers. In phase 3, COCO-Wizard-of-Oz achieved a mean SUS score of 75.6% (SD 12.9%), reflecting acceptable usability. Participants demonstrated significant improvement in negative affect, sadness, guilt, and fatigue following PST sessions (

arXiv:2605.25984v1 Announce Type: cross Abstract: Ensuring safe and contextually appropriate behaviour in Large Language Models (LLMs) remains a critical challenge for real-world deployment. We present \textbf{SafeCtrl-RL}, an inference-time behavioural control framework that enables adaptive safety regulation without model retraining or parameter modification. The method formulates dialogue generation as a sequential decision process, where a reinforcement learning agent dynamically selects prompt adjustment strategies based on contextual feedback. This allows unsafe behaviours to be suppressed through iterative refinement, which we conceptualise as inference-time behavioural unlearning. Evaluated across multiple LLMs and unsafe dialogue scenarios, SafeCtrl-RL consistently improves safety and response quality, outperforms existing prompt-based optimisation methods, and achieves favourable performance--efficiency trade-offs. **Warning: This paper may contain examples of harmful language, and reader discretion is recommended.

Oncogene, Published online: 16 April 2026; doi:10.1038/s41388-026-03794-w

XPO1 inhibitor KPT-330 disrupts the core transcriptional regulatory circuitry of dedifferentiated liposarcoma by modulating the translation process

npj Digital Medicine, Published online: 07 April 2026; doi:10.1038/s41746-026-02593-7

Lightweight liquid neural networks decipher salivary metabolic fingerprinting for high-risk periodontitis screening in diabetes

arXiv:2604.01618v1 Announce Type: cross Abstract: Vision-language-action (VLA) models have shown strong performance in robotic manipulation, yet their robustness to physically realizable adversarial attacks remains underexplored. Existing studies reveal vulnerabilities through language perturbations and 2D visual attacks, but these attack surfaces are either less representative of real deployment or limited in physical realism. In contrast, adversarial 3D textures pose a more physically plausible and damaging threat, as they are naturally attached to manipulated objects and are easier to deploy in physical environments. Bringing adversarial 3D textures to VLA systems is nevertheless nontrivial. A central obstacle is that standard 3D simulators do not provide a differentiable optimization path from the VLA objective function back to object appearance, making it difficult to optimize through an end-to-end manner. To address this, we introduce Foreground-Background Decoupling (FBD), which enables differentiable texture optimization through dual-renderer alignment while preserving the original simulation environment. To further ensure that the attack remains effective across long-horizon and diverse viewpoints in the physical world, we propose Trajectory-Aware Adversarial Optimization (TAAO), which prioritizes behaviorally critical frames and stabilizes optimization with a vertex-based parameterization. Built on these designs, we present Tex3D, the first framework for end-to-end optimization of 3D adversarial textures directly within the VLA simulation environment. Experiments in both simulation and real-robot settings show that Tex3D significantly degrades VLA performance across multiple manipulation tasks, achieving task failure rates of up to 96.7\%. Our empirical results expose critical vulnerabilities of VLA systems to physically grounded 3D adversarial attacks and highlight the need for robustness-aware training.

arXiv:2603.29112v1 Announce Type: new Abstract: We introduce GISTBench, a benchmark for evaluating Large Language Models' (LLMs) ability to understand users from their interaction histories in recommendation systems. Unlike traditional RecSys benchmarks that focus on item prediction accuracy, our benchmark evaluates how well LLMs can extract and verify user interests from engagement data. We propose two novel metric families: Interest Groundedness (IG), decomposed into precision and recall components to separately penalize hallucinated interest categories and reward coverage, and Interest Specificity (IS), which assesses the distinctiveness of verified LLM-predicted user profiles. We release a synthetic dataset constructed on real user interactions on a global short-form video platform. Our dataset contains both implicit and explicit engagement signals and rich textual descriptions. We validate our dataset fidelity against user surveys, and evaluate eight open-weight LLMs spanning 7B to 120B parameters. Our findings reveal performance bottlenecks in current LLMs, particularly their limited ability to accurately count and attribute engagement signals across heterogeneous interaction types.

arXiv:2603.06664v1 Announce Type: cross Abstract: Diffusion Transformer (DiT)-based video generation models inherently suffer from bottlenecks in long video synthesis and real-time inference, which can be attributed to the use of full spatiotemporal attention. Specifically, this mechanism leads to explosive O(N^2) memory consumption and high first-frame latency. To address these issues, we implement system-level inference optimizations for a causal autoregressive video generation pipeline. We adapt the Self-Forcing causal autoregressive framework to sequence parallel inference and implement a sequence-parallel variant of the causal rotary position embedding which we refer to as Causal-RoPE SP. This adaptation enables localized computation and reduces cross-rank communication in sequence parallel execution. In addition, computation and communication pipelines are optimized through operator fusion and RoPE precomputation. Experiments conducted on an eight GPU A800 cluster show that the optimized system achieves comparable generation quality, sub-second first-frame latency, and near real-time inference speed. For generating five second 480P videos, a 1.58x speedup is achieved, thereby providing effective support for real-time interactive applications.

arXiv:2602.13977v1 Announce Type: cross Abstract: Reinforcement learning (RL) promises to unlock capabilities beyond imitation learning for Vision-Language-Action (VLA) models, but its requirement for massive real-world interaction prevents direct deployment on physical robots. Recent work attempts to use learned world models as simulators for policy optimization, yet closed-loop imagined rollouts inevitably suffer from hallucination and long-horizon error accumulation. Such errors do not merely degrade visual fidelity; they corrupt the optimization signal, encouraging policies to exploit model inaccuracies rather than genuine task progress. We propose WoVR, a reliable world-model-based reinforcement learning framework for post-training VLA policies. Instead of assuming a faithful world model, WoVR explicitly regulates how RL interacts with imperfect imagined dynamics. It improves rollout stability through a controllable action-conditioned video world model, reshapes imagined interaction to reduce effective error depth via Keyframe-Initialized Rollouts, and maintains policy-simulator alignment through World Model-Policy co-evolution. Extensive experiments on LIBERO benchmarks and real-world robotic manipulation demonstrate that WoVR enables stable long-horizon imagined rollouts and effective policy optimization, improving average LIBERO success from 39.95% to 69.2% (+29.3 points) and real-robot success from 61.7% to 91.7% (+30.0 points). These results show that learned world models can serve as practical simulators for reinforcement learning when hallucination is explicitly controlled.